Setup your entire back office in one place with 10x better workflows using AI. Banking, Cards, Bill Payments, Bookkeeping, Tax, HR, Payroll, and Benefits. www.every.io

Every

every

Every AI Prediction That Came True in 2024—and the Ones That Didn't

Todas las predicciones de IA que se cumplieron en 2024 (y las que no)

Toutes les prédictions de l'IA qui se sont réalisées en 2024, et celles qui ne se sont pas réalisées

Todas as predicións da IA que se fixeron realidade en 2024 e as que non o fixeron

2024-ին իրականություն դարձած յուրաքանչյուր AI կանխատեսում, և նրանք, որոնք չեն իրականացել

כל תחזית AI שהתגשמה בשנת 2024 - ואלה שלא

2024 年に実現したすべての AI 予測と実現しなかった予測

ხელოვნური ინტელექტის ყველა პროგნოზი, რომელიც ახდა 2024 წელს და ის, რაც არ შესრულდა

រាល់ការទស្សន៍ទាយ AI ដែលបានក្លាយជាការពិតនៅឆ្នាំ 2024 ហើយអ្វីដែលមិនបានកើតឡើង

2024년에 실현된 모든 AI 예측과 실현되지 않은 예측

Wszystkie prognozy dotyczące sztucznej inteligencji, które spełniły się w 2024 r. — i te, które się nie spełniły

Todas as previsões de IA que se concretizaram em 2024 — e as que não se concretizaram

2024 දී සැබෑ වූ සෑම AI අනාවැකියක්ම - සහ නොකළ ඒවා

The HackerNoon Newsletter: Next Holiday Season, Ignore Everyone Except One Customer (12/21/2024)

Holiday marketing, as it’s preached, is a con. A ritualized scam sold to startups by ad platforms, agencies, and marketers who profit from your desperation. Do you really think you can outspend Amazon, out-charm Coca-Cola, or outmaneuver Nike? Unless you’re secretly printing money in your garage, no. Christmas isn’t for playing Santa with jingles and 20% off codes—it’s for hunting a whale. Next year, turn your entire marketing strategy into a public obsession with one person. Not a “demographic market.” Not a target persona. A single, breathing, flesh-and-blood human being whose “yes”could grant you permanent access to Cuban’s Cabo cocktail club. Bet it all on a campaign so precise it’s borderline psychotic. Dangerous? Stupid? Obviously, and probably. But boring? Never. If it crashes, people will talk. If it flies, you’ll own the year. Either is better than begging the Internet for clicks, plus it’ll give you a great story (or a great excuse to drink). Holiday Marketing Sucks. This is a classic David vs. Goliath scenario—except Goliath has a machine gun and David forgot his slingshot. Here’s why it’s a losing battle: You’re Outgunned And Outclassed Every big brand with deep pockets is jacking up CPMs. Ad costs more than double, competition spikes, and algorithms prioritize whoever spends the most. Yeah, obviously not you. If your $10K campaign could talk, it would whisper: "Help." You’re Selling To Zombies By mid-December, your CTA is competing with drunken relatives, shipping delays, and 87 ads about the perfect gift for canine therapists. Your audience is fried and their brains are coasting on autopilot—barely capable of choosing between two brands of wrapping paper. You’re Competing With Santa Claus—Not Other Businesses Know why Apple kills it every Christmas? It’s not with Siri, but with sentiment. The holiday giants use their war chests to craft masterpieces that ride on the emotional meaning of the season—nostalgia, tradition, family values—and if you’re not on par, you come off like a door-to-door salesman in a Saint Nick suit. ROI = Regret On Investment December sales bring January sobs. Whatever you make from your festive push will be eaten alive by your CPA—and you’ll wake up next month drowning in remorse, invoices, and wondering where the cash went. (Hint: Zuck’s back pocket) Why Serenading One Person Is The Smartest Play The best presents don’t come in mass-produced packages. Neither should your holiday marketing. A custom campaign is as much a gift to them as it is to you. It’s Psychological Warfare People hate being sold to, but they love feeling chosen. Focusing on one person creates a psychological bind: If they respond, they know they’re winning because you’ve done all the work to court them.If they don’t respond, they wonder if someone else will—and if they’ll miss their shot. It’s The Ultimate Litmus Test For Your Product If you can’t convince one perfect-fit customer, how will you convince hundreds? A campaign por uno is the fastest way to figure out if your pitch, product, and strategy are sharp enough to scale. It’s The Best Marketing Boot Camp A one-person campaign is combat training. No crutches of generic strategies, no Canva, no “but our CTR was decent” excuses—survival here relies on raw ingenuity, superhuman agility, and uncompromising precision. Success isn’t guaranteed, but clarity is. The hard lessons will stick, and your team becomes immune to mediocrity. After this, they’ll be more than competent: they’ll be dangerous. It Turns Failure Into Success Even if it “fails,” it doesn’t. People appreciate balls. The audacity of your move becomes its own currency, generating buzz, controversy, and respect—all of which could bag you a new client, partner, or investor. In a one-person campaign, failure is just another kind of success. How To Execute A Campaign That Borders On Maniacal Step 1: Pick Your Whale A Dream Customer: E.g., the enterprise client whose logo would give your sales team goosebumps.A Key Partner: E.g., the CMO that could double your reach if they said yes.An Influential Investor: Not just for the check, but for the name on the cap table. Be ambitious, but realistic. Elon’s too busy running Government Efficiency. Aim for the guys two rungs below him. Step 2: Stalk Your Prey Just because you’re targeting one person doesn’t mean you get to skip the user insight phase. Oh no, this research matters even more when you’re professing love, and it doesn’t stop at their LinkedIn profile. Do they like having their ego massaged? Have a fetish for cryptic puzzles? Are they nosing around for gaming collaborations? Stalk (ethically), snoop (strategically), and build a dossier so good it’d make the FBI sweat. Just remember there’s a fine line between research and “why is there a black van outside my house?” Step 3: Unleash Controlled Madness. Now that you understand your whale’s appetite, cook up the bait. Go brazen, or don’t bother. But your bravado and performance art means nothing without an unapologetic CTA and irresistible pitch. That ask needs to be so direct it’s borderline confrontational, and that pitch better feel like a bespoke Italian suit. Here are some thought-starters: Sponsor Their Daily Coffee: Find out where they get their morning coffee and prepay for a week of their orders. Include a personalized note on every receipt that says: “Fuel for your next great decision—working with us.” Ship A Literal Door: Send them a door (yes, an actual door) with a key attached and a note: “This is the key to unlocking something extraordinary. Let’s talk.” Buy An Editorial: Call up some prestigious publication they subscribe to and splurge on a paid piece. Write an open letter addressed to them, full of specifics about why they should engage with you. Be cheeky: “Hey [Their Name], Playing Small Looks Weird on You.” Step 4: Make Sure The World Knows You’re Crazy. Exclusivity is no fun if no one’s watching. Turn your serenade into a spectacle. More thought-starters: Build a Public Countdown: Launch a countdown clock on social media, with posts teasing, “6 Days Until [Target’s Name] Changes Everything.” Make sure they’re mentioned, their employees are tagged, their peers are targeted. Go Guerrilla (Literally): Hire street performers or actors to “protest” outside their HQ, carrying signs with slogans like, “[Target’s Name], Take the Meeting!”. Extra karma if it’s themed to their industry or product. Rent Their Attention: Buy ad space outside their office. A billboard. Graffiti a building (legally). Be loud to the point you start the water cooler gossip: “Dear [Name], If You Don’t Call Us, We’ll Have to Buy More Billboards. And That’s Awkward for Everyone.” Final Thoughts: Maybe It’s Not You—Maybe It’s Just Dumb Sometimes, you’re not bad at the game—the game is (likely) a waste of time. Holiday marketing for startups? That’s the house always winning, with you just the festive confetti. Stop groveling for attention in a system that turns startups into cannon fodder for ad giants. That’s not persistence—it’s herd mentality dressed up as hustle. Do something irrationally ambitious. Go unhinged, go nuclear. Not because it’s guaranteed success, but because it’s not preordained defeat. At least you’ll have had some fun, and failed in style—on your own terms, not theirs. It’s Christmas. Use the season for what it’s meant for—walking away from bad ideas. Holiday marketing, as it’s preached, is a con. A ritualized scam sold to startups by ad platforms, agencies, and marketers who profit from your desperation. Do you really think you can outspend Amazon, out-charm Coca-Cola, or outmaneuver Nike? Unless you’re secretly printing money in your garage, no. Holiday marketing, as it’s preached, is a con. Christmas isn’t for playing Santa with jingles and 20% off codes—it’s for hunting a whale. Next year, turn your entire marketing strategy into a public obsession with one person. Not a “demographic market.” Not a target persona. A single, breathing, flesh-and-blood human being whose “yes”could grant you permanent access to Cuban’s Cabo cocktail club. Next year, turn your entire marketing strategy into a public obsession with one person. Bet it all on a campaign so precise it’s borderline psychotic . Dangerous? Stupid? Obviously, and probably. But boring? Never. so precise it’s borderline psychotic so precise it’s borderline psychotic But boring? Never. If it crashes, people will talk. If it flies, you’ll own the year. Either is better than begging the Internet for clicks, plus it’ll give you a great story (or a great excuse to drink). Holiday Marketing Sucks. Holiday Marketing Sucks. This is a classic David vs. Goliath scenario— except Goliath has a machine gun and David forgot his slingshot . Here’s why it’s a losing battle: except Goliath has a machine gun and David forgot his slingshot except Goliath has a machine gun and David forgot his slingshot You’re Outgunned And Outclassed You’re Outgunned And Outclassed Every big brand with deep pockets is jacking up CPMs. Ad costs more than double, competition spikes, and algorithms prioritize whoever spends the most. Yeah, obviously not you. If your $10K campaign could talk, it would whisper: "Help." If your $10K campaign could talk, it would whisper: "Help." You’re Selling To Zombies You’re Selling To Zombies By mid-December, your CTA is competing with drunken relatives, shipping delays, and 87 ads about the perfect gift for canine therapists. Your audience is fried and their brains are coasting on autopilot—barely capable of choosing between two brands of wrapping paper. You’re Competing With Santa Claus—Not Other Businesses You’re Competing With Santa Claus—Not Other Businesses Know why Apple kills it every Christmas? It’s not with Siri, but with sentiment . The holiday giants use their war chests to craft masterpieces that ride on the emotional meaning of the season —nostalgia, tradition, family values—and if you’re not on par, you come off like a door-to-door salesman in a Saint Nick suit. sentiment that ride on the emotional meaning of the season that ride on the emotional meaning of the season ROI = Regret On Investment ROI = Regret On Investment December sales bring January sobs. Whatever you make from your festive push will be eaten alive by your CPA —and you’ll wake up next month drowning in remorse, invoices, and wondering where the cash went. (Hint: Zuck’s back pocket) Whatever you make from your festive push will be eaten alive by your CPA Whatever you make from your festive push will be eaten alive by your CPA Why Serenading One Person Is The Smartest Play Why Serenading One Person Is The Smartest Play The best presents don’t come in mass-produced packages. Neither should your holiday marketing. A custom campaign is as much a gift to them as it is to you. A custom campaign is as much a gift to them as it is to you. A custom campaign is as much a gift to them as it is to you. It’s Psychological Warfare It’s Psychological Warfare People hate being sold to, but they love feeling chosen. Focusing on one person creates a psychological bind: If they respond, they know they’re winning because you’ve done all the work to court them. If they don’t respond, they wonder if someone else will—and if they’ll miss their shot. If they respond, they know they’re winning because you’ve done all the work to court them. If they don’t respond, they wonder if someone else will—and if they’ll miss their shot. It’s The Ultimate Litmus Test For Your Product It’s The Ultimate Litmus Test For Your Product If you can’t convince one perfect-fit customer, how will you convince hundreds? A campaign por uno is the fastest way to figure out if your pitch, product, and strategy are sharp enough to scale. It’s The Best Marketing Boot Camp It’s The Best Marketing Boot Camp A one-person campaign is combat training. No crutches of generic strategies, no Canva, no “but our CTR was decent” excuses—survival here relies on raw ingenuity, superhuman agility, and uncompromising precision. Success isn’t guaranteed, but clarity is. The hard lessons will stick, and your team becomes immune to mediocrity. After this, they’ll be more than competent: they’ll be dangerous. Success isn’t guaranteed, but clarity is. It Turns Failure Into Success It Turns Failure Into Success Even if it “fails,” it doesn’t. People appreciate balls. The audacity of your move becomes its own currency, generating buzz, controversy, and respect— all of which could bag you a new client, partner, or investor. In a one-person campaign, failure is just another kind of success. all of which could bag you a new client, partner, or investor. all of which could bag you a new client, partner, or investor. How To Execute A Campaign That Borders On Maniacal How To Execute A Campaign That Borders On Maniacal Step 1: Pick Your Whale Step 1: Pick Your Whale A Dream Customer: E.g., the enterprise client whose logo would give your sales team goosebumps. A Key Partner: E.g., the CMO that could double your reach if they said yes. An Influential Investor: Not just for the check, but for the name on the cap table. A Dream Customer: E.g., the enterprise client whose logo would give your sales team goosebumps. A Dream Customer: A Key Partner: E.g., the CMO that could double your reach if they said yes. A Key Partner: An Influential Investor: Not just for the check, but for the name on the cap table. An Influential Investor: Be ambitious, but realistic. Elon’s too busy running Government Efficiency. Aim for the guys two rungs below him. Step 2: Stalk Your Prey Step 2: Stalk Your Prey Just because you’re targeting one person doesn’t mean you get to skip the user insight phase . Oh no, this research matters even more when you’re professing love, and it doesn’t stop at their LinkedIn profile. Do they like having their ego massaged? Have a fetish for cryptic puzzles? Are they nosing around for gaming collaborations? doesn’t mean you get to skip the user insight phase doesn’t mean you get to skip the user insight phase Stalk (ethically), snoop (strategically), and build a dossier so good it’d make the FBI sweat. Just remember there’s a fine line between research and “why is there a black van outside my house?” Just remember there’s a fine line between research and “why is there a black van outside my house?” Step 3: Unleash Controlled Madness. Step 3: Unleash Controlled Madness. Now that you understand your whale’s appetite, cook up the bait. Go brazen, or don’t bother. But your bravado and performance art means nothing without an unapologetic CTA and irresistible pitch. That ask needs to be so direct it’s borderline confrontational , and that pitch better feel like a bespoke Italian suit. But your bravado and performance art means nothing without an unapologetic CTA and irresistible pitch. so direct it’s borderline confrontational Here are some thought-starters: Here are some thought-starters: Sponsor Their Daily Coffee: Find out where they get their morning coffee and prepay for a week of their orders. Include a personalized note on every receipt that says: “ Fuel for your next great decision—working with us. ” Sponsor Their Daily Coffee: Fuel for your next great decision—working with us. Ship A Literal Door: Send them a door (yes, an actual door) with a key attached and a note: “ This is the key to unlocking something extraordinary. Let’s talk. ” Ship A Literal Door: This is the key to unlocking something extraordinary. Let’s talk. Buy An Editorial: Call up some prestigious publication they subscribe to and splurge on a paid piece. Write an open letter addressed to them, full of specifics about why they should engage with you. Be cheeky: “ Hey [Their Name], Playing Small Looks Weird on You .” Buy An Editorial: Hey [Their Name], Playing Small Looks Weird on You Step 4: Make Sure The World Knows You’re Crazy. Step 4: Make Sure The World Knows You’re Crazy. Exclusivity is no fun if no one’s watching. Turn your serenade into a spectacle. More thought-starters: More thought-starters: Build a Public Countdown: Launch a countdown clock on social media, with posts teasing, “ 6 Days Until [Target’s Name] Changes Everything. ” Make sure they’re mentioned, their employees are tagged, their peers are targeted. Build a Public Countdown: 6 Days Until [Target’s Name] Changes Everything. Go Guerrilla (Literally) : Hire street performers or actors to “protest” outside their HQ, carrying signs with slogans like, “ [Target’s Name], Take the Meeting! ”. Extra karma if it’s themed to their industry or product. Go Guerrilla (Literally) [Target’s Name], Take the Meeting! Rent Their Attention: Buy ad space outside their office. A billboard. Graffiti a building (legally). Be loud to the point you start the water cooler gossip: “ Dear [Name], If You Don’t Call Us, We’ll Have to Buy More Billboards. And That’s Awkward for Everyone. ” Rent Their Attention: Dear [Name], If You Don’t Call Us, We’ll Have to Buy More Billboards. And That’s Awkward for Everyone. Final Thoughts: Maybe It’s Not You—Maybe It’s Just Dumb Final Thoughts: Maybe It’s Not You—Maybe It’s Just Dumb Sometimes, you’re not bad at the game—the game is (likely) a waste of time. Holiday marketing for startups? That’s the house always winning, with you just the festive confetti. Stop groveling for attention in a system that turns startups into cannon fodder for ad giants. That’s not persistence— it’s herd mentality dressed up as hustle . Do something irrationally ambitious. Go unhinged, go nuclear . Not because it’s guaranteed success, but because it’s not preordained defeat. At least you’ll have had some fun, and failed in style— on your own terms, not theirs . it’s herd mentality dressed up as hustle it’s herd mentality dressed up as hustle Go unhinged, go nuclear Not because it’s guaranteed success, but because it’s not preordained defeat. on your own terms, not theirs It’s Christmas. Use the season for what it’s meant for—walking away from bad ideas. It’s Christmas. Use the season for what it’s meant for—walking away from bad ideas.

Next Holiday Season, Ignore Everyone Except One Customer

I næste feriesæson, ignorer alle undtagen én kunde

La próxima temporada navideña, ignore a todos excepto a un cliente

Sljedeće blagdanske sezone zanemarite sve osim jednog kupca

次のホリデーシーズンは、1人の顧客以外全員を無視する

შემდეგი სადღესასწაულო სეზონი, იგნორირება გაუკეთეთ ყველას გარდა ერთი მომხმარებლისა

រដូវកាលឈប់សម្រាកបន្ទាប់ មិនអើពើនឹងអ្នកគ្រប់គ្នា លើកលែងតែអតិថិជនម្នាក់

다음 휴가 시즌에는 한 명의 고객을 제외한 모든 사람을 무시하세요

Saison ya ba fêtes oyo ekoya, Ignorer mutu nionso sauf Client moko

Hamuq Fiesta pacha, Llapa runata mana kasuychu, huk rantiqmanta aswanta

Следеће празничне сезоне, игноришите све осим једног купца

Susunod na Holiday Season, Balewalain ang Lahat Maliban sa Isang Customer

Keyingi bayram mavsumida bitta mijozdan tashqari hammaga e'tibor bermang

The TechBeat: Everything You Need to Know About the Lumoz Token Generation Event (12/18/2024)

Table of Links Abstract and 1. Introduction 2 Related Work 3 Preliminaries 4 Method 4.1 Key Sample and Joint Editing 4.2 Edit Propagation Via TokenFlow 5 Results 5.1 Qualitative Evaluation and 5.2 Quantitative Evaluation 5.3 Ablation Study 6 Discussion 7 Acknowledgement and References A Implementation Details A IMPLEMENTATION DETAILS StableDiffusion. We use Stable Diffusion as our pre-trained text-to-image model; we use the StableDiffusion-v-2-1 checkpoint provided via official HuggingFace webpage. DDIM inversion. In all of our experiments, we use DDIM deterministic sampling with 50 steps. For inverting the video, we follow Tumanyan et al. (2023) and use DDIM inversion with classifierfree guidance scale of 1 and 1000 forward steps; and extract the self-attention input tokens from this process similarly to Qi et al. (2023). Runtime. Since we don’t compute the attention module on most video frames (i.e., we only compute the self-attention output on the keyframes) our method is efficient in run-time, and the sampling of the video reduces the time of per-frame editing by 20%. The inversion process with 1000 steps is the main bottleneck of our method in terms of run-time, and in many cases a significantly smaller amount of steps is suffieicent (e.g. 50). Table 3 reports runtime comparisons using 50 steps in all methods. Notably, our sampling time is indeed faster than that of per-frame editing (PnP). Baselines. For running the baseline of Tune-a-video (Wu et al., 2022) we used their official repository. For Gen-1 (Esser et al., 2023) we used their platform on Runaway website. This platform outputs a video that is not in the same length and frame-rate as the input video; therefore, we could not compute the warping error on their results. For text-to-video-zero (Khachatryan et al., 2023b) we used their official repository, with their depth conditioning configuration. For Fate-Zero (Qi et al., 2023) with used their official repository, and verified the run configurations with the authors. This paper is available on arxiv under CC BY 4.0 DEED DEED license. Authors:
(1) Michal Geyer, Weizmann Institute of Science and Indicates equal contribution;
(2) Omer Bar-Tal, Weizmann Institute of Science and Indicates equal contribution;
(3) Shai Bagon, Weizmann Institute of Science;
(4) Tali Dekel, Weizmann Institute of Science. Table of Links Abstract and 1. Introduction Abstract and 1. Introduction 2 Related Work 2 Related Work 3 Preliminaries 3 Preliminaries 4 Method 4 Method 4.1 Key Sample and Joint Editing 4.1 Key Sample and Joint Editing 4.2 Edit Propagation Via TokenFlow 4.2 Edit Propagation Via TokenFlow 5 Results 5 Results 5.1 Qualitative Evaluation and 5.2 Quantitative Evaluation 5.1 Qualitative Evaluation and 5.2 Quantitative Evaluation 5.3 Ablation Study 5.3 Ablation Study 6 Discussion 6 Discussion 7 Acknowledgement and References 7 Acknowledgement and References A Implementation Details A Implementation Details A IMPLEMENTATION DETAILS StableDiffusion. We use Stable Diffusion as our pre-trained text-to-image model; we use the StableDiffusion-v-2-1 checkpoint provided via official HuggingFace webpage. StableDiffusion. DDIM inversion. In all of our experiments, we use DDIM deterministic sampling with 50 steps. For inverting the video, we follow Tumanyan et al. (2023) and use DDIM inversion with classifierfree guidance scale of 1 and 1000 forward steps; and extract the self-attention input tokens from this process similarly to Qi et al. (2023). DDIM inversion. Runtime. Since we don’t compute the attention module on most video frames (i.e., we only compute the self-attention output on the keyframes) our method is efficient in run-time, and the sampling of the video reduces the time of per-frame editing by 20%. The inversion process with 1000 steps is the main bottleneck of our method in terms of run-time, and in many cases a significantly smaller amount of steps is suffieicent (e.g. 50). Table 3 reports runtime comparisons using 50 steps in all methods. Notably, our sampling time is indeed faster than that of per-frame editing (PnP). Runtime. Baselines. For running the baseline of Tune-a-video (Wu et al., 2022) we used their official repository. For Gen-1 (Esser et al., 2023) we used their platform on Runaway website. This platform outputs a video that is not in the same length and frame-rate as the input video; therefore, we could not compute the warping error on their results. For text-to-video-zero (Khachatryan et al., 2023b) we used their official repository, with their depth conditioning configuration. For Fate-Zero (Qi et al., 2023) with used their official repository, and verified the run configurations with the authors. Baselines. This paper is available on arxiv under CC BY 4.0 DEED DEED license. This paper is available on arxiv under CC BY 4.0 DEED DEED license. available on arxiv Authors: (1) Michal Geyer, Weizmann Institute of Science and Indicates equal contribution; (2) Omer Bar-Tal, Weizmann Institute of Science and Indicates equal contribution; (3) Shai Bagon, Weizmann Institute of Science; (4) Tali Dekel, Weizmann Institute of Science. Authors: Authors: (1) Michal Geyer, Weizmann Institute of Science and Indicates equal contribution; (2) Omer Bar-Tal, Weizmann Institute of Science and Indicates equal contribution; (3) Shai Bagon, Weizmann Institute of Science; (4) Tali Dekel, Weizmann Institute of Science.

TokenFlow's Implementation Details: Everything That We Used

The TechBeat: Everything You Need to Know About the Lumoz Token Generation Event (12/16/2024)

Everything Apache Iceberg-Related Announced This Year

Authors:
(1) An Yan, UC San Diego, ayan@ucsd.edu;
(2) Zhengyuan Yang, Microsoft Corporation, zhengyang@microsoft.com with equal contributions;
(3) Wanrong Zhu, UC Santa Barbara, wanrongzhu@ucsb.edu;
(4) Kevin Lin, Microsoft Corporation, keli@microsoft.com;
(5) Linjie Li, Microsoft Corporation, lindsey.li@mocrosoft.com;
(6) Jianfeng Wang, Microsoft Corporation, jianfw@mocrosoft.com;
(7) Jianwei Yang, Microsoft Corporation, jianwei.yang@mocrosoft.com;
(8) Yiwu Zhong, University of Wisconsin-Madison, yzhong52@wisc.edu;
(9) Julian McAuley, UC San Diego, jmcauley@ucsd.edu;
(10) Jianfeng Gao, Microsoft Corporation, jfgao@mocrosoft.com;
(11) Zicheng Liu, Microsoft Corporation, zliu@mocrosoft.com;
(12) Lijuan Wang, Microsoft Corporation, lijuanw@mocrosoft.com. Editor’s note: This is the part 6 of 13 of a paper evaluating the use of a generative AI to navigate smartphones. You can read the rest of the paper via the table of links below. Table of Links Abstract and 1 Introduction
2 Related Work
3 MM-Navigator
3.1 Problem Formulation and 3.2 Screen Grounding and Navigation via Set of Mark
3.3 History Generation via Multimodal Self Summarization
4 iOS Screen Navigation Experiment
4.1 Experimental Setup
4.2 Intended Action Description
4.3 Localized Action Execution and 4.4 The Current State with GPT-4V
5 Android Screen Navigation Experiment
5.1 Experimental Setup
5.2 Performance Comparison
5.3 Ablation Studies
5.4 Error Analysis
6 Discussion
7 Conclusion and References 4.2 Intended Action Description Table 1 reports an accuracy of 90.9% on generating the correct intended action description, quantitatively supporting GPT-4V’s capability in understanding the screen actions to perform (Yang et al., 2023c; Lin et al., 2023). Figure 1 showcases representative screen understanding examples. Given a screen and a text instruction, GPT-4V gives a text description of its intended next move. For example, in Figure 1(a), GPT-4V understands the Safari browser limits of “the limit of 500 open tabs,” and suggests “Try closing a few tabs and then see if the "+" button becomes clickable.” Another example is telling the procedure for iOS update: “You should click on "General" and then look for an option labeled "Software Update” in (b). GPT-4V also effectively understands complicated screens with multiple images and icons. For example, in (c), GPT-4V mentions, “For information on road closures and other alerts at Mt. Rainier, you should click on "6 Alerts" at the top of the screen.” Figure 1(d) gives an example in online shopping, where GPT-4V suggests the correct product to check based on the user input of the desired “wet cat food.” This paper is available on arxiv under CC BY 4.0 DEED license. Authors: (1) An Yan, UC San Diego, ayan@ucsd.edu; (2) Zhengyuan Yang, Microsoft Corporation, zhengyang@microsoft.com with equal contributions; (3) Wanrong Zhu, UC Santa Barbara, wanrongzhu@ucsb.edu; (4) Kevin Lin, Microsoft Corporation, keli@microsoft.com; (5) Linjie Li, Microsoft Corporation, lindsey.li@mocrosoft.com; (6) Jianfeng Wang, Microsoft Corporation, jianfw@mocrosoft.com; (7) Jianwei Yang, Microsoft Corporation, jianwei.yang@mocrosoft.com; (8) Yiwu Zhong, University of Wisconsin-Madison, yzhong52@wisc.edu; (9) Julian McAuley, UC San Diego, jmcauley@ucsd.edu; (10) Jianfeng Gao, Microsoft Corporation, jfgao@mocrosoft.com; (11) Zicheng Liu, Microsoft Corporation, zliu@mocrosoft.com; (12) Lijuan Wang, Microsoft Corporation, lijuanw@mocrosoft.com. Authors: Authors: (1) An Yan, UC San Diego, ayan@ucsd.edu; (2) Zhengyuan Yang, Microsoft Corporation, zhengyang@microsoft.com with equal contributions; (3) Wanrong Zhu, UC Santa Barbara, wanrongzhu@ucsb.edu; (4) Kevin Lin, Microsoft Corporation, keli@microsoft.com; (5) Linjie Li, Microsoft Corporation, lindsey.li@mocrosoft.com; (6) Jianfeng Wang, Microsoft Corporation, jianfw@mocrosoft.com; (7) Jianwei Yang, Microsoft Corporation, jianwei.yang@mocrosoft.com; (8) Yiwu Zhong, University of Wisconsin-Madison, yzhong52@wisc.edu; (9) Julian McAuley, UC San Diego, jmcauley@ucsd.edu; (10) Jianfeng Gao, Microsoft Corporation, jfgao@mocrosoft.com; (11) Zicheng Liu, Microsoft Corporation, zliu@mocrosoft.com; (12) Lijuan Wang, Microsoft Corporation, lijuanw@mocrosoft.com. Editor’s note: This is the part 6 of 13 of a paper evaluating the use of a generative AI to navigate smartphones. You can read the rest of the paper via the table of links below. Editor’s note: This is the part 6 of 13 of a paper evaluating the use of a generative AI to navigate smartphones. You can read the rest of the paper via the table of links below. Editor’s note: This is the part 6 of 13 of a paper evaluating the use of a generative AI to navigate smartphones. You can read the rest of the paper via the table of links below. Editor’s note: This is the part 6 of 13 of a paper evaluating the use of a generative AI to navigate smartphones. You can read the rest of the paper via the table of links below. Table of Links Abstract and 1 Introduction 2 Related Work 3 MM-Navigator 3.1 Problem Formulation and 3.2 Screen Grounding and Navigation via Set of Mark 3.3 History Generation via Multimodal Self Summarization 4 iOS Screen Navigation Experiment 4.1 Experimental Setup 4.2 Intended Action Description 4.3 Localized Action Execution and 4.4 The Current State with GPT-4V 5 Android Screen Navigation Experiment 5.1 Experimental Setup 5.2 Performance Comparison 5.3 Ablation Studies 5.4 Error Analysis 6 Discussion 7 Conclusion and References Abstract and 1 Introduction Abstract and 1 Introduction 2 Related Work 2 Related Work 3 MM-Navigator 3.1 Problem Formulation and 3.2 Screen Grounding and Navigation via Set of Mark 3.1 Problem Formulation and 3.2 Screen Grounding and Navigation via Set of Mark 3.3 History Generation via Multimodal Self Summarization 3.3 History Generation via Multimodal Self Summarization 4 iOS Screen Navigation Experiment 4.1 Experimental Setup 4.1 Experimental Setup 4.2 Intended Action Description 4.2 Intended Action Description 4.3 Localized Action Execution and 4.4 The Current State with GPT-4V 4.3 Localized Action Execution and 4.4 The Current State with GPT-4V 5 Android Screen Navigation Experiment 5.1 Experimental Setup 5.1 Experimental Setup 5.2 Performance Comparison 5.2 Performance Comparison 5.3 Ablation Studies 5.3 Ablation Studies 5.4 Error Analysis 5.4 Error Analysis 6 Discussion 6 Discussion 7 Conclusion and References 7 Conclusion and References 4.2 Intended Action Description Table 1 reports an accuracy of 90.9% on generating the correct intended action description, quantitatively supporting GPT-4V’s capability in understanding the screen actions to perform (Yang et al., 2023c; Lin et al., 2023). Figure 1 showcases representative screen understanding examples. Given a screen and a text instruction, GPT-4V gives a text description of its intended next move. For example, in Figure 1(a), GPT-4V understands the Safari browser limits of “the limit of 500 open tabs,” and suggests “Try closing a few tabs and then see if the "+" button becomes clickable.” Another example is telling the procedure for iOS update: “You should click on "General" and then look for an option labeled "Software Update” in (b). GPT-4V also effectively understands complicated screens with multiple images and icons. For example, in (c), GPT-4V mentions, “For information on road closures and other alerts at Mt. Rainier, you should click on "6 Alerts" at the top of the screen.” Figure 1(d) gives an example in online shopping, where GPT-4V suggests the correct product to check based on the user input of the desired “wet cat food.” This paper is available on arxiv under CC BY 4.0 DEED license. This paper is available on arxiv under CC BY 4.0 DEED license. available on arxiv

Researchers Successfully Develop AI Model That Can Handle Everyday Tasks on Your iPhone

Albert Einstein famously defined insanity as doing the same thing over and over again and expecting different results. From my perspective, this principle, often cited to highlight counterproductive behavior, aptly describes the current state of B2B marketing, especially at the middle market and enterprise levels. Despite evolving technologies and market dynamics, many companies persist with traditional strategies relying on volume and consistency, hoping for a breakthrough that sometimes never arrives. The Current Challenges In B2B Marketing The B2B marketing landscape is fraught with challenges. From saturated channels to evolving buyer expectations, businesses find it increasingly difficult to stand out. Email marketing, social media, blogs and content creation have become ubiquitous tools. Yet, their effectiveness diminishes as every company follows the same playbook. Let's explore some of the core challenges. Email Marketing Fatigue: Despite sophisticated segmentation and personalization techniques, email inboxes are flooded. Decision makers often overlook even the most well-crafted emails amidst the clutter. • Social Media Overload: Platforms like LinkedIn are brimming with posts, articles and ads. The noise level is high, making it challenging for any single message to resonate. • Content Saturation: Every company has a blog, eBook or whitepaper. The internet is awash with content, much of it repetitive and lacking in originality. Potential clients struggle to find valuable insights amid the deluge. • Diminishing Returns On Traditional Tactics: Webinars, whitepapers and case studies are staples in B2B marketing. However, as more businesses use these tactics, their impact lessens. The same strategies are recycled without yielding new results. The Insanity Principle In Action The prevailing B2B marketing approach exemplifies the Insanity Principle. Companies stick to the same strategies, hoping for different outcomes. The cycle perpetuates because change is daunting, and there is comfort in the familiar. Yet, breaking free from this cycle is essential for meaningful progress. Rethinking B2B Marketing Strategies To overcome these challenges and escape the insanity cycle, B2B marketers must embrace innovation and adaptability. Here are some steps to consider. • Embrace Account-Based Marketing (ABM): ABM focuses on personalized, high-touch engagement with key accounts. This approach contrasts with the broad, often impersonal methods of traditional marketing. By tailoring efforts to specific needs and pain points, companies can foster deeper relationships with customers and drive better results. • Leverage AI And Data Analytics: AI can offer predictive insights, automate mundane tasks and personalize interactions at scale. Data analytics provide a clearer understanding of buyer behavior, enabling more targeted and effective marketing efforts. • Humanize The Brand: Authenticity resonates. User-generated content and influencer partnerships can inject credibility and relatability into marketing efforts. Meeting customers where they are—whether on TikTok, Instagram or LinkedIn—ensures messages are delivered in a context that feels natural and engaging. • Innovate With Interactive Content: Static content is losing its luster. Interactive content like quizzes, calculators and virtual events can capture attention and engage audiences in more meaningful ways. • Foster Community And Engagement: Building a community around your brand can drive loyalty and advocacy. Creating platforms for discussion, feedback and shared experiences can help cultivate a sense of belonging and trust. • Experiment And Iterate: Agile marketing allows for continuous testing and optimization. By embracing a mindset of experimentation, companies can quickly adapt to changing market conditions and refine their strategies based on real-time feedback. Looking Forward The Insanity Principle serves as a cautionary tale for B2B marketers. Sticking to outdated strategies while expecting different outcomes is a recipe for stagnation. By embracing innovation, leveraging new technologies and prioritizing authentic engagement, companies can break free from the cycle and achieve sustainable growth. Truly escaping the trappings of the Insanity Principle requires bold steps and a willingness to challenge the status quo. Here are some ideas. • Adopt A Growth Mindset: Encourage a culture that embraces change and continuous learning. A growth mindset allows teams to view failures as opportunities for learning and improvement. • Empower Teams To Take Risks: Innovation often comes from taking calculated risks. Leadership should create an environment where teams feel safe to experiment without fear of failure. • Invest In Continuous Improvement: Regularly assess and refine marketing strategies based on data-driven insights. Continuous improvement ensures that marketing efforts evolve with market conditions and buyer expectations. • Encourage Cross-Functional Collaboration: Breaking down silos and fostering collaboration across departments can lead to more integrated and effective marketing strategies. Diverse perspectives often lead to more innovative solutions. • Stay Customer-Centric: Keep the customer at the heart of all marketing efforts. Understanding and addressing their evolving needs and preferences can drive more effective and resonant marketing campaigns. By moving beyond the comfort of the familiar and pioneering new paths in B2B marketing, companies can not only avoid the pitfalls of the Insanity Principle but also position themselves for long-term success in a rapidly changing landscape. Follow me on LinkedIn. Albert Einstein famously defined insanity as doing the same thing over and over again and expecting different results. From my perspective, this principle, often cited to highlight counterproductive behavior, aptly describes the current state of B2B marketing, especially at the middle market and enterprise levels. Despite evolving technologies and market dynamics, many companies persist with traditional strategies relying on volume and consistency, hoping for a breakthrough that sometimes never arrives. The Current Challenges In B2B Marketing The B2B marketing landscape is fraught with challenges. From saturated channels to evolving buyer expectations, businesses find it increasingly difficult to stand out. Email marketing, social media, blogs and content creation have become ubiquitous tools. Yet, their effectiveness diminishes as every company follows the same playbook. Let's explore some of the core challenges. Email Marketing Fatigue: Despite sophisticated segmentation and personalization techniques, email inboxes are flooded. Decision makers often overlook even the most well-crafted emails amidst the clutter. Email Marketing Fatigue: Despite sophisticated segmentation and personalization techniques, email inboxes are flooded. Decision makers often overlook even the most well-crafted emails amidst the clutter. Email Marketing Fatigue: • Social Media Overload: Platforms like LinkedIn are brimming with posts, articles and ads. The noise level is high, making it challenging for any single message to resonate. • Social Media Overload: • Content Saturation: Every company has a blog, eBook or whitepaper. The internet is awash with content, much of it repetitive and lacking in originality. Potential clients struggle to find valuable insights amid the deluge. • Content Saturation: • Diminishing Returns On Traditional Tactics: Webinars, whitepapers and case studies are staples in B2B marketing. However, as more businesses use these tactics, their impact lessens. The same strategies are recycled without yielding new results. • Diminishing Returns On Traditional Tactics: The Insanity Principle In Action The prevailing B2B marketing approach exemplifies the Insanity Principle. Companies stick to the same strategies, hoping for different outcomes. The cycle perpetuates because change is daunting, and there is comfort in the familiar. Yet, breaking free from this cycle is essential for meaningful progress. Rethinking B2B Marketing Strategies To overcome these challenges and escape the insanity cycle, B2B marketers must embrace innovation and adaptability. Here are some steps to consider. • Embrace Account-Based Marketing (ABM): ABM focuses on personalized, high-touch engagement with key accounts. This approach contrasts with the broad, often impersonal methods of traditional marketing. By tailoring efforts to specific needs and pain points, companies can foster deeper relationships with customers and drive better results. • Embrace Account-Based Marketing (ABM): • Leverage AI And Data Analytics: AI can offer predictive insights, automate mundane tasks and personalize interactions at scale. Data analytics provide a clearer understanding of buyer behavior, enabling more targeted and effective marketing efforts. • Leverage AI And Data Analytics: • Humanize The Brand: Authenticity resonates. User-generated content and influencer partnerships can inject credibility and relatability into marketing efforts. Meeting customers where they are—whether on TikTok, Instagram or LinkedIn—ensures messages are delivered in a context that feels natural and engaging. • Humanize The Brand: • Innovate With Interactive Content: Static content is losing its luster. Interactive content like quizzes, calculators and virtual events can capture attention and engage audiences in more meaningful ways. • Innovate With Interactive Content: • Foster Community And Engagement: Building a community around your brand can drive loyalty and advocacy. Creating platforms for discussion, feedback and shared experiences can help cultivate a sense of belonging and trust. • Foster Community And Engagement: • Experiment And Iterate: Agile marketing allows for continuous testing and optimization. By embracing a mindset of experimentation, companies can quickly adapt to changing market conditions and refine their strategies based on real-time feedback. • Experiment And Iterate: Looking Forward The Insanity Principle serves as a cautionary tale for B2B marketers. Sticking to outdated strategies while expecting different outcomes is a recipe for stagnation. By embracing innovation, leveraging new technologies and prioritizing authentic engagement, companies can break free from the cycle and achieve sustainable growth. Truly escaping the trappings of the Insanity Principle requires bold steps and a willingness to challenge the status quo. Here are some ideas. • Adopt A Growth Mindset: Encourage a culture that embraces change and continuous learning. A growth mindset allows teams to view failures as opportunities for learning and improvement. • Adopt A Growth Mindset: • Empower Teams To Take Risks: Innovation often comes from taking calculated risks. Leadership should create an environment where teams feel safe to experiment without fear of failure. • Empower Teams To Take Risks: • Invest In Continuous Improvement: Regularly assess and refine marketing strategies based on data-driven insights. Continuous improvement ensures that marketing efforts evolve with market conditions and buyer expectations. • Invest In Continuous Improvement: • Encourage Cross-Functional Collaboration: Breaking down silos and fostering collaboration across departments can lead to more integrated and effective marketing strategies. Diverse perspectives often lead to more innovative solutions. • Encourage Cross-Functional Collaboration: • Stay Customer-Centric: Keep the customer at the heart of all marketing efforts. Understanding and addressing their evolving needs and preferences can drive more effective and resonant marketing campaigns. • Stay Customer-Centric: By moving beyond the comfort of the familiar and pioneering new paths in B2B marketing, companies can not only avoid the pitfalls of the Insanity Principle but also position themselves for long-term success in a rapidly changing landscape. Follow me on LinkedIn . LinkedIn

B2B Marketing Is Stuck in a Loop, and It’s Driving Everyone Insane

B2B Marketing ukax mä Loop ukan ch’amanchatawa, ukatx Taqiniruw Loco ukham tukuyi

El marketing B2B está atrapado en un bucle y está volviendo locos a todos

Maketing B2B kole nan yon bouk, epi li fè tout moun fou

B2B マーケティングはループに陥り、誰もが気が狂いそうになっている

B2B მარკეტინგი გაჭედილია და ის ყველას გიჟდება

B2B 마케팅은 루프에 갇혀 있고 모든 사람을 미치게 만들고 있습니다.

B2B маркетинги айлампага тыгылып, бардыгын жинди кылып жатат

B2B mārketings ir iestrēdzis, un tas visus padara neprātīgus

Marketingul B2B este blocat și îi înnebunește pe toți

B2B අලෙවිකරණය ලූපයක සිරවී ඇති අතර එය සෑම කෙනෙකුම උමතු කරවයි

Б2Б маркетинг је заглављен у петљи и све излуђује

การตลาด B2B ติดอยู่ในวงจร และทำให้ทุกคนคลั่งไคล้

Every Reason You Should be Getting Started With .Net Aspire

"If you have four hours to cut a tree with an axe, spend three sharpening the axe. If you have six hours, find a chainsaw." ~ Praise J.J. The world doesn’t reward effort. It rewards results. Yet, effort can lead to results if applied correctly. The secret? leverage. Let’s start with the question that matters: Are you right enough? Success Isn’t About Dreams—It’s About Systems Unlike the feel-good narrative Hollywood sells, success isn’t about bad childhoods and big aspirations. It’s about having the right systems. Everyone running with Usain Bolt has the same goal: finish first. The winner isn’t the one who wants it most—it’s the one with the right combination of diet, training, and strategy. The world operates efficiently. Unlike school, where you get points for showing your work, the real world only cares about the answer. Effort ≠ Justification. Results = justification. $10,000 in value earns $10,000, whether it takes 20 minutes or 200 hours. Even if it’s 10 lines of code. The question isn’t whether you’re working hard or smart. The question is: Are you doing what it takes? “Work Hard" Is Overrated Let’s break down the world’s advice: 49% say: Work hard. I disagree.
Another 49% say: Work smart. I disagree.
The final 1% say: Work hard AND smart. I still disagree. The problem? These phrases are overloaded—they mean different things to different people. The better word is diligence: doing the work required, with or without recognition, for as long as it takes. Not your best, not your hardest, not your smartest. Just what the problem demands. Work = Volume × Leverage The work equation is simple: Volume is the intensity of inputs: 100 cold emails, 10,000 words, 500 squats.
Leverage is the multiplier: the tools or systems that amplify your efforts. To achieve success, you need the right balance of volume and leverage. Increase volume: "Do more."
Find leverage: tools, systems, or strategies that multiply your inputs. There are two types of leverage: Permissioned: Somebody has to give it to you (e.g., labor, capital).
Permissionless: You don't need consent (e.g., code, content creation). Leverage is non-negotiable. Without it, you’re playing the wrong game. Linear vs. Exponential Growth You either grow or die. The stock market grows by 9% annually. Money is inflating. Everything is dynamic. So, being stagnant means you're growing at the same rate the world moves. Anything less is dying. There are two types of growth paths: Linear Growth
Exponential Growth Linear Growth: Climbing the Ladder Linear growth is when the value of leverage is 1. Most people are taught to grow linearly: Climb the corporate ladder.
Add one rung at a time.
Earn slightly more each year. Linear growth is predictable: Input X → Output X.
Input 2X → Output 2X. The problem? You always have to push. If you stop, you die. Linear growth works for specialists: surgeons, lawyers, and pilots. They’re expensive because they provide immense value. But their value depends entirely on them showing up. No surgeon, no surgery. If your goal is to build something bigger than yourself, linear growth won’t cut it. Exponential Growth: Leveraging Multipliers Exponential growth is where the magic happens. Here the value of leverage skyrockets. You can work for five years and see nothing, then one weekend in a garage changes everything. Examples: Facebook isn’t about the code—it’s the network effects.
Tesla’s worth isn’t in cars—it’s in the gigafactories and IP.
MrBeast’s success isn’t from one viral video—it’s consistency meeting exponential reach. Exponential growth is uncomfortable because it's not easy to predict. In exponential growth, you will always get the big payoff, but you can't know when. If you know when, you've already had the payoff. In linear growth, you will usually be right because you're not doing anything new. 1+1 will always equal 2. Millions of people have already lived their lives and gotten the same results. It's a framework that always delivers. It's based on simple and concrete truths. You'll only be wrong a few times (accidents and unfortunate circumstances). In exponential growth, you will be wrong 99% of the time. Just like Thomas Edison, just like the Wright brothers, just like Elon Musk, just like MrBeast and thousands of others that are building things bigger than themselves.But the only thing that matters is the big one at the end. That's what makes you 100% right. Linear people are 99% right but 100% wrong. The 1% that works overshadows all the failures. Exponential growth rewards patience, persistence, and adaptability. Every failed iteration builds the foundation for eventual success. Start Young, Start Small, Start Now The earlier you start chasing exponential growth, the easier it is. Early exponential growth: Late Exponential Growth: Here’s why: When you’re young, failure costs less.
As responsibilities grow, the cost of risk skyrockets. Consider this: If you earn $500/hour, risking 100 hours feels like losing $50,000.
If you earn $50/hour, those same 100 hours only risk $5,000.
If you earn $0/hour, you have nothing to lose. Start today. Every day you wait makes it harder to escape the linear trap. Leverage: The Skeleton Key to Tech Giants Leverage simply takes a decision that linear people would have to make repeatedly and does it at scale while you sleep. Tech giants leveraged systems that multiplied their efforts. Here’s how: Intellectual Property


Protect unique assets others can’t copy.


Example: Google’s search algorithm, Coca-Cola’s secret formula.


Takeaway: Build moats around your core value.



Network Effects


Users amplify the product’s value.


Example: Uber, LinkedIn, and Airbnb thrive because every new user strengthens the ecosystem.


Takeaway: Create flywheels where users drive growth.



Vertical Integration


Control your supply chain. Keep profits in-house.


Example: Apple builds hardware and software; Shopify owns its e-commerce stack.


Takeaway: Maximize profits by owning the stack.



Government Contracts


Secure recurring revenue and legitimacy.


Example: Palantir thrives on defense contracts.


Takeaway: Look for opportunities in regulated industries.



Productized Simplicity

Nail one thing.
Example: The "Yo" app (just sends “Yo”), yet it dominated for its simplicity.
Takeaway: Complexity isn’t always an asset. The Catapult Principle: Consistency Beats Entropy Success is like a catapult. The more you consistently pull, the farther you shoot. But inconsistency lets the stone fall off. Be diligent. Show up every day, especially when it feels pointless. Closing Thoughts: The Best Time to Start The best time to chase exponential growth was when you were a child. The second-best time is today. Linear growth gives comfort, but exponential growth gives freedom. Leverage your unique skills. Start small. Stay consistent. And one day, people will call you an overnight success. Rooting for you. Enjoy the rest of your day. PS: Want to unlock exponential growth and mental clarity? Join my newsletter: https://crive.substack.com "If you have four hours to cut a tree with an axe, spend three sharpening the axe. If you have six hours, find a chainsaw." ~ Praise J.J. "If you have four hours to cut a tree with an axe, spend three sharpening the axe. If you have six hours, find a chainsaw." ~ Praise J.J. "If you have four hours to cut a tree with an axe, spend three sharpening the axe. If you have six hours, find a chainsaw." The world doesn’t reward effort. It rewards results. Yet, effort can lead to results if applied correctly. The secret? leverage. Let’s start with the question that matters: Are you right enough? Are you right enough? Success Isn’t About Dreams—It’s About Systems Success Isn’t About Dreams—It’s About Systems Unlike the feel-good narrative Hollywood sells, success isn’t about bad childhoods and big aspirations. It’s about having the right systems. Everyone running with Usain Bolt has the same goal: finish first . The winner isn’t the one who wants it most—it’s the one with the right combination of diet, training, and strategy. finish first The world operates efficiently. Unlike school, where you get points for showing your work, the real world only cares about the answer. Effort ≠ Justification. Results = justification. $10,000 in value earns $10,000, whether it takes 20 minutes or 200 hours. Even if it’s 10 lines of code. The question isn’t whether you’re working hard or smart. The question is: Are you doing what it takes? Are you doing what it takes? “Work Hard" Is Overrated “Work Hard" Is Overrated Let’s break down the world’s advice: 49% say: Work hard. I disagree. Another 49% say: Work smart. I disagree. The final 1% say: Work hard AND smart. I still disagree. 49% say: Work hard. I disagree. Work hard. Another 49% say: Work smart. I disagree. Work smart. The final 1% say: Work hard AND smart. I still disagree. Work hard AND smart. The problem? These phrases are overloaded—they mean different things to different people. The better word is diligence : doing the work required, with or without recognition, for as long as it takes. Not your best, not your hardest, not your smartest. Just what the problem demands. diligence Work = Volume × Leverage Work = Volume × Leverage The work equation is simple: Volume is the intensity of inputs: 100 cold emails, 10,000 words, 500 squats. Leverage is the multiplier: the tools or systems that amplify your efforts. Volume is the intensity of inputs: 100 cold emails, 10,000 words, 500 squats. Volume Leverage is the multiplier: the tools or systems that amplify your efforts. Leverage To achieve success, you need the right balance of volume and leverage. Increase volume: "Do more." Find leverage: tools, systems, or strategies that multiply your inputs. Increase volume : "Do more." volume Find leverage : tools, systems, or strategies that multiply your inputs. leverage There are two types of leverage: Permissioned: Somebody has to give it to you (e.g., labor, capital). Permissionless: You don't need consent (e.g., code, content creation). Permissioned: Somebody has to give it to you (e.g., labor, capital). Permissioned: Permissionless: You don't need consent (e.g., code, content creation). Permissionless: Leverage is non-negotiable. Without it, you’re playing the wrong game. Linear vs. Exponential Growth Linear vs. Exponential Growth You either grow or die. The stock market grows by 9% annually. Money is inflating. Everything is dynamic. So, being stagnant means you're growing at the same rate the world moves. Anything less is dying. There are two types of growth paths: Linear Growth Exponential Growth Linear Growth Linear Growth Exponential Growth Exponential Growth Linear Growth: Climbing the Ladder Linear Growth: Climbing the Ladder Linear growth is when the value of leverage is 1. Most people are taught to grow linearly: Climb the corporate ladder. Add one rung at a time. Earn slightly more each year. Climb the corporate ladder. Add one rung at a time. Earn slightly more each year. Linear growth is predictable: Input X → Output X. Input 2X → Output 2X. Input X → Output X. Input 2X → Output 2X. The problem? You always have to push. If you stop, you die. Linear growth works for specialists: surgeons, lawyers, and pilots. They’re expensive because they provide immense value. But their value depends entirely on them showing up. No surgeon, no surgery. If your goal is to build something bigger than yourself, linear growth won’t cut it. Exponential Growth: Leveraging Multipliers Exponential Growth: Leveraging Multipliers Exponential growth is where the magic happens. Here the value of leverage skyrockets. You can work for five years and see nothing, then one weekend in a garage changes everything. Examples: Facebook isn’t about the code—it’s the network effects. Tesla’s worth isn’t in cars—it’s in the gigafactories and IP. MrBeast’s success isn’t from one viral video—it’s consistency meeting exponential reach. Facebook isn’t about the code—it’s the network effects. Tesla’s worth isn’t in cars—it’s in the gigafactories and IP. MrBeast’s success isn’t from one viral video—it’s consistency meeting exponential reach. Exponential growth is uncomfortable because it's not easy to predict. In exponential growth, you will always get the big payoff, but you can't know when. If you know when, you've already had the payoff. In linear growth, you will usually be right because you're not doing anything new. 1+1 will always equal 2. Millions of people have already lived their lives and gotten the same results. It's a framework that always delivers. It's based on simple and concrete truths. You'll only be wrong a few times (accidents and unfortunate circumstances). In exponential growth, you will be wrong 99% of the time. Just like Thomas Edison, just like the Wright brothers, just like Elon Musk, just like MrBeast and thousands of others that are building things bigger than themselves. But the only thing that matters is the big one at the end. That's what makes you 100% right. Linear people are 99% right but 100% wrong. The 1% that works overshadows all the failures. Exponential growth rewards patience, persistence, and adaptability. Every failed iteration builds the foundation for eventual success. Start Young, Start Small, Start Now Start Young, Start Small, Start Now The earlier you start chasing exponential growth, the easier it is. Early exponential growth: Early exponential growth: Late Exponential Growth: Late Exponential Growth: Here’s why: When you’re young, failure costs less. As responsibilities grow, the cost of risk skyrockets. When you’re young, failure costs less. As responsibilities grow, the cost of risk skyrockets. Consider this: If you earn $500/hour, risking 100 hours feels like losing $50,000. If you earn $50/hour, those same 100 hours only risk $5,000. If you earn $0/hour, you have nothing to lose. If you earn $500/hour, risking 100 hours feels like losing $50,000. If you earn $50/hour, those same 100 hours only risk $5,000. If you earn $0/hour, you have nothing to lose. Start today. Every day you wait makes it harder to escape the linear trap. Leverage: The Skeleton Key to Tech Giants Leverage: The Skeleton Key to Tech Giants Leverage simply takes a decision that linear people would have to make repeatedly and does it at scale while you sleep. Tech giants leveraged systems that multiplied their efforts. Here’s how: Intellectual Property


Protect unique assets others can’t copy.


Example: Google’s search algorithm, Coca-Cola’s secret formula.


Takeaway: Build moats around your core value. Network Effects


Users amplify the product’s value.


Example: Uber, LinkedIn, and Airbnb thrive because every new user strengthens the ecosystem.


Takeaway: Create flywheels where users drive growth. Vertical Integration


Control your supply chain. Keep profits in-house.


Example: Apple builds hardware and software; Shopify owns its e-commerce stack.


Takeaway: Maximize profits by owning the stack. Government Contracts


Secure recurring revenue and legitimacy.


Example: Palantir thrives on defense contracts.


Takeaway: Look for opportunities in regulated industries. Productized Simplicity

Nail one thing.
Example: The "Yo" app (just sends “Yo”), yet it dominated for its simplicity.
Takeaway: Complexity isn’t always an asset. Intellectual Property Protect unique assets others can’t copy.


Example: Google’s search algorithm, Coca-Cola’s secret formula.


Takeaway: Build moats around your core value. Intellectual Property Protect unique assets others can’t copy. Example: Google’s search algorithm, Coca-Cola’s secret formula. Takeaway: Build moats around your core value. Protect unique assets others can’t copy. Protect unique assets others can’t copy. Example: Google’s search algorithm, Coca-Cola’s secret formula. Example: Google’s search algorithm, Coca-Cola’s secret formula. Takeaway: Build moats around your core value. Takeaway: Build moats around your core value. Takeaway: Network Effects Users amplify the product’s value.


Example: Uber, LinkedIn, and Airbnb thrive because every new user strengthens the ecosystem.


Takeaway: Create flywheels where users drive growth. Network Effects Users amplify the product’s value. Example: Uber, LinkedIn, and Airbnb thrive because every new user strengthens the ecosystem. Takeaway: Create flywheels where users drive growth. Users amplify the product’s value. Users amplify the product’s value. Example: Uber, LinkedIn, and Airbnb thrive because every new user strengthens the ecosystem. Example: Uber, LinkedIn, and Airbnb thrive because every new user strengthens the ecosystem. Takeaway: Create flywheels where users drive growth. Takeaway: Create flywheels where users drive growth. Takeaway: Vertical Integration Control your supply chain. Keep profits in-house.


Example: Apple builds hardware and software; Shopify owns its e-commerce stack.


Takeaway: Maximize profits by owning the stack. Vertical Integration Control your supply chain. Keep profits in-house. Example: Apple builds hardware and software; Shopify owns its e-commerce stack. Takeaway: Maximize profits by owning the stack. Control your supply chain. Keep profits in-house. Control your supply chain. Keep profits in-house. Example: Apple builds hardware and software; Shopify owns its e-commerce stack. Example: Apple builds hardware and software; Shopify owns its e-commerce stack. Takeaway: Maximize profits by owning the stack. Takeaway: Maximize profits by owning the stack. Takeaway: Government Contracts Secure recurring revenue and legitimacy.


Example: Palantir thrives on defense contracts.


Takeaway: Look for opportunities in regulated industries. Government Contracts Secure recurring revenue and legitimacy. Example: Palantir thrives on defense contracts. Takeaway: Look for opportunities in regulated industries. Secure recurring revenue and legitimacy. Secure recurring revenue and legitimacy. Example: Palantir thrives on defense contracts. Example: Palantir thrives on defense contracts. Takeaway: Look for opportunities in regulated industries. Takeaway: Look for opportunities in regulated industries. Takeaway: Productized Simplicity Nail one thing.
Example: The "Yo" app (just sends “Yo”), yet it dominated for its simplicity.
Takeaway: Complexity isn’t always an asset. Productized Simplicity Nail one thing. Example: The "Yo" app (just sends “Yo”), yet it dominated for its simplicity. Takeaway: Complexity isn’t always an asset. Nail one thing. Example: The "Yo" app (just sends “Yo”), yet it dominated for its simplicity. Takeaway: Complexity isn’t always an asset. Takeaway: The Catapult Principle: Consistency Beats Entropy The Catapult Principle: Consistency Beats Entropy Success is like a catapult. The more you consistently pull, the farther you shoot. But inconsistency lets the stone fall off. Be diligent. Show up every day, especially when it feels pointless. Closing Thoughts: The Best Time to Start Closing Thoughts: The Best Time to Start The best time to chase exponential growth was when you were a child. The second-best time is today. Linear growth gives comfort, but exponential growth gives freedom. Leverage your unique skills. Start small. Stay consistent. And one day, people will call you an overnight success. Rooting for you. Enjoy the rest of your day. PS: Want to unlock exponential growth and mental clarity? Join my newsletter: https://crive.substack.com PS: Want to unlock exponential growth and mental clarity?

Leverage: The Shortcut to Outperforming Everyone

Today, we are excited to announce that the modular compute layer & RaaS platform Lumoz will launch its mainnet and conduct the TGE within a week. A month ago, Lumoz opened esMOZ token airdrop queries and launched the Lumoz OG NFT campaign. To date, over 3 million users have participated in the airdrop queries, and more than 300,000 NFTs have been claimed. The scale of these activities and the large number of participants further demonstrate the community's immense enthusiasm and support for Lumoz's future development. In this article, we will provide everything you need to know about Lumoz and the essential details about the unmissable TGE! Lumoz Development Journey
TGE Timeline
MOZ Tokenomics
Potential Participation Opportunities
Conversion Between esMOZ and MOZ
Conclusion Lumoz Development Journey Lumoz is a leading modular compute layer and ZK-RaaS platform, with a total funding of $14 million. Participating investors include OKX Ventures, HashKey Capital, IDG, Web3port, and Polygon. In terms of market performance, since the launch of its testnet in May 2023, Lumoz has successfully completed three testnet phases: Pre-Alpha, Alpha, and Quidditch. The platform has attracted over 3 million users and gained support from more than 100 top-tier ecosystem projects, growing Lumoz's community to 3 million members. On the technological front, Lumoz remains at the forefront of the industry. Leveraging its "one-click Layer 2 deployment" technology, Lumoz has provided technical services and support to over 20 high-quality projects and Layer 2 platforms, including CARV, UXLINK, ZKFair, Merlin Chain, Matr1X, and Ultiverse. TGE Timeline With the launch and conclusion of the large-scale airdrop and OG NFT campaigns, the Lumoz mainnet launch and TGE are fast approaching. We will also announce the listing timeline soon, so please stay tuned! MOZ Tokenomics On December 5, Lumoz announced the tokenomics for MOZ. The total supply of MOZ tokens is 10 billion, with 66% allocated to the community, ecosystem, nodes, and miners. The initial circulating supply is approximately 11%. The token distribution is as follows: 25% to the zkProver network, 25% to zkVerifier nodes, 16% to early contributors, 18% to early investors, 10% to the ecosystem, and 6% to the community. MOZ tokens will be gradually unlocked over the next 10 years, after which the entire supply will be in circulation, with no additional issuance. More info：https://x.com/LumozOrg/status/1864688725273530507 Potential Participation Opportunities esMOZ Airdrop Query On November 5, Lumoz officially announced the esMOZ airdrop event. The esMOZ airdrop accounts for 10% of the total supply, valued at over $30 million based on a $300 million valuation. The airdrop covers Lumoz node holders, early Lumoz participants, users in the Ethereum and Move ecosystems, including Layer 2 solutions such as Aptos and SUI, PoW testnet participants, as well as users from the Lumoz ecosystem chains like Merlin Chain, ZKFair, Ultiverse, and Matr1x. Additionally, Lumoz conducted airdrops for high-quality project partners and community users from CARV, UXLINK, and other prominent projects. Airdrop query portal: https://lumoz.org/airdrop Lumoz OG NFT Events Since its launch, the Lumoz OG NFT Events has consistently dominated the rankings on major NFT marketplaces, with the current floor price at 15 USDT and Claw units reaching as high as 150 USDT at one point. What makes Lumoz OG NFTs special is that they allow users to unlock and convert esMOZ to MOZ at a 1:1 ratio, providing immediate value returns. Currently, the official platform supports multiple channels for users to obtain NFTs for free. Lumoz OG NFT acquisition portal: https://lumoz.org/og-nft Lumoz zkVerifier Node Public Sale The zkVerifier node, as an essential component of the Lumoz network's verification layer, greatly facilitates ordinary users' participation in the Zero-Knowledge (ZK) computation network while also bearing the critical responsibility of maintaining the security of network data. Lumoz launched node sales six months ago, and node holders are eligible for various rewards, including 25% of Lumoz Tokens, potential incentives from over 20 ecosystem L2s, esMOZ airdrops, Lumoz OG NFTs, and more. Currently, 80% of the nodes have been sold, and the public sales and KOL node incentive programs are still open. If you're interested, you can click here: https://node.lumoz.org/ Conversion Between esMOZ and MOZ Lumoz employs a dual-token economic model, consisting of the functional token MOZ and the incentive token esMOZ. MOZ is primarily used for transaction gas fees, resource usage, and similar purposes, while esMOZ can be used for staking rewards, delegating zkVerifier nodes, and converting into MOZ. Both tokens are integral to the Lumoz protocol. After the mainnet launch, esMOZ can be converted into MOZ under specific conditions. The conversion ratio is 1:1, and conversions typically require a vesting period. However, users can bypass the lock-up period by using Lumoz OG NFTs for conversion. Upon completion of the conversion, one token is burned, and its corresponding token is minted. If a user converts esMOZ using an OG NFT, the OG NFT will also be destroyed after the conversion is completed. For more information on obtaining Lumoz OG NFTs, visit: https://lumoz.org/airdrop. For users who do not use OG NFTs, the following standard conversion rules apply: Shortest conversion period: 30 days, with a conversion ratio of 1:0.25 (25%). The remaining 75% of esMOZ will be burned during conversion.
Medium conversion period: 90 days, with a conversion ratio of 1:0.5 (50%). The remaining 50% of esMOZ will be burned during conversion.
Longest conversion period: 180 days, with a conversion ratio of 1:1 (100%). Conclusion What is past is prologue. Currently, Lumoz is in the final stages of preparation, and we hope everyone can seize this last opportunity to become a part of the Lumoz ecosystem. This is not just an investment in the future, but also the best choice to participate in the revolution of cryptocurrency technology. For those familiar with Lumoz, you’ll know that we upgraded from the Opside brand to Lumoz. The name Lumoz comes from the Latin word "Lumos," meaning "light" or "brightness." It is referenced in J.K. Rowling's bestselling Harry Potter series, where "Lumos" is a spell used to light the tip of a magic wand and create light. By replacing the "s" with "z," "Lumoz" cleverly associates itself with "zk" (Zero-Knowledge Proof) technology. With the mainnet launch and TGE approaching, we sincerely hope Lumoz will become a beacon of light for the ZK track, illuminating the path for the widespread adoption of ZK-Rollups. Now, let us witness this miracle together! Today, we are excited to announce that the modular compute layer & RaaS platform Lumoz will launch its mainnet and conduct the TGE within a week. A month ago, Lumoz opened esMOZ token airdrop queries and launched the Lumoz OG NFT campaign. To date, over 3 million users have participated in the airdrop queries, and more than 300,000 NFTs have been claimed. The scale of these activities and the large number of participants further demonstrate the community's immense enthusiasm and support for Lumoz's future development. In this article, we will provide everything you need to know about Lumoz and the essential details about the unmissable TGE! Lumoz Development Journey TGE Timeline MOZ Tokenomics Potential Participation Opportunities Conversion Between esMOZ and MOZ Conclusion Lumoz Development Journey Lumoz Development Journey Lumoz Development Journey TGE Timeline TGE Timeline TGE Timeline MOZ Tokenomics MOZ Tokenomics MOZ Tokenomics Potential Participation Opportunities Potential Participation Opportunities Potential Participation Opportunities Conversion Between esMOZ and MOZ Conversion Between esMOZ and MOZ Conversion Between esMOZ and MOZ Conclusion Conclusion Conclusion Lumoz Development Journey Lumoz Development Journey Lumoz is a leading modular compute layer and ZK-RaaS platform, with a total funding of $14 million. Participating investors include OKX Ventures, HashKey Capital, IDG, Web3port, and Polygon. In terms of market performance, since the launch of its testnet in May 2023, Lumoz has successfully completed three testnet phases: Pre-Alpha, Alpha, and Quidditch. The platform has attracted over 3 million users and gained support from more than 100 top-tier ecosystem projects, growing Lumoz's community to 3 million members. On the technological front, Lumoz remains at the forefront of the industry. Leveraging its "one-click Layer 2 deployment" technology, Lumoz has provided technical services and support to over 20 high-quality projects and Layer 2 platforms, including CARV, UXLINK, ZKFair, Merlin Chain, Matr1X, and Ultiverse. TGE Timeline With the launch and conclusion of the large-scale airdrop and OG NFT campaigns, the Lumoz mainnet launch and TGE are fast approaching. We will also announce the listing timeline soon, so please stay tuned! MOZ Tokenomics On December 5, Lumoz announced the tokenomics for MOZ. The total supply of MOZ tokens is 10 billion, with 66% allocated to the community, ecosystem, nodes, and miners. The initial circulating supply is approximately 11%. The token distribution is as follows: 25% to the zkProver network, 25% to zkVerifier nodes, 16% to early contributors, 18% to early investors, 10% to the ecosystem, and 6% to the community. MOZ tokens will be gradually unlocked over the next 10 years, after which the entire supply will be in circulation, with no additional issuance. More info：https://x.com/LumozOrg/status/1864688725273530507 Potential Participation Opportunities esMOZ Airdrop Query esMOZ Airdrop Query On November 5, Lumoz officially announced the esMOZ airdrop event. The esMOZ airdrop accounts for 10% of the total supply, valued at over $30 million based on a $300 million valuation. The airdrop covers Lumoz node holders, early Lumoz participants, users in the Ethereum and Move ecosystems, including Layer 2 solutions such as Aptos and SUI, PoW testnet participants, as well as users from the Lumoz ecosystem chains like Merlin Chain, ZKFair, Ultiverse, and Matr1x. Additionally, Lumoz conducted airdrops for high-quality project partners and community users from CARV, UXLINK, and other prominent projects. Airdrop query portal: https://lumoz.org/airdrop https://lumoz.org/airdrop Lumoz OG NFT Events Lumoz OG NFT Events Lumoz OG NFT Events Since its launch, the Lumoz OG NFT Events has consistently dominated the rankings on major NFT marketplaces, with the current floor price at 15 USDT and Claw units reaching as high as 150 USDT at one point. What makes Lumoz OG NFTs special is that they allow users to unlock and convert esMOZ to MOZ at a 1:1 ratio, providing immediate value returns. Currently, the official platform supports multiple channels for users to obtain NFTs for free. Lumoz OG NFT acquisition portal: https://lumoz.org/og-nft https://lumoz.org/og-nft Lumoz zkVerifier Node Public Sale Lumoz zkVerifier Node Public Sale Lumoz zkVerifier Node Public Sale The zkVerifier node, as an essential component of the Lumoz network's verification layer, greatly facilitates ordinary users' participation in the Zero-Knowledge (ZK) computation network while also bearing the critical responsibility of maintaining the security of network data. Lumoz launched node sales six months ago, and node holders are eligible for various rewards, including 25% of Lumoz Tokens, potential incentives from over 20 ecosystem L2s, esMOZ airdrops, Lumoz OG NFTs, and more. Currently, 80% of the nodes have been sold, and the public sales and KOL node incentive programs are still open. If you're interested, you can click here: https://node.lumoz.org/ https://node.lumoz.org/ Conversion Between esMOZ and MOZ Lumoz employs a dual-token economic model, consisting of the functional token MOZ and the incentive token esMOZ. MOZ is primarily used for transaction gas fees, resource usage, and similar purposes, while esMOZ can be used for staking rewards, delegating zkVerifier nodes, and converting into MOZ. Both tokens are integral to the Lumoz protocol. After the mainnet launch, esMOZ can be converted into MOZ under specific conditions. The conversion ratio is 1:1, and conversions typically require a vesting period. However, users can bypass the lock-up period by using Lumoz OG NFTs for conversion. Upon completion of the conversion, one token is burned, and its corresponding token is minted. If a user converts esMOZ using an OG NFT, the OG NFT will also be destroyed after the conversion is completed. For more information on obtaining Lumoz OG NFTs, visit: https://lumoz.org/airdrop . https://lumoz.org/airdrop For users who do not use OG NFTs, the following standard conversion rules apply: Shortest conversion period: 30 days, with a conversion ratio of 1:0.25 (25%). The remaining 75% of esMOZ will be burned during conversion. Medium conversion period: 90 days, with a conversion ratio of 1:0.5 (50%). The remaining 50% of esMOZ will be burned during conversion. Longest conversion period: 180 days, with a conversion ratio of 1:1 (100%). Shortest conversion period: 30 days, with a conversion ratio of 1:0.25 (25%). The remaining 75% of esMOZ will be burned during conversion. Medium conversion period: 90 days, with a conversion ratio of 1:0.5 (50%). The remaining 50% of esMOZ will be burned during conversion. Longest conversion period: 180 days, with a conversion ratio of 1:1 (100%). Conclusion Conclusion What is past is prologue. Currently, Lumoz is in the final stages of preparation, and we hope everyone can seize this last opportunity to become a part of the Lumoz ecosystem. This is not just an investment in the future, but also the best choice to participate in the revolution of cryptocurrency technology. For those familiar with Lumoz, you’ll know that we upgraded from the Opside brand to Lumoz. The name Lumoz comes from the Latin word "Lumos," meaning "light" or "brightness." It is referenced in J.K. Rowling's bestselling Harry Potter series, where "Lumos" is a spell used to light the tip of a magic wand and create light. By replacing the "s" with "z," "Lumoz" cleverly associates itself with "zk" (Zero-Knowledge Proof) technology. With the mainnet launch and TGE approaching, we sincerely hope Lumoz will become a beacon of light for the ZK track, illuminating the path for the widespread adoption of ZK-Rollups. Harry Potter Now, let us witness this miracle together!

Everything You Need to Know About the Lumoz Token Generation Event

Taqi kunatix Lumoz Token Generation Event ukan yatiñax wakiski ukanak yatiñataki

Todo lo que necesitas saber sobre el evento de generación de tokens Lumoz

Tout ce que vous devez savoir sur l'événement de génération de jetons Lumoz

כל מה שאתה צריך לדעת על אירוע דור האסימונים של לומוז

लुमोज़ टोकन जेनरेशन इवेंट के बारे में आपको जो कुछ भी जानना चाहिए

Lumoz トークン生成イベントについて知っておくべきことすべて

ყველაფერი რაც თქვენ უნდა იცოდეთ Lumoz Token Generation ღონისძიების შესახებ

អ្វីគ្រប់យ៉ាងដែលអ្នកត្រូវដឹងអំពីព្រឹត្តិការណ៍ Lumoz Token Generation

Všetko, čo potrebujete vedieť o udalosti generovania tokenov Lumoz

Wax kasta oo aad u baahan tahay inaad ka ogaato Dhacdada Jiilka Lumoz Token

Allt du behöver veta om Lumoz Token Generation Event

Lahat ng Kailangan Mong Malaman Tungkol sa Lumoz Token Generation Event

Using Cursor Can Feel Like Magic—Until It Doesn’t AI-generated code can feel revolutionary—until it fails you. Why does this happen? Often, the issue isn't with the AI itself, but with the inputs we provide. AI thrives on precise and comprehensive instructions. If your prompts are unclear or lack essential context, the fault lies not with the AI but with the user. The Developer’s Dilemma: Context Is Everything Imagine this scenario: you submit a prompt, and Cursor, the AI code editor, delivers what looks like a flawless solution. But then you run the code, and nothing works. Why? Because Cursor lacked the necessary context to understand your unique environment. Here’s what typically goes wrong: Mismatch in Operating Systems: You’re on Linux, but Cursor assumes macOS.
Conflicting Ports or Docker Errors: Cursor generates code oblivious to the state of your running containers.
Python Environment Issues: Your specific dependencies or versions don’t align with the generated code. These are not Cursor’s failures—they’re gaps in the context we provide. Without environmental awareness, even the smartest AI can only guess at solutions. The Solution: Introducing CursorBoost To bridge this gap, I created CursorBoost—a lightweight agent designed to ensure AI has the context it needs to work effectively. It captures critical details about your system in real-time and integrates them seamlessly into your AI interactions. Here’s how CursorBoost transforms your experience: Environment Monitoring: CursorBoost actively tracks vital system details, including:

Operating System
Open Ports
Docker Logs
Python Versions


Dynamic Updates: The agent writes this information into a .cursorrules file—a centralized snapshot of your system’s state.
Enhanced Prompts: By integrating this context, CursorBoost ensures that the AI generates solutions tailored to your specific environment. Without CursorBoost, the AI might suggest using port 8080—only to clash with an active process. With CursorBoost, it knows which ports are open and offers an alternative, saving you from debugging frustration. Why CursorBoost Matters CursorBoost takes the guesswork out of AI-assisted development, delivering practical benefits: Accurate Solutions: Code that aligns with your system from the start.
Time Saved: Fewer iterations and clarifications.
Reduced Frustration: Debugging becomes smoother when the AI understands your setup. This isn’t just about convenience—it’s about making AI truly work for you. Real-Life Success: CursorBoost at the Nvidia & Vercel Hackathon CursorBoost was born out of necessity during an Nvidia and Vercel hackathon. The problem was personal—AI-generated solutions that didn’t align with my workflow. By solving it, I not only saved hours of development time but also won an RTX 4080 GPU signed by Jensen Huang. However, the real prize wasn’t the hardware—it was seeing how much more effective AI could be with the right context. CursorBoost turned my daily frustrations into a scalable solution for developers everywhere. The Golden Rule of Prompt Engineering: Context Is Key The quality of AI output is directly proportional to the clarity of your input. CursorBoost ensures that your AI assistant always has the context it needs to succeed. With better tools, we can spend less time troubleshooting and more time building. Let’s Build Smarter Tools—Together The future of coding lies in collaboration between developers and AI. Tools like CursorBoost bridge the gap, making AI more effective and workflows more efficient. If you’re passionate about improving developer tools or workflows, let’s connect. Together, we can make AI an even more powerful ally in coding. Check out CursorBoost on GitHub: https://github.com/grp06/cursor-boost Using Cursor Can Feel Like Magic—Until It Doesn’t AI-generated code can feel revolutionary—until it fails you. Why does this happen? Often, the issue isn't with the AI itself, but with the inputs we provide. AI thrives on precise and comprehensive instructions. If your prompts are unclear or lack essential context, the fault lies not with the AI but with the user. The Developer’s Dilemma: Context Is Everything Imagine this scenario: you submit a prompt, and Cursor, the AI code editor, delivers what looks like a flawless solution. But then you run the code, and nothing works. Why? Because Cursor lacked the necessary context to understand your unique environment. Here’s what typically goes wrong: Mismatch in Operating Systems: You’re on Linux, but Cursor assumes macOS. Conflicting Ports or Docker Errors: Cursor generates code oblivious to the state of your running containers. Python Environment Issues: Your specific dependencies or versions don’t align with the generated code. Mismatch in Operating Systems : You’re on Linux, but Cursor assumes macOS. Mismatch in Operating Systems Conflicting Ports or Docker Errors : Cursor generates code oblivious to the state of your running containers. Conflicting Ports or Docker Errors Python Environment Issues : Your specific dependencies or versions don’t align with the generated code. Python Environment Issues These are not Cursor’s failures—they’re gaps in the context we provide. Without environmental awareness, even the smartest AI can only guess at solutions. The Solution: Introducing CursorBoost To bridge this gap, I created CursorBoost —a lightweight agent designed to ensure AI has the context it needs to work effectively. It captures critical details about your system in real-time and integrates them seamlessly into your AI interactions. CursorBoost Here’s how CursorBoost transforms your experience: Environment Monitoring: CursorBoost actively tracks vital system details, including:

Operating System
Open Ports
Docker Logs
Python Versions Dynamic Updates: The agent writes this information into a .cursorrules file—a centralized snapshot of your system’s state. Enhanced Prompts: By integrating this context, CursorBoost ensures that the AI generates solutions tailored to your specific environment. Environment Monitoring : CursorBoost actively tracks vital system details, including: Operating System
Open Ports
Docker Logs
Python Versions Environment Monitoring Operating System Open Ports Docker Logs Python Versions Operating System Open Ports Docker Logs Python Versions Dynamic Updates : The agent writes this information into a .cursorrules file—a centralized snapshot of your system’s state. Dynamic Updates .cursorrules Enhanced Prompts : By integrating this context, CursorBoost ensures that the AI generates solutions tailored to your specific environment. Enhanced Prompts Without CursorBoost, the AI might suggest using port 8080—only to clash with an active process. With CursorBoost, it knows which ports are open and offers an alternative, saving you from debugging frustration. Why CursorBoost Matters CursorBoost takes the guesswork out of AI-assisted development, delivering practical benefits: Accurate Solutions: Code that aligns with your system from the start. Time Saved: Fewer iterations and clarifications. Reduced Frustration: Debugging becomes smoother when the AI understands your setup. Accurate Solutions : Code that aligns with your system from the start. Accurate Solutions Time Saved : Fewer iterations and clarifications. Time Saved Reduced Frustration : Debugging becomes smoother when the AI understands your setup. Reduced Frustration This isn’t just about convenience—it’s about making AI truly work for you. Real-Life Success: CursorBoost at the Nvidia & Vercel Hackathon CursorBoost was born out of necessity during an Nvidia and Vercel hackathon . The problem was personal—AI-generated solutions that didn’t align with my workflow. By solving it, I not only saved hours of development time but also won an RTX 4080 GPU signed by Jensen Huang . Nvidia and Vercel hackathon RTX 4080 GPU signed by Jensen Huang However, the real prize wasn’t the hardware—it was seeing how much more effective AI could be with the right context. CursorBoost turned my daily frustrations into a scalable solution for developers everywhere. The Golden Rule of Prompt Engineering: Context Is Key The quality of AI output is directly proportional to the clarity of your input. CursorBoost ensures that your AI assistant always has the context it needs to succeed. With better tools, we can spend less time troubleshooting and more time building. Let’s Build Smarter Tools—Together The future of coding lies in collaboration between developers and AI. Tools like CursorBoost bridge the gap, making AI more effective and workflows more efficient. If you’re passionate about improving developer tools or workflows, let’s connect. Together, we can make AI an even more powerful ally in coding. Check out CursorBoost on GitHub: https://github.com/grp06/cursor-boost Check out CursorBoost on GitHub: https://github.com/grp06/cursor-boost https://github.com/grp06/cursor-boost

Bad Prompts, Bad Results: Why Every LLM Failure Is Your Fault

Дрэнныя падказкі, дрэнныя вынікі: чаму кожны збой LLM - ваша віна

Лоши подкани, лоши резултати: Защо всеки провал на LLM е по ваша вина

Loši upiti, loši rezultati: Zašto je svaki neuspjeh LLM vaša greška

Malos avisos, malos resultados: por qué cada fracaso en un LLM es culpa tuya

درخواست‌های بد، نتایج بد: چرا هر شکست LLM تقصیر شماست

Վատ հուշումներ, վատ արդյունքներ. ինչու է LLM-ի յուրաքանչյուր ձախողում ձեր մեղքը

הנחיות רעות, תוצאות רעות: מדוע כל כשל ב-LLM הוא באשמתך

悪いプロンプト、悪い結果: すべての LLM 不合格はあなたのせいである理由

Жаман сунуштар, жаман натыйжалар: Эмне үчүн ар бир LLM ийгиликсиздиги сиздин күнөөңүз

Maoni Mbaya, Matokeo Mabaya: Kwa Nini Kila Kushindwa kwa LLM ni Kosa Lako

Switsundzuxo Swo Biha, Vuyelo Byo Biha: Ha Yini Ku Hluleka Hinkwako Ka LLM Ku Ri Xihoxo Xa Wena

糟糕的提示，糟糕的结果：为什么每次法学硕士失败都是你的错

Behind Every Question-Answer AI Is a Data Pipeline Built for Scale — Here's How to Build Your Own

Table of Links Abstract and 1 Introduction 2 Background & Problem Statement 2.1 How can we use MLLMs for Diffusion Synthesis that Synergizes both sides? 3 DreamLLM 3.1 End-to-End Interleaved generative Pretraining (I-GPT) 3.2 Model Training 4 Experiments and 4.1 Multimodal Comprehension 4.2 Text-Conditional Image Synthesis 4.3 Multimodal Joint Creation & Comprehension 5 Discussions 5.1 Synergy between creation & Comprehension? 5. 2 What is learned by DreamLLM? 6 Related Works 7 Conclusions and References A Additional Experiments B Additional Qualitative Examples C Implementation Details D Additional Related Works E Limitations, Failure Cases & Future Works ABSTRACT This paper presents DREAMLLM, a learning framework that first achieves versatile Multimodal Large Language Models (MLLMs) empowered with frequently overlooked synergy between multimodal comprehension and creation. DREAMLLM operates on two fundamental principles. The first focuses on the generative modeling of both language and image posteriors by direct sampling in the raw multimodal space. This approach circumvents the limitations and information loss inherent to external feature extractors like CLIP, and a more thorough multimodal understanding is obtained. Second, DREAMLLM fosters the generation of raw, interleaved documents, modeling both text and image contents, along with unstructured layouts. This allows DREAMLLM to learn all conditional, marginal, and joint multimodal distributions effectively. As a result, DREAMLLM is the first MLLM capable of generating free-form interleaved content. Comprehensive experiments highlight DREAMLLM’s superior performance as a zero-shot multimodal generalist, reaping from the enhanced learning synergy. 1 INTRODUCTION “What I cannot create, I do not understand.” Richard P. Feynman, on his blackboard at the time of his death, 1988 Content comprehension and creation in multimodality are crucial and among the ultimate courses of machine intelligence (Sternberg, 1985; Legg & Hutter, 2007). To this end, Multimodal Large Language Models (MLLMs) (Alayrac et al., 2022; Hao et al., 2022; Huang et al., 2023) have emerged as extensions of the successful GPT-style Large Language Models (LLMs) (Brown et al., 2020; Zhang et al., 2022; OpenAI, 2022; 2023; Chen et al., 2023b; Touvron et al., 2023a;b) into visual realm. Recognized as foundation models (Bommasani et al., 2021), MLLMs have achieved unprecedented progress in multimodal comprehension capabilities. These advanced models typically enhance LLMs by incorporating images as multimodal inputs, such as CLIP features (Radford et al., 2021), to facilitate language-output multimodal comprehension. Their aim is to capture multimodal conditional or marginal distributions via a language posterior. However, multimodal creation, which involves generating images, texts, or both, necessitates a universal generative model that simultaneously learns language and image posteriors—currently underexplored. Until very recently, some concurrent works have shown success in conditional image generation using MLLMs (Koh et al., 2023; Sun et al., 2023b). As depicted in Fig. 1, these methods compel MLLMs to produce either discrete or continuous conditional embeddings that explicitly align with a pretrained CLIP encoder, which could later be used by a pretrained Stable Diffusion (SD) (Rombach et al., 2022) model for image generation. However, due to an inherent modality gap (Liang et al., 2022), CLIP semantics focus predominantly on modality-shared information, often overlooking modality-specific knowledge that could enhance multimodal comprehension. Consequently, these studies have not fully realized the potential learning synergy between multimodal creation and comprehension, have shown only marginal improvements in creativity, and remain deficient in multimodal comprehension. In this work, we introduce DREAMLLM, universally learning image and text posteriors with expected creation & comprehension synergy, based on the following two de-facto designing principles: i. Generate Everything as It Is Different from existing works that generate intermediate image representations like CLIP embeddings during training, DREAMLLM not only takes all modalities raw data as inputs but also as outputs in a truly end-to-end fashion (i.e., outputs are identical to inputs, see Fig. 1). The challenge lies in enabling MLLMs to learn the image posterior without compromising their comprehension capabilities. To address this, we introduce dream queries, a set of learnable embeddings that encapsulate the semantics encoded by MLLMs. This approach avoids altering the output space of MLLMs. Raw images are then decoded by the SD image decoder conditioned on these semantics. In this fashion, the pretrained SD acts as the score function (Ho et al., 2020). The image posterior is thus modeled by direct sampling in the pixel space, facilitated by score distillation (van den Oord et al., 2018; Poole et al., 2023). ii. Interleaved Generative Pre-Training (I-GPT) DREAMLLM is trained to generate interleaved multimodal corpora from the internet (Zhu et al., 2023b), both encoding and decoding interleaved image-text multimodal inputs. Unlike encoding multimodal inputs as in existing methods, decoding interleaved multimodal outputs is challenging due to the complex interleaving layout structures and the long-context requirement of image. Our approach tackles the interleaved layout learning using a unique token that predicts the placement of images within texts. Harnessing DREAMLLM’s causal nature, all contents are generated with history multimodal contexts of any length. This interleaved generative pretraining (I-GPT) inherently forms all joint, marginal, and conditional distributions of images and texts in the document, leading to a learning synergy that grounds DREAMLLM’s comprehension in creation and vice versa. Extensive experiments across various vision-language comprehension, content creation, and languageonly tasks demonstrate DREAMLLM’s superior performance as a zero-shot multimodal generalist. For instance, DREAMLLM-7B achieves an 8.46 FID on MS-COCO and sets a new standard with 49.1/35.9 scores on MMBench and MM-Vet evaluations, respectively. Moreover, we delve into the learning synergy between comprehension and creation, revealing decent in-context generation capabilities. With I-GPT pretraining, DREAMLLM generates interleaved documents following human prompts after supervised fine-tuning on instruction-following data, curated with GPT-4. To our knowledge, this work is the first to enable MLLMs to create free-form interleaved content with a learning synergy on both sides. As a foundational learning framework, DREAMLLM is adaptable across all modalities, laying a promising foundation for future multimodal learning research. This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. Authors:
(1) Runpei Dong, Xi’an Jiaotong University and Internship at MEGVII;
(2) Chunrui Han, MEGVII Technology;
(3) Yuang Peng, Tsinghua University and Internship at MEGVII;
(4) Zekun Qi, Xi’an Jiaotong University and Internship at MEGVII;
(5) Zheng Ge, MEGVII Technology;
(6) Jinrong Yang, HUST and Internship at MEGVII;
(7) Liang Zhao, MEGVII Technology;
(8) Jianjian Sun, MEGVII Technology;
(9) Hongyu Zhou, MEGVII Technology;
(10) Haoran Wei, MEGVII Technology;
(11) Xiangwen Kong, MEGVII Technology;
(12) Xiangyu Zhang, MEGVII Technology and a Project leader;
(13) Kaisheng Ma, Tsinghua University and a Corresponding author;
(14) Li Yi, Tsinghua University, a Corresponding authors and Project leader. Table of Links Abstract and 1 Introduction Abstract and 1 Introduction 2 Background & Problem Statement 2 Background & Problem Statement 2.1 How can we use MLLMs for Diffusion Synthesis that Synergizes both sides? 2.1 How can we use MLLMs for Diffusion Synthesis that Synergizes both sides? 3 DreamLLM 3 DreamLLM 3.1 End-to-End Interleaved generative Pretraining (I-GPT) 3.1 End-to-End Interleaved generative Pretraining (I-GPT) 3.2 Model Training 3.2 Model Training 4 Experiments and 4.1 Multimodal Comprehension 4 Experiments and 4.1 Multimodal Comprehension 4.2 Text-Conditional Image Synthesis 4.2 Text-Conditional Image Synthesis 4.3 Multimodal Joint Creation & Comprehension 4.3 Multimodal Joint Creation & Comprehension 5 Discussions 5.1 Synergy between creation & Comprehension? 5.1 Synergy between creation & Comprehension? 5. 2 What is learned by DreamLLM? 5. 2 What is learned by DreamLLM? 6 Related Works 6 Related Works 7 Conclusions and References 7 Conclusions and References A Additional Experiments A Additional Experiments B Additional Qualitative Examples B Additional Qualitative Examples C Implementation Details C Implementation Details D Additional Related Works D Additional Related Works E Limitations, Failure Cases & Future Works E Limitations, Failure Cases & Future Works ABSTRACT This paper presents DREAMLLM, a learning framework that first achieves versatile Multimodal Large Language Models (MLLMs) empowered with frequently overlooked synergy between multimodal comprehension and creation. DREAMLLM operates on two fundamental principles. The first focuses on the generative modeling of both language and image posteriors by direct sampling in the raw multimodal space. This approach circumvents the limitations and information loss inherent to external feature extractors like CLIP, and a more thorough multimodal understanding is obtained. Second, DREAMLLM fosters the generation of raw, interleaved documents, modeling both text and image contents, along with unstructured layouts. This allows DREAMLLM to learn all conditional, marginal, and joint multimodal distributions effectively. As a result, DREAMLLM is the first MLLM capable of generating free-form interleaved content. Comprehensive experiments highlight DREAMLLM’s superior performance as a zero-shot multimodal generalist, reaping from the enhanced learning synergy. 1 INTRODUCTION “What I cannot create, I do not understand.” Richard P. Feynman, on his blackboard at the time of his death, 1988 Richard P. Feynman, on his blackboard at the time of his death, 1988 Content comprehension and creation in multimodality are crucial and among the ultimate courses of machine intelligence (Sternberg, 1985; Legg & Hutter, 2007). To this end, Multimodal Large Language Models (MLLMs) (Alayrac et al., 2022; Hao et al., 2022; Huang et al., 2023) have emerged as extensions of the successful GPT-style Large Language Models (LLMs) (Brown et al., 2020; Zhang et al., 2022; OpenAI, 2022; 2023; Chen et al., 2023b; Touvron et al., 2023a;b) into visual realm. Recognized as foundation models (Bommasani et al., 2021), MLLMs have achieved unprecedented progress in multimodal comprehension capabilities. These advanced models typically enhance LLMs by incorporating images as multimodal inputs, such as CLIP features (Radford et al., 2021), to facilitate language-output multimodal comprehension. Their aim is to capture multimodal conditional or marginal distributions via a language posterior. However, multimodal creation, which involves generating images, texts, or both, necessitates a universal generative model that simultaneously learns language and image posteriors—currently underexplored. Until very recently, some concurrent works have shown success in conditional image generation using MLLMs (Koh et al., 2023; Sun et al., 2023b). As depicted in Fig. 1, these methods compel MLLMs to produce either discrete or continuous conditional embeddings that explicitly align with a pretrained CLIP encoder, which could later be used by a pretrained Stable Diffusion (SD) (Rombach et al., 2022) model for image generation. However, due to an inherent modality gap (Liang et al., 2022), CLIP semantics focus predominantly on modality-shared information, often overlooking modality-specific knowledge that could enhance multimodal comprehension. Consequently, these studies have not fully realized the potential learning synergy between multimodal creation and comprehension, have shown only marginal improvements in creativity, and remain deficient in multimodal comprehension. In this work, we introduce DREAMLLM, universally learning image and text posteriors with expected creation & comprehension synergy, based on the following two de-facto designing principles: i. Generate Everything as It Is Different from existing works that generate intermediate image representations like CLIP embeddings during training, DREAMLLM not only takes all modalities raw data as inputs but also as outputs in a truly end-to-end fashion (i.e., outputs are identical to inputs, see Fig. 1). i. Generate Everything as It Is The challenge lies in enabling MLLMs to learn the image posterior without compromising their comprehension capabilities. To address this, we introduce dream queries, a set of learnable embeddings that encapsulate the semantics encoded by MLLMs. This approach avoids altering the output space of MLLMs. Raw images are then decoded by the SD image decoder conditioned on these semantics. In this fashion, the pretrained SD acts as the score function (Ho et al., 2020). The image posterior is thus modeled by direct sampling in the pixel space, facilitated by score distillation (van den Oord et al., 2018; Poole et al., 2023). ii. Interleaved Generative Pre-Training (I-GPT) DREAMLLM is trained to generate interleaved multimodal corpora from the internet (Zhu et al., 2023b), both encoding and decoding interleaved image-text multimodal inputs. Unlike encoding multimodal inputs as in existing methods, decoding interleaved multimodal outputs is challenging due to the complex interleaving layout structures and the long-context requirement of image. ii. Interleaved Generative Pre-Training (I-GPT) Our approach tackles the interleaved layout learning using a unique token that predicts the placement of images within texts. Harnessing DREAMLLM’s causal nature, all contents are generated with history multimodal contexts of any length. This interleaved generative pretraining (I-GPT) inherently forms all joint, marginal, and conditional distributions of images and texts in the document, leading to a learning synergy that grounds DREAMLLM’s comprehension in creation and vice versa. Extensive experiments across various vision-language comprehension, content creation, and languageonly tasks demonstrate DREAMLLM’s superior performance as a zero-shot multimodal generalist. For instance, DREAMLLM-7B achieves an 8.46 FID on MS-COCO and sets a new standard with 49.1/35.9 scores on MMBench and MM-Vet evaluations, respectively. Moreover, we delve into the learning synergy between comprehension and creation, revealing decent in-context generation capabilities. With I-GPT pretraining, DREAMLLM generates interleaved documents following human prompts after supervised fine-tuning on instruction-following data, curated with GPT-4. To our knowledge, this work is the first to enable MLLMs to create free-form interleaved content with a learning synergy on both sides. As a foundational learning framework, DREAMLLM is adaptable across all modalities, laying a promising foundation for future multimodal learning research. This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. available on arxiv Authors: (1) Runpei Dong, Xi’an Jiaotong University and Internship at MEGVII; (2) Chunrui Han, MEGVII Technology; (3) Yuang Peng, Tsinghua University and Internship at MEGVII; (4) Zekun Qi, Xi’an Jiaotong University and Internship at MEGVII; (5) Zheng Ge, MEGVII Technology; (6) Jinrong Yang, HUST and Internship at MEGVII; (7) Liang Zhao, MEGVII Technology; (8) Jianjian Sun, MEGVII Technology; (9) Hongyu Zhou, MEGVII Technology; (10) Haoran Wei, MEGVII Technology; (11) Xiangwen Kong, MEGVII Technology; (12) Xiangyu Zhang, MEGVII Technology and a Project leader; (13) Kaisheng Ma, Tsinghua University and a Corresponding author; (14) Li Yi, Tsinghua University, a Corresponding authors and Project leader. Authors: Authors: (1) Runpei Dong, Xi’an Jiaotong University and Internship at MEGVII; (2) Chunrui Han, MEGVII Technology; (3) Yuang Peng, Tsinghua University and Internship at MEGVII; (4) Zekun Qi, Xi’an Jiaotong University and Internship at MEGVII; (5) Zheng Ge, MEGVII Technology; (6) Jinrong Yang, HUST and Internship at MEGVII; (7) Liang Zhao, MEGVII Technology; (8) Jianjian Sun, MEGVII Technology; (9) Hongyu Zhou, MEGVII Technology; (10) Haoran Wei, MEGVII Technology; (11) Xiangwen Kong, MEGVII Technology; (12) Xiangyu Zhang, MEGVII Technology and a Project leader; (13) Kaisheng Ma, Tsinghua University and a Corresponding author; (14) Li Yi, Tsinghua University, a Corresponding authors and Project leader.

What Is DreamLLM? Everything You Need to Know About the Learning Framework

Why Everyone is Wrong About the Future of Work™

কেন সবাই কাজের ভবিষ্যত সম্পর্কে ভুল ™

Warum alle falsch liegen, wenn sie über die Zukunft der Arbeit sprechen™

Por qué todo el mundo está equivocado sobre el futuro del trabajo™

Pourquoi tout le monde se trompe sur l’avenir du travail™

काम के भविष्य के बारे में हर कोई गलत क्यों है™

誰もが仕事の未来について間違っている理由™

왜 모든 사람이 미래의 일에 대해 틀렸는가?™

Por que todos estão errados sobre o futuro do trabalho™

Почему все ошибаются относительно будущего работы™

Neden Herkes Çalışmanın Geleceği Hakkında Yanılıyor?™

Tại sao mọi người đều sai về tương lai của công việc™

为什么每个人对工作的未来都有误解™

Now that we've covered the US presidential elections and most of the Q3 earnings, I want to take a moment to discuss Intel and some of the excellent reporting from Reuters about the company's downfall. Intel, as it is today, appears very insignificant when compared to AMD and Nvidia, both of whom have made some very shrewd bets on artificial intelligence, especially Nvidia. Though in Nvidia's case, I've argued that its technology was simply at the right place at the right time. As I said so in March: Nvidia never really started as an artificial intelligence company the way OpenAI and so many others like it did. It started out as a company that produced graphic cards for video games, and for a long, long time, this is what Nvidia was known for. But as an offshoot of all the research it did in producing better and better graphic cards, it ended up creating tensor cores for real-time deep learning image enhancement and upscaling technologies for use in video games. Little did Nvidia know that this technology would then serve as the backbone of LLMs, instantly transforming it as the arms dealer of choice in corporations' pursuit of the perfect AI. At the time Nvidia was developing its technology, however, Intel engineers were scoffing at video gaming chip architecture and believed that a CPU could more effectively handle processing tasks needed to build and run AI models. As time has bore witness, Intel and its engineers were wrong. Very, very wrong. So when demand for AI chips skyrocketed, Intel was left with its pants down, scrambling to get a product out. Ironically, Intel was this close to hitting the jackpot that Nvidia eventually ended up striking. In 2017 and 2018, the company had the opportunity to invest in OpenAI but decided against it because then-CEO Bob Swan did not believe generative AI would make it to the market soon enough for it to be worthwhile the $1 billion Intel would have had to pay for a 15% stake. OpenAI also offered a further 15% stake if Intel agreed to make hardware for it at cost price — but the company refused, leaving OpenAI to rely on Nvidia's chips instead. GET PUBLISHED ON HACKERNOON.
Join a global network of 45,000+ published devs, builders,   founders, makers, VCs, hodlrs, and hackers. Start submitting your tech   stories and tutorials to get published FREE on HackerNoon — no pop-ups,   no paywalls. But that was then. Intel hasn't fared any better under current CEO Pat Gelsinger's leadership either. Intel rehired Gelsinger in 2021 to help lead the company out of a tough spot: it was losing out to the likes of Taiwan Semiconductor Manufacturing and Samsung Electronics on chip-making technology, and the idea was to have Gelsinger, who had been with Intel for most of his career and helped develop some of its most successful personal-computer chips, lead a turn around. While Gelsinger's heart was in the right place, his turnaround plan didn't really do much to turn the company's fortunes around. In fact, things came to a head this past summer when Intel stunned both employees and investors with its poor quarterly results and announced sweeping measures (cost cuts, headcount reduction, etc.) to become profitable again. All of this could have been avoided had it not been for some key mistakes Gelsinger made after taking reins of the company. According to Reuters, soon after becoming CEO, Gelsinger offended Taiwan Semiconductor Manufacturing by calling Taiwan unstable. Gelsinger's "Taiwan is not a stable place" comment had more to do with his push to develop a US-based supply chain for chip manufacturing, but was so poorly received by TSMC that it reneged on a 40% discount it was willing to offer Intel on chips it designed but could not produce. That discount would have really helped Intel's margins, but, nope, Gelsinger was too busy commenting on Taiwan. Making matters worse, Gelsinger kept doubling down on his plan to develop a manufacturing footprint that could rival TSMC's, but on American and European soil instead. Problem is, customers have no reason to shy away from TSMC, particularly considering that Intel's foundry ambitions have run into technical challenges. Combined with the fact that Intel kept sticking to producing chips for desktop computers and servers, and you have a clearer picture of why the company is in such dire straits today. The list of fumbles goes on, but what must really make this sting for Gelsinger is that just a few weeks ago it was announced that Nvidia would take the company's place on the Dow Jones Industrial Average. Today, Nvidia is the most valuable company in the world. The company has a pile of cash. And CEO Jensen Huang's 3.8% stake in Nvidia is worth more than Intel's market value combined, Reuters reports. And to think Intel could have prevented all of this from happening. The way I see it, Intel is either going to be bought or Gelsinger will be fired, possibly both. I'll be here to report on it when it happens. Intel was #198 on HackerNoon's Tech Company Rankings this week. In Other News.. 📰 Bitcoin ETF Options Attract $2B on Day One, Shifting BTC’s Market Structure — via CoinDesk
H, the AI startup that raised $220M, launches its first product: Runner H for ‘agentic’ applications — via TechCrunch
With Musk in tow, Trump eyes changes to government policies on AI and its dangers — via CNN
Apple to urge judge to end US smartphone monopoly case — via Reuters
Trump 2.0 will face Big Tech antitrust test — via Axios
FTX co-founder Gary Wang avoids prison time for role in crypto fraud — via CNBC And that's a wrap! Don't forget to share this newsletter with your family and friends! See y'all next week. PEACE! ☮️ — Sheharyar Khan, Editor, Business Tech @ HackerNoon *All rankings are current as of Monday. To see how the rankings have changed, please visit HackerNoon's Tech Company Rankings page. Tech, What the Heck!? is a once-weekly newsletter written by HackerNoon editors that combine HackerNoon's proprietary data with news-worthy tech stories from around the internet. Humorous and insightful, the newsletter recaps trending events that are shaping the world of tech. Subscribe here. Now that we've covered the US presidential elections and most of the Q3 earnings , I want to take a moment to discuss Intel and some of the excellent reporting from Reuters about the company's downfall. US presidential elections US presidential elections Q3 earnings Q3 earnings Intel Intel Intel, as it is today, appears very insignificant when compared to AMD and Nvidia , both of whom have made some very shrewd bets on artificial intelligence, especially Nvidia. Though in Nvidia's case, I've argued that its technology was simply at the right place at the right time. AMD AMD Nvidia Nvidia As I said so in March: said said Nvidia never really started as an artificial intelligence company the way OpenAI and so many others like it did. It started out as a company that produced graphic cards for video games, and for a long, long time, this is what Nvidia was known for. But as an offshoot of all the research it did in producing better and better graphic cards, it ended up creating tensor cores for real-time deep learning image enhancement and upscaling technologies for use in video games. Little did Nvidia know that this technology would then serve as the backbone of LLMs, instantly transforming it as the arms dealer of choice in corporations' pursuit of the perfect AI. Nvidia never really started as an artificial intelligence company the way OpenAI and so many others like it did. It started out as a company that produced graphic cards for video games, and for a long, long time, this is what Nvidia was known for. But as an offshoot of all the research it did in producing better and better graphic cards, it ended up creating tensor cores for real-time deep learning image enhancement and upscaling technologies for use in video games. Little did Nvidia know that this technology would then serve as the backbone of LLMs, instantly transforming it as the arms dealer of choice in corporations' pursuit of the perfect AI. Nvidia never really started as an artificial intelligence company the way OpenAI and so many others like it did. It started out as a company that produced graphic cards for video games, and for a long, long time, this is what Nvidia was known for. But as an offshoot of all the research it did in producing better and better graphic cards, it ended up creating tensor cores for real-time deep learning image enhancement and upscaling technologies for use in video games. Little did Nvidia know that this technology would then serve as the backbone of LLMs, instantly transforming it as the arms dealer of choice in corporations' pursuit of the perfect AI. At the time Nvidia was developing its technology, however, Intel engineers were scoffing at video gaming chip architecture and believed that a CPU could more effectively handle processing tasks needed to build and run AI models. scoffing scoffing As time has bore witness, Intel and its engineers were wrong. Very, very wrong. So when demand for AI chips skyrocketed, Intel was left with its pants down, scrambling to get a product out. Ironically, Intel was this close to hitting the jackpot that Nvidia eventually ended up striking. In 2017 and 2018, the company had the opportunity to invest in OpenAI but decided against it because then-CEO Bob Swan did not believe generative AI would make it to the market soon enough for it to be worthwhile the $1 billion Intel would have had to pay for a 15% stake. this OpenAI OpenAI OpenAI also offered a further 15% stake if Intel agreed to make hardware for it at cost price — but the company refused, leaving OpenAI to rely on Nvidia's chips instead. GET PUBLISHED ON HACKERNOON. Join a global network of 45,000+ published devs, builders,   founders, makers, VCs, hodlrs, and hackers. Start submitting your tech   stories and tutorials to get published FREE on HackerNoon — no pop-ups,   no paywalls. GET PUBLISHED ON HACKERNOON. GET PUBLISHED ON HACKERNOON . GET PUBLISHED ON HACKERNOON Join a global network of 45,000+ published devs, builders,   founders, makers, VCs, hodlrs, and hackers. Start submitting your tech   stories and tutorials to get published FREE on HackerNoon — no pop-ups,   no paywalls. Join a global network of 45,000+ published devs, builders,   founders, makers, VCs, hodlrs, and hackers. Start submitting your tech   stories and tutorials to get published FREE on HackerNoon — no pop-ups,   no paywalls. But that was then. Intel hasn't fared any better under current CEO Pat Gelsinger's leadership either. Intel rehired Gelsinger in 2021 to help lead the company out of a tough spot: it was losing out to the likes of Taiwan Semiconductor Manufacturing and Samsung Electronics on chip-making technology, and the idea was to have Gelsinger, who had been with Intel for most of his career and helped develop some of its most successful personal-computer chips, lead a turn around. rehired rehired While Gelsinger's heart was in the right place, his turnaround plan didn't really do much to turn the company's fortunes around. In fact, things came to a head this past summer when Intel stunned both employees and investors with its poor quarterly results and announced sweeping measures (cost cuts, headcount reduction, etc.) to become profitable again. stunned stunned All of this could have been avoided had it not been for some key mistakes Gelsinger made after taking reins of the company. According to Reuters, soon after becoming CEO, Gelsinger offended Taiwan Semiconductor Manufacturing by calling Taiwan unstable. According According Gelsinger's "Taiwan is not a stable place" comment had more to do with his push to develop a US-based supply chain for chip manufacturing, but was so poorly received by TSMC that it reneged on a 40% discount it was willing to offer Intel on chips it designed but could not produce. That discount would have really helped Intel's margins, but, nope, Gelsinger was too busy commenting on Taiwan. Making matters worse, Gelsinger kept doubling down on his plan to develop a manufacturing footprint that could rival TSMC's, but on American and European soil instead. Problem is, customers have no reason to shy away from TSMC, particularly considering that Intel's foundry ambitions have run into technical challenges. Combined with the fact that Intel kept sticking to producing chips for desktop computers and servers, and you have a clearer picture of why the company is in such dire straits today. The list of fumbles goes on, but what must really make this sting for Gelsinger is that just a few weeks ago it was announced that Nvidia would take the company's place on the Dow Jones Industrial Average. announced announced Today, Nvidia is the most valuable company in the world. The company has a pile of cash. And CEO Jensen Huang's 3.8% stake in Nvidia is worth more than Intel's market value combined, Reuters reports . reports reports And to think Intel could have prevented all of this from happening. The way I see it, Intel is either going to be bought or Gelsinger will be fired, possibly both. I'll be here to report on it when it happens. Intel was #198 on HackerNoon's Tech Company Rankings this week. In Other News.. 📰 In Other News.. 📰 Bitcoin ETF Options Attract $2B on Day One, Shifting BTC’s Market Structure — via CoinDesk H, the AI startup that raised $220M, launches its first product: Runner H for ‘agentic’ applications — via TechCrunch With Musk in tow, Trump eyes changes to government policies on AI and its dangers — via CNN Apple to urge judge to end US smartphone monopoly case — via Reuters Trump 2.0 will face Big Tech antitrust test — via Axios FTX co-founder Gary Wang avoids prison time for role in crypto fraud — via CNBC Bitcoin ETF Options Attract $2B on Day One, Shifting BTC’s Market Structure — via CoinDesk CoinDesk CoinDesk H, the AI startup that raised $220M, launches its first product: Runner H for ‘agentic’ applications — via TechCrunch TechCrunch TechCrunch With Musk in tow, Trump eyes changes to government policies on AI and its dangers — via CNN CNN CNN Apple to urge judge to end US smartphone monopoly case — via Reuters Reuters Reuters Trump 2.0 will face Big Tech antitrust test — via Axios Axios Axios FTX co-founder Gary Wang avoids prison time for role in crypto fraud — via CNBC CNBC CNBC And that's a wrap! Don't forget to share this newsletter with your family and friends! See y'all next week. PEACE! ☮️ — Sheharyar Khan , Editor, Business Tech @ HackerNoon Sheharyar Khan Sheharyar Khan *All rankings are current as of Monday. To see how the rankings have changed, please visit HackerNoon's Tech Company Rankings page. *All rankings are current as of Monday. To see how the rankings have changed, please visit HackerNoon's Tech Company Rankings page. *All rankings are current as of Monday. To see how the rankings have changed, please visit HackerNoon's Tech Company Rankings page. Tech Company Rankings Tech Company Rankings Tech, What the Heck!? is a once-weekly newsletter written by HackerNoon editors that combine HackerNoon's proprietary data with news-worthy tech stories from around the internet. Humorous and insightful, the newsletter recaps trending events that are shaping the world of tech. Subscribe here. Tech, What the Heck!? is a once-weekly newsletter written by HackerNoon editors that combine HackerNoon's proprietary data with news-worthy tech stories from around the internet. Humorous and insightful, the newsletter recaps trending events that are shaping the world of tech. Subscribe here. Tech, What the Heck!? is a once-weekly newsletter written by HackerNoon editors that combine HackerNoon's proprietary data with news-worthy tech stories from around the internet. Humorous and insightful, the newsletter recaps trending events that are shaping the world of tech. Subscribe here . here here

Intel (No Longer) Inside, or How One CEO Messed Everything Up

ইন্টেল (আর নেই) ভিতরে, বা কীভাবে একজন সিইও সবকিছুকে গোলমাল করেছে

Intel (nicht mehr) Inside oder wie ein CEO alles vermasselt hat

Intel (ya no está) dentro, o cómo un CEO lo arruinó todo

Intel (n'est plus) à l'intérieur, ou comment un PDG a tout gâché

इंटेल (अब नहीं) अंदर, या कैसे एक सीईओ ने सब कुछ गड़बड़ कर दिया

インテル（もはや）内部、あるいはある CEO がいかにしてすべてを台無しにしたか

Intel(더 이상) 내부 또는 한 CEO가 모든 것을 어떻게 망쳤는지

Intel (não mais) dentro, ou como um CEO bagunçou tudo

Intel (больше не) Inside, или как один генеральный директор все испортил

Intel (Artık) İçeride Değil veya Bir CEO Her Şeyi Nasıl Mahvetti

Intel (Không Còn) Bên Trong, hay Một CEO Đã Làm Hỏng Mọi Thứ Như Thế Nào

英特尔（不再）内部，或者一位首席执行官如何把一切都搞砸了

In today's tech world with fierce competition and tight deadlines, it is useful to explore new and exciting tools to improve the workflow. Open-source tools are there to help; with some significant advantages over other tools - they're customizable, offer more control, and are community-driven. I decided to handpick 9 powerful open-source projects that will enhance your cloud workflow, project management, user feedback, automation, and more. Whether you are a beginner or a seasoned developer there should be something for everyone to elevate your toolkit and inspire your next big project. Let’s dive in and hopefully, you will find some tools that you can use in the future! 1. StackQL – Manage cloud services with SQL StackQL offers a revolutionary approach to deploy, query, and interact with cloud services and resources using Structured query language (SQL). That means you no longer need to learn different APIs for each provider. StackQL parses your SQL statements and transpiles them into API requests that all your favorite major cloud providers would understand. Some of the best features include: SQL-based querying: unified SQL semantics for various cloud providers lets you save time and manage cloud services more efficiently.


Multi-cloud support: manage Google Cloud, AWS, Azure, Digital Ocean, etc. resources all from a single interface.


Several usage modes: use StackQL via the interactive REPL shell, via the command line utility, or execute StackQL statements in an input file (IQL file).


Powerful operations: perform live queries against cloud resources, make configuration changes, manage cloud infrastructure, and more.


Available for various OS: available to download for Windows, MacOS, and Linux. Ready to reduce setup times and simplify your cloud configurations? Try StackQL yourself and explore all the awesome features it comes with! ⭐ Star StackQL on GitHub: github.com/stackql/stackql 🌎 Website: stackql.io Thanks to the StackQL team for sponsoring this article. 2. Flagsmith – Feature flag management Flagsmith is an open-source feature flag management tool. You can use hosted API, deploy to your private cloud, or run as on-premise software. It enables gradual feature rollouts for safer deployments and easier A/B testing, reducing the risk of bugs and ensuring faster iteration cycles. 👨‍💻 GitHub Repository: https://github.com/Flagsmith/flagsmith 🌎 Website: flagsmith.com 3. Formbricks – User feedback and analytics Formbricks is an open-source tool for collecting user feedback and form analytics. It helps devs understand user behavior, allowing them to improve performance, reduce friction, and enhance the overall UX. 👨‍💻 GitHub Repository: https://github.com/formbricks/formbricks 🌎 Website: formbricks.com 4. Dub – URL shortener with analytics Dub is an open-source URL shortener with built-in analytics. It simplifies link management for websites, marketing, and projects. By allowing developers to manage and analyze shortened URLs, Dub streamlines project workflows, making link tracking and data gathering much easier. 👨‍💻 GitHub Repository: https://github.com/dubinc/dub 🌎 Website: dub.co 5. AppFlowy – Project & notes management AppFlowy is an open-source, privacy-focused alternative to Notion. It provides a flexible workspace for managing tasks, notes, and general knowledge. AppFlowy offers teams control over their data and customizations, allowing them to create a tailored workspace while benefiting from privacy and scalability. 👨‍💻 GitHub Repository: https://github.com/AppFlowy-IO/AppFlowy 🌎 Website: appflowy.io 6. APITable – Visual database for collaborative apps APITable combines the ease of spreadsheets with the power of databases. It helps teams collaborate on data-driven projects with visual and interactive tables. APITable simplifies data management and collaboration, allowing developers to organize and automate workflows with minimal setup and technical knowledge. 👨‍💻 GitHub Repository: https://github.com/apitable/apitable 🌎 Website: aitable.ai 7. Flowise – Build LLM workflows with flowcharts Flowise is an open-source visual tool that simplifies building workflows for large language models (LLMs). It’s designed to integrate AI with apps easily. Flowise allows developers to visually manage complex LLM workflows, speeding up the integration of AI into their applications without deep coding knowledge. 👨‍💻 GitHub Repository: https://github.com/FlowiseAI/Flowise 🌎 Website: flowiseai.com 8. Jan – Use AI with full control and privacy Jan is an open-source tool; a ChatGPT alternative that runs 100% offline on your device but with more control, flexibility, and no privacy concerns. Jan enhances productivity by allowing you to ask questions, brainstorm, and learn from the AI running on your device to be more productive. 👨‍💻 GitHub Repository: https://github.com/janhq/jan 🌎 Website: jan.ai 9. Docsify – Documentation site generator Docsify is an open-source documentation site generator that creates beautiful, responsive docs from markdown files. It simplifies the process of generating documentation, making it easy for developers to maintain and update their docs without complex builds. 👨‍💻 GitHub Repository: https://github.com/docsifyjs/docsify 🌎 Website: docsify.js.org Did you like the resources? Here is more 👇 Join 6000+ others to receive the best DEV resources, tools, productivity tips, and career growth advice I discover by subscribing to my newsletter! Also, connect with me on Twitter, LinkedIn, and GitHub! Writing has always been my passion and it gives me pleasure to help and inspire people. If you want to get featured or partner up, feel free to get in touch! Also, published here. In today's tech world with fierce competition and tight deadlines, it is useful to explore new and exciting tools to improve the workflow. Open-source tools are there to help; with some significant advantages over other tools - they're customizable, offer more control, and are community-driven. I decided to handpick 9 powerful open-source projects that will enhance your cloud workflow, project management, user feedback, automation, and more. Whether you are a beginner or a seasoned developer there should be something for everyone to elevate your toolkit and inspire your next big project. Let’s dive in and hopefully, you will find some tools that you can use in the future! 1. StackQL – Manage cloud services with SQL StackQL StackQL offers a revolutionary approach to deploy, query, and interact with cloud services and resources using Structured query language (SQL). That means you no longer need to learn different APIs for each provider. StackQL parses your SQL statements and transpiles them into API requests that all your favorite major cloud providers would understand. Some of the best features include: Some of the best features include: SQL-based querying: unified SQL semantics for various cloud providers lets you save time and manage cloud services more efficiently. Multi-cloud support: manage Google Cloud, AWS, Azure, Digital Ocean, etc. resources all from a single interface. Several usage modes: use StackQL via the interactive REPL shell, via the command line utility, or execute StackQL statements in an input file (IQL file). Powerful operations: perform live queries against cloud resources, make configuration changes, manage cloud infrastructure, and more. Available for various OS: available to download for Windows, MacOS, and Linux. SQL-based querying: unified SQL semantics for various cloud providers lets you save time and manage cloud services more efficiently. SQL-based querying : unified SQL semantics for various cloud providers lets you save time and manage cloud services more efficiently. SQL-based querying Multi-cloud support: manage Google Cloud, AWS, Azure, Digital Ocean, etc. resources all from a single interface. Multi-cloud support: manage Google Cloud, AWS, Azure, Digital Ocean, etc. resources all from a single interface. Multi-cloud support: Several usage modes: use StackQL via the interactive REPL shell, via the command line utility, or execute StackQL statements in an input file (IQL file). Several usage modes: use StackQL via the interactive REPL shell, via the command line utility, or execute StackQL statements in an input file (IQL file). Several usage modes: Powerful operations: perform live queries against cloud resources, make configuration changes, manage cloud infrastructure, and more. Powerful operations: perform live queries against cloud resources, make configuration changes, manage cloud infrastructure, and more. Powerful operations: Available for various OS: available to download for Windows, MacOS, and Linux. Available for various OS : available to download for Windows, MacOS, and Linux. Available for various OS Ready to reduce setup times and simplify your cloud configurations? Try StackQL yourself and explore all the awesome features it comes with! ⭐ Star StackQL on GitHub: github.com/stackql/stackql Star StackQL on GitHub: github.com/stackql/stackql github.com/stackql/stackql 🌎 Website: stackql.io 🌎 Website: stackql.io stackql.io Thanks to the StackQL team for sponsoring this article. 2. Flagsmith – Feature flag management Flagsmith Flagsmith is an open-source feature flag management tool. You can use hosted API, deploy to your private cloud, or run as on-premise software. It enables gradual feature rollouts for safer deployments and easier A/B testing, reducing the risk of bugs and ensuring faster iteration cycles. 👨‍💻 GitHub Repository: https://github.com/Flagsmith/flagsmith GitHub Repository: https://github.com/Flagsmith/flagsmith 🌎 Website: flagsmith.com 🌎 Website: flagsmith.com 3. Formbricks – User feedback and analytics Formbricks Formbricks is an open-source tool for collecting user feedback and form analytics. It helps devs understand user behavior, allowing them to improve performance, reduce friction, and enhance the overall UX. 👨‍💻 GitHub Repository: https://github.com/formbricks/formbricks GitHub Repository: https://github.com/formbricks/formbricks 🌎 Website: formbricks.com 🌎 Website: formbricks.com 4. Dub – URL shortener with analytics Dub Dub is an open-source URL shortener with built-in analytics. It simplifies link management for websites, marketing, and projects. By allowing developers to manage and analyze shortened URLs, Dub streamlines project workflows, making link tracking and data gathering much easier. 👨‍💻 GitHub Repository: https://github.com/dubinc/dub GitHub Repository: https://github.com/dubinc/dub 🌎 Website: dub.co 🌎 Website: dub.co 5. AppFlowy – Project & notes management AppFlowy AppFlowy is an open-source, privacy-focused alternative to Notion. It provides a flexible workspace for managing tasks, notes, and general knowledge. AppFlowy offers teams control over their data and customizations, allowing them to create a tailored workspace while benefiting from privacy and scalability. 👨‍💻 GitHub Repository: https://github.com/AppFlowy-IO/AppFlowy GitHub Repository: https://github.com/AppFlowy-IO/AppFlowy 🌎 Website: appflowy.io 🌎 Website: appflowy.io 6. APITable – Visual database for collaborative apps APITable APITable combines the ease of spreadsheets with the power of databases. It helps teams collaborate on data-driven projects with visual and interactive tables. APITable simplifies data management and collaboration, allowing developers to organize and automate workflows with minimal setup and technical knowledge. 👨‍💻 GitHub Repository: https://github.com/apitable/apitable GitHub Repository: https://github.com/apitable/apitable 🌎 Website: aitable.ai 🌎 Website: aitable.ai 7. Flowise – Build LLM workflows with flowcharts Flowise Flowise is an open-source visual tool that simplifies building workflows for large language models (LLMs). It’s designed to integrate AI with apps easily. Flowise allows developers to visually manage complex LLM workflows, speeding up the integration of AI into their applications without deep coding knowledge. 👨‍💻 GitHub Repository: https://github.com/FlowiseAI/Flowise GitHub Repository: https://github.com/FlowiseAI/Flowise 🌎 Website: flowiseai.com 🌎 Website: flowiseai.com 8. Jan – Use AI with full control and privacy Jan Jan is an open-source tool; a ChatGPT alternative that runs 100% offline on your device but with more control, flexibility, and no privacy concerns. Jan enhances productivity by allowing you to ask questions, brainstorm, and learn from the AI running on your device to be more productive. 👨‍💻 GitHub Repository: https://github.com/janhq/jan GitHub Repository: https://github.com/janhq/jan 🌎 Website: jan.ai 🌎 Website: jan.ai 9. Docsify – Documentation site generator Docsify Docsify is an open-source documentation site generator that creates beautiful, responsive docs from markdown files. It simplifies the process of generating documentation, making it easy for developers to maintain and update their docs without complex builds. 👨‍💻 GitHub Repository: https://github.com/docsifyjs/docsify GitHub Repository: https://github.com/docsifyjs/docsify 🌎 Website: docsify.js.org 🌎 Website: docsify.js.org Did you like the resources? Here is more 👇 Join 6000+ others to receive the best DEV resources, tools, productivity tips, and career growth advice I discover by subscribing to my newsletter ! my newsletter Also, connect with me on Twitter , LinkedIn , and GitHub ! Twitter LinkedIn GitHub Writing has always been my passion and it gives me pleasure to help and inspire people. If you want to get featured or partner up, feel free to get in touch ! get in touch Also, published here . here

9 Open Source Projects Every Developer Needs to Bookmark for Their Workflow

Stories, News & Mentions about EVERY

Every’s Stories on HackerNoon