正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

🐣 Introducing `smol-developer`! ▸ Human-centric, coherent whole program synthesis ▸ your own junior developer ▸ develop, debug, decompile ▸ open source: ▸ 200 LOC, half english Insights: 💡 100k context can summarize both content and codebases 💡 Markdown is the best prompting DSL 💡 Copy and paste your errors... show more

swyx

189,993 subscribers

1,591,045 次观看 • 3 年前 •via X (Twitter)

喜剧教育科学技术

Anya Rossi• Live Now

Private livecam show

10 条评论

swyx 🔜 @aidotEngineer NYC 的头像

swyx 🔜 @aidotEngineer NYC3 年前

uploaded full resolution onto youtube for those who prefer non potato quality

swyx 🔜 @aidotEngineer NYC 的头像

swyx 🔜 @aidotEngineer NYC3 年前

another more serious run at this problem

swyx 🔜 @aidotEngineer NYC 的头像

swyx 🔜 @aidotEngineer NYC3 年前

We've just added a prompt gallery, thanks to users sending in more examples! thanks to @ultrasoundchad, @RobertCaracaus, @rodavelli, @atlantis2point0 for all the feedback, and @talboren and @thatliuser for adding local options!

swyx 🔜 @aidotEngineer NYC 的头像

swyx 🔜 @aidotEngineer NYC3 年前

ported to JS by @picocreator!

swyx 🔜 @aidotEngineer NYC 的头像

swyx 🔜 @aidotEngineer NYC3 年前

ported to C# by @colhountech!

swyx 🔜 @aidotEngineer NYC 的头像

swyx 🔜 @aidotEngineer NYC3 年前

ported to Go by @traviscline!

swyx 🔜 @aidotEngineer NYC 的头像

swyx 🔜 @aidotEngineer NYC3 年前

smol-developer projects are trending on both HN and YouTube today

PicoCreator - AI Model Builder 🌉 的头像

PicoCreator - AI Model Builder 🌉3 年前

you had to lit a fire in me, to build on this for the next level ✊🔥🔥🔥 complete rewrite has begun for targeting the 8k context size (since most folks would not have 32k/claude access)

swyx 🔜 @aidotEngineer NYC 的头像

swyx 🔜 @aidotEngineer NYC3 年前

thanks for all the support! this project now has more github stars than lines of python as a wise Slovakian-Canadian Youtube influencer once said, English is the hottest new programming language

swyx 🔜 @aidotEngineer NYC 的头像

swyx 🔜 @aidotEngineer NYC3 年前

first mention by a conference speaker

相关视频

🤖 We are entering the second revolution of Personal AI. AI agents aren’t just answering questions, they’re working for you 24/7 to help you stay ahead. Imagine a virtual team that continuously processes data, monitors systems, and executes tasks automatically even while you sleep. With AMD Agent Computers, your AI agents collaborate like a virtual team, maintaining context, sharing insights, and executing tasks as part of a coordinated agentic workflow, something that wasn’t possible before. Everything runs locally, keeping your data private and secure 🛡️ They think deeply, reasoning through problems, connecting insights, and making decisions that multiply your impact. Powered by Ryzen AI Max and Radeon AI PRO, your models, context, and tools live together, no token limits, fully local and private. 💡 Experience what a 24/7 AI team can do to multiply your impact beyond what you thought possible. 🔥 Join us April 30 in San Francisco at AMD Developer Day:

🤖 We are entering the second revolution of Personal AI. AI agents aren’t just answering questions, they’re working for you 24/7 to help you stay ahead. Imagine a virtual team that continuously processes data, monitors systems, and executes tasks automatically even while you sleep. With AMD Agent Computers, your AI agents collaborate like a virtual team, maintaining context, sharing insights, and executing tasks as part of a coordinated agentic workflow, something that wasn’t possible before. Everything runs locally, keeping your data private and secure 🛡️ They think deeply, reasoning through problems, connecting insights, and making decisions that multiply your impact. Powered by Ryzen AI Max and Radeon AI PRO, your models, context, and tools live together, no token limits, fully local and private. 💡 Experience what a 24/7 AI team can do to multiply your impact beyond what you thought possible. 🔥 Join us April 30 in San Francisco at AMD Developer Day:

Jack Huynh

24,609 次观看 • 4 个月前

Claude Skills are a cheat code for DTC creative teams 🤯 One setup, reusable forever. Claude automatically follows your exact creative process — briefs, hooks, ad copy, research — without you explaining anything twice. Perfect for e-comm brands and agencies who are using Claude for creative work but wasting time re-explaining context every single conversation. Here's the problem: You open Claude: you paste in your brand guidelines, explain your brief format, write the copy, you close the chat. Next day, you do it all over again. Every conversation starts from zero. You're burning 20 minutes on setup before you even get to the actual work. Claude Skills fix this: → Write your creative process once as a Skill (a simple markdown file) → Claude reads it automatically whenever the task comes up → Skills compose — research triggers the research Skill, briefs trigger the brief Skill, copy triggers the copy Skill → All in one conversation, all building on each other → Share across your team so everyone gets the same quality output No re-explaining your brand voice. No pasting the same context every chat. No siloed projects that don't talk to each other. What's in the playbook: → The full architecture (how Skills trigger, chain, and compose) → 5 ready-to-use Skill templates built for DTC and agency creative teams → Step-by-step setup from zero to working Skills → How to write instructions that produce consistent output every time → The composability framework for running multi-step creative workflows in one conversation I put together the complete Claude Skills Playbook for DTC brands and creative agencies. Want it for free? >Like this post >Comment "SKILLS" And I'll send it over (must be following so I can DM)

Mike Futia

31,988 次观看 • 7 个月前

I built a Morning Brew-style daily newsletter that writes itself with AI (now at 10,000 daily readers) (and I’m nuts for open-sourcing it) The AI system behind @therecapai clones how a human writer would function, but each step is packaged into an agentic workstream. The full automation handles: - Daily scrapping hundreds of reddit threads, hackernews, twitter posts, Google news API to build a massive data lake of 'daily AI news' - Dozens of custom prompts to pull the top daily stories from this data lake, write short breakdowns, & format the newsletter - Builds custom images based on each story with ChatGPT image generator This was 5 MONTHS of iterating and fine-tuning prompts to get the output and content to an extremely high quality state (no AI slop) and call me crazy, but I'm giving it away for free. Just Retweet, follow me, and comment on this thread "NEWSLETTER" and I'll DM you the full n8n template that you can copy/paste and fine-tune to your use-case as well as a full hour long video breakdown explaining the build. Also you can see the actual contents of the newsletter for yourself with the link below—proof of the quality that is possible with AI. RIP media companies 💀 AI is here.

I built a Morning Brew-style daily newsletter that writes itself with AI (now at 10,000 daily readers) (and I’m nuts for open-sourcing it) The AI system behind @therecapai clones how a human writer would function, but each step is packaged into an agentic workstream. The full automation handles: - Daily scrapping hundreds of reddit threads, hackernews, twitter posts, Google news API to build a massive data lake of 'daily AI news' - Dozens of custom prompts to pull the top daily stories from this data lake, write short breakdowns, & format the newsletter - Builds custom images based on each story with ChatGPT image generator This was 5 MONTHS of iterating and fine-tuning prompts to get the output and content to an extremely high quality state (no AI slop) and call me crazy, but I'm giving it away for free. Just Retweet, follow me, and comment on this thread "NEWSLETTER" and I'll DM you the full n8n template that you can copy/paste and fine-tune to your use-case as well as a full hour long video breakdown explaining the build. Also you can see the actual contents of the newsletter for yourself with the link below—proof of the quality that is possible with AI. RIP media companies 💀 AI is here.

David Roberts

219,613 次观看 • 1 年前

sarah achieves the most impossible developer events - she got a bunch of developers at a writers meetup to sit together…. … and actually write 🤯 thanks to anhtho 🍊 for being our 3rd Writing Mentor! allen helped with AV and will upload both her and Corey Quinn video soon

sarah achieves the most impossible developer events - she got a bunch of developers at a writers meetup to sit together…. … and actually write 🤯 thanks to anhtho 🍊 for being our 3rd Writing Mentor! allen helped with AV and will upload both her and Corey Quinn video soon

swyx

14,508 次观看 • 5 个月前

I got curious how compaction works as a PM, so I did some brain surgery on Claude Code: (Anthropic's been doing really interesting work on context editing - they showed Claude Opus playing Settlers of Catan for 75+ minutes in a single thread by constantly editing the context instead of starting fresh. When I saw that Claude Code has a compaction command with optional custom instructions, I wanted to understand what's actually happening.) Abhishek Katiyar and Aman Khan gave me the key tip: Claude Code stores all your conversation history as text files on your computer. Open a new directory and give Claude Code a task. Here's how to watch compaction happening: 1. Go to your user's root directory 2. Press Command+Shift+Period (Mac) to show hidden folders 3. Navigate to ~/.claude/projects/ 4. Find your project folder and use Cursor/VSCode to open it (there's a reason) 5. Install the JSONL Gazelle plugin (open source, thank you Gabor Cselle!) 6. Open the most recent JSONL file - each row is a message in your conversation 7. Run the compact command in Claude Code with custom instructions 8. Watch what happens in the file What I learned: When you compact, Claude Code doesn't just summarize and delete everything. It creates a "compact boundary" in the conversation file, writes a summary of what happened before, but keeps the full original conversation (!!!!) The new thread can still retrieve any details from before compaction if needed. That is so damn cool. Why this matters: What you're getting in Claude Code is similar to what Anthropic ships in their developer SDK - so inspecting your daily tools is how you build real product intuition. The best way to understand AI systems is to open them up and look inside. Everything is text files.

I got curious how compaction works as a PM, so I did some brain surgery on Claude Code: (Anthropic's been doing really interesting work on context editing - they showed Claude Opus playing Settlers of Catan for 75+ minutes in a single thread by constantly editing the context instead of starting fresh. When I saw that Claude Code has a compaction command with optional custom instructions, I wanted to understand what's actually happening.) Abhishek Katiyar and Aman Khan gave me the key tip: Claude Code stores all your conversation history as text files on your computer. Open a new directory and give Claude Code a task. Here's how to watch compaction happening: 1. Go to your user's root directory 2. Press Command+Shift+Period (Mac) to show hidden folders 3. Navigate to ~/.claude/projects/ 4. Find your project folder and use Cursor/VSCode to open it (there's a reason) 5. Install the JSONL Gazelle plugin (open source, thank you Gabor Cselle!) 6. Open the most recent JSONL file - each row is a message in your conversation 7. Run the compact command in Claude Code with custom instructions 8. Watch what happens in the file What I learned: When you compact, Claude Code doesn't just summarize and delete everything. It creates a "compact boundary" in the conversation file, writes a summary of what happened before, but keeps the full original conversation (!!!!) The new thread can still retrieve any details from before compaction if needed. That is so damn cool. Why this matters: What you're getting in Claude Code is similar to what Anthropic ships in their developer SDK - so inspecting your daily tools is how you build real product intuition. The best way to understand AI systems is to open them up and look inside. Everything is text files.

Tal Raviv

57,910 次观看 • 8 个月前

Why haven't AI agents gone mainstream yet? We asked Jean-Denis Greze 💡, co-founder of Town, to break down what's missing. "OpenClaw is fantastic, but think about what it asks you to do. Get this open source software working on your machine. Connect MCPs to it. Write prompts or workflows. It can break all the time. A normal person is gonna do none of those things." "No one has packaged what a power user gets out of OpenClaw and made that available for everybody." "Before the iPhone there were other very powerful devices. They just didn't have the whole package, you went from single digit millions to billions of users once someone figured out how to make a consumer grade version." "Products today still ask you to be creative about what you wanna do with AI. People don't know what to put in that box, let's reverse the relationship. Let AI figure out what you're already doing and automate that."

Why haven't AI agents gone mainstream yet? We asked Jean-Denis Greze 💡, co-founder of Town, to break down what's missing. "OpenClaw is fantastic, but think about what it asks you to do. Get this open source software working on your machine. Connect MCPs to it. Write prompts or workflows. It can break all the time. A normal person is gonna do none of those things." "No one has packaged what a power user gets out of OpenClaw and made that available for everybody." "Before the iPhone there were other very powerful devices. They just didn't have the whole package, you went from single digit millions to billions of users once someone figured out how to make a consumer grade version." "Products today still ask you to be creative about what you wanna do with AI. People don't know what to put in that box, let's reverse the relationship. Let AI figure out what you're already doing and automate that."

MTS

11,713 次观看 • 3 个月前

Broccoli Quanto Tutorial Video 📽️ 🥦Stack Broccoli (0x343E66bAD706EED4350fE96385ffc6336bd653Ef) and use it as collateral to long/short BNB with leverage: 💡If you’re trading Broccoli Quanto on mobile devices, make sure to copy and paste the in your wallet browser instead of chrome/safari. 📖If you’re new to quanto trading👇 -What is quanto perpetual trading? Quanto perpetual contracts let you trade assets (like BNB) with leverage without holding them, instead committing a collateral asset (like Broccoli #3Ef). Profits settle in the collateral asset (like Broccoli 3Ef). This means you can speculate on BNB’s price movements by committing Broccoli 3Ef as collateral asset to long/short BNB from a leveraged position. If BNB price goes up, you’ll gain multiplied Broccoli 3Ef. With higher leverage, your gains multiply more —but risk liquidation if the market moves against you. -Position limit on Broccoli Quanto When trading, ensure that your total position size after applying your customized leverage does not exceed 20% of the total market liquidity. Please review available liquidity before placing orders to maintain seamless trading operations.

Broccoli Quanto Tutorial Video 📽️ 🥦Stack Broccoli (0x343E66bAD706EED4350fE96385ffc6336bd653Ef) and use it as collateral to long/short BNB with leverage: 💡If you’re trading Broccoli Quanto on mobile devices, make sure to copy and paste the in your wallet browser instead of chrome/safari. 📖If you’re new to quanto trading👇 -What is quanto perpetual trading? Quanto perpetual contracts let you trade assets (like BNB) with leverage without holding them, instead committing a collateral asset (like Broccoli #3Ef). Profits settle in the collateral asset (like Broccoli 3Ef). This means you can speculate on BNB’s price movements by committing Broccoli 3Ef as collateral asset to long/short BNB from a leveraged position. If BNB price goes up, you’ll gain multiplied Broccoli 3Ef. With higher leverage, your gains multiply more —but risk liquidation if the market moves against you. -Position limit on Broccoli Quanto When trading, ensure that your total position size after applying your customized leverage does not exceed 20% of the total market liquidity. Please review available liquidity before placing orders to maintain seamless trading operations.

Bounce Brand

22,672 次观看 • 1 年前

Introducing my newest app, TethrX, made with Grok 4.5 to access Grok Build on your phone. TethrX connects to Grok Build running on your own computer, so you can start a task from anywhere, follow Grok's reasoning as it works, approve every command before it runs, and review the code it writes. Your code never leaves your machine. The public TestFlight is now open, and a demonstration is below. You pair your phone by scanning a QR code, either on your local network or from anywhere through Tailscale. From there TethrX streams Grok's reasoning, tool calls, command output and file changes as they happen, and asks your approval before anything runs. Plan mode lets you read the plan before the work begins. When a task finishes you can review exactly what changed. TethrX lists the modified files in your project, shows a diff for each one, and lets you commit or discard the work without leaving your phone. The app supports slash commands, including /compact and any skills you have installed, along with voice dictation, queued messages and reusable prompts. Sessions can be searched and organised into folders, and you can pair several computers and switch between them. Siri can start a task or tell you what Grok is doing without opening the app, a home screen widget shows whether Grok is working, and a Live Activity tracks progress on your lock screen and Dynamic Island. Every session reports its context window, token usage and cost, the app can be locked behind Face ID, and your computer is kept awake for as long as a task is running. TethrX requires Grok Build installed and signed in on your computer, together with Node.js 20 or newer. A single command starts the bridge: npx tethrx-bridge TethrX is open source under the Apache License 2.0. Both the iOS client and the local bridge are available here: Grok 4.5 helped a lot, thanks to SpaceXAI for making Grok 4.5 exceptional.

Introducing my newest app, TethrX, made with Grok 4.5 to access Grok Build on your phone. TethrX connects to Grok Build running on your own computer, so you can start a task from anywhere, follow Grok's reasoning as it works, approve every command before it runs, and review the code it writes. Your code never leaves your machine. The public TestFlight is now open, and a demonstration is below. You pair your phone by scanning a QR code, either on your local network or from anywhere through Tailscale. From there TethrX streams Grok's reasoning, tool calls, command output and file changes as they happen, and asks your approval before anything runs. Plan mode lets you read the plan before the work begins. When a task finishes you can review exactly what changed. TethrX lists the modified files in your project, shows a diff for each one, and lets you commit or discard the work without leaving your phone. The app supports slash commands, including /compact and any skills you have installed, along with voice dictation, queued messages and reusable prompts. Sessions can be searched and organised into folders, and you can pair several computers and switch between them. Siri can start a task or tell you what Grok is doing without opening the app, a home screen widget shows whether Grok is working, and a Live Activity tracks progress on your lock screen and Dynamic Island. Every session reports its context window, token usage and cost, the app can be locked behind Face ID, and your computer is kept awake for as long as a task is running. TethrX requires Grok Build installed and signed in on your computer, together with Node.js 20 or newer. A single command starts the bridge: npx tethrx-bridge TethrX is open source under the Apache License 2.0. Both the iOS client and the local bridge are available here: Grok 4.5 helped a lot, thanks to SpaceXAI for making Grok 4.5 exceptional.

Myrhe𝕩

13,099 次观看 • 2 个月前

Q: What's better than Mixture of Experts? A: Mixture of Mixture of Experts! Introducing GodMode: the AI Chat Browser Fast, Free, Access to ChatGPT, Bing, Bard, Claude, YouChat, Poe, Perplexity, Phind, and Local/GGML Models like Vicuna and Alpaca No API keys, Full Webapp Features We've been heads down rewriting the Smol AI menubar into a full fledged "AI Chat Browser" this past month. Our sincere belief is that every chat in the world should be one keyboard shortcut away and run simultaneously to tap into the global superintelligence. And ofc, free and open source. Releasing this as a "beta" today because a bunch of polish features aren't yet done, but we want to get something out there for feedback. Roadmap: - model voting -> global/personal leaderboard - add even more models, eg Inflection - PromptCritic - AI assisted prompt improvement Heavy thanks to Sean Oliver and Alec Dewitz for working with me this past month on this and our upcoming Super Secret next project (see if you can find it in the menu!)

Q: What's better than Mixture of Experts? A: Mixture of Mixture of Experts! Introducing GodMode: the AI Chat Browser Fast, Free, Access to ChatGPT, Bing, Bard, Claude, YouChat, Poe, Perplexity, Phind, and Local/GGML Models like Vicuna and Alpaca No API keys, Full Webapp Features We've been heads down rewriting the Smol AI menubar into a full fledged "AI Chat Browser" this past month. Our sincere belief is that every chat in the world should be one keyboard shortcut away and run simultaneously to tap into the global superintelligence. And ofc, free and open source. Releasing this as a "beta" today because a bunch of polish features aren't yet done, but we want to get something out there for feedback. Roadmap: - model voting -> global/personal leaderboard - add even more models, eg Inflection - PromptCritic - AI assisted prompt improvement Heavy thanks to Sean Oliver and Alec Dewitz for working with me this past month on this and our upcoming Super Secret next project (see if you can find it in the menu!)

swyx

341,292 次观看 • 3 年前

For the first time we are fundamentally changing how humans can collaborate with ChatGPT since it launched two years ago. We’re introducing canvas, a new interface for working with ChatGPT on writing and coding projects that go beyond simple chat. Product and model features: 1/ Ask for in-line feedback. With canvas, ChatGPT can better understand the context of what you’re trying to accomplish. You can highlight specific sections to indicate exactly what you want ChatGPT to focus on. Like a copy editor or code reviewer, it can give in-line feedback and suggestions with the entire project in mind. 2/ Directly edit the model's output and select a specific area for targeted editing. You control your creative work on canvas. You can directly edit text or code. 3/ Menu of shortcuts. There’s a menu of shortcuts for you to ask ChatGPT to adjust writing length, debug your code, and quickly perform other useful actions. You can also restore previous versions of your work by using the back button in canvas. 4/ Use search with canvas for research writing! As we are moving towards the new paradigm of reasoning we are fundamentally evolving the chat interface into a more collaborative human-AI interaction. Today you can say “browse / use browsing to find XYZ on the internet and write a report in canvas”

For the first time we are fundamentally changing how humans can collaborate with ChatGPT since it launched two years ago. We’re introducing canvas, a new interface for working with ChatGPT on writing and coding projects that go beyond simple chat. Product and model features: 1/ Ask for in-line feedback. With canvas, ChatGPT can better understand the context of what you’re trying to accomplish. You can highlight specific sections to indicate exactly what you want ChatGPT to focus on. Like a copy editor or code reviewer, it can give in-line feedback and suggestions with the entire project in mind. 2/ Directly edit the model's output and select a specific area for targeted editing. You control your creative work on canvas. You can directly edit text or code. 3/ Menu of shortcuts. There’s a menu of shortcuts for you to ask ChatGPT to adjust writing length, debug your code, and quickly perform other useful actions. You can also restore previous versions of your work by using the back button in canvas. 4/ Use search with canvas for research writing! As we are moving towards the new paradigm of reasoning we are fundamentally evolving the chat interface into a more collaborative human-AI interaction. Today you can say “browse / use browsing to find XYZ on the internet and write a report in canvas”

Karina

845,887 次观看 • 2 年前

THIS GUY BUILT AN AUTONOMOUS AI AGENT OUT OF CLAUDE CODE + OBSIDIAN and this is way more interesting than another “use AI to take notes” demo the trick is simple: Obsidian is not the writing app here. it becomes the agent’s memory, task board, and context folder. Claude Code is not just answering prompts. it reads the vault, edits files, follows instructions, and keeps moving through the work like a junior operator with a filesystem. the reusable setup looks like this: 1. create an Obsidian vault for one project 2. keep goals, rules, tasks, decisions, and references as markdown files 3. point Claude Code at the folder 4. give it a clear operating loop: read context → choose next task → execute → write back what changed 5. use the notes as persistent memory instead of re-explaining the project every chat that’s the part people miss. the “agent” is not magic. it’s the boring combination of: - local files - explicit rules - task state - write access - a model that can run through the repo/vault Obsidian makes the memory human-readable. Claude Code makes the memory executable. that combo is why the video worked: it turns a notes app into an operating surface for actual work. best use cases: - content systems - research vaults - coding projects - client ops docs - personal knowledge bases that need actions, not just storage the caveat: if your vault is messy, your agent becomes messy too. folders, naming, “done” criteria, and forbidden actions matter more than the prompt. but once the structure is clean, this is one of the easiest ways to build an agent that remembers what happened yesterday without paying for a full custom app.

THIS GUY BUILT AN AUTONOMOUS AI AGENT OUT OF CLAUDE CODE + OBSIDIAN and this is way more interesting than another “use AI to take notes” demo the trick is simple: Obsidian is not the writing app here. it becomes the agent’s memory, task board, and context folder. Claude Code is not just answering prompts. it reads the vault, edits files, follows instructions, and keeps moving through the work like a junior operator with a filesystem. the reusable setup looks like this: 1. create an Obsidian vault for one project 2. keep goals, rules, tasks, decisions, and references as markdown files 3. point Claude Code at the folder 4. give it a clear operating loop: read context → choose next task → execute → write back what changed 5. use the notes as persistent memory instead of re-explaining the project every chat that’s the part people miss. the “agent” is not magic. it’s the boring combination of: - local files - explicit rules - task state - write access - a model that can run through the repo/vault Obsidian makes the memory human-readable. Claude Code makes the memory executable. that combo is why the video worked: it turns a notes app into an operating surface for actual work. best use cases: - content systems - research vaults - coding projects - client ops docs - personal knowledge bases that need actions, not just storage the caveat: if your vault is messy, your agent becomes messy too. folders, naming, “done” criteria, and forbidden actions matter more than the prompt. but once the structure is clean, this is one of the easiest ways to build an agent that remembers what happened yesterday without paying for a full custom app.

kocer

30,403 次观看 • 2 个月前

THIS GUY TURNED 5 PROMPTING TIPS INTO A FREE AI CEO CHALLENGE The useful part is treating every prompt like you are briefing a very fast employee who has zero context. Most people open ChatGPT and type a wish. Pros give it a job. Try this instead: 1. Give it a role Not “help me with marketing.” Say: “Act as a B2B SaaS growth operator reviewing a landing page.” 2. Give it the real context Who is the customer? What are they buying? What have you already tried? What does success look like? 3. Give it constraints Length, tone, format, audience, banned words, examples to copy, examples to avoid. A vague prompt gets a vague answer. A constrained prompt gets something you can edit. 4. Ask for options before answers “Give me 5 angles, rank them, then explain the tradeoff.” This turns AI from an autocomplete box into a thinking partner. 5. Force it to show assumptions Before it writes, ask: “What are you assuming, what info is missing, and what would change your answer?” That one line saves a lot of fake confidence. Dan Martell’s video works because the promise is simple: 5 prompting habits that make AI feel less random. The reusable move is even simpler: Stop prompting for outputs. Start prompting for decisions. Bad: “Write me a post.” Better: “Here is the source, here is the reader, here is the angle, give me 3 hooks, choose the strongest, then draft in this style.” That is the difference between getting content-shaped noise and getting work you can actually ship. Caveat: prompts do not fix weak taste, bad data, or unclear strategy. But they do expose those problems faster. If your AI answers are generic, your prompt probably has no job, no context, no constraints, and no standard for what “good” means.

THIS GUY TURNED 5 PROMPTING TIPS INTO A FREE AI CEO CHALLENGE The useful part is treating every prompt like you are briefing a very fast employee who has zero context. Most people open ChatGPT and type a wish. Pros give it a job. Try this instead: 1. Give it a role Not “help me with marketing.” Say: “Act as a B2B SaaS growth operator reviewing a landing page.” 2. Give it the real context Who is the customer? What are they buying? What have you already tried? What does success look like? 3. Give it constraints Length, tone, format, audience, banned words, examples to copy, examples to avoid. A vague prompt gets a vague answer. A constrained prompt gets something you can edit. 4. Ask for options before answers “Give me 5 angles, rank them, then explain the tradeoff.” This turns AI from an autocomplete box into a thinking partner. 5. Force it to show assumptions Before it writes, ask: “What are you assuming, what info is missing, and what would change your answer?” That one line saves a lot of fake confidence. Dan Martell’s video works because the promise is simple: 5 prompting habits that make AI feel less random. The reusable move is even simpler: Stop prompting for outputs. Start prompting for decisions. Bad: “Write me a post.” Better: “Here is the source, here is the reader, here is the angle, give me 3 hooks, choose the strongest, then draft in this style.” That is the difference between getting content-shaped noise and getting work you can actually ship. Caveat: prompts do not fix weak taste, bad data, or unclear strategy. But they do expose those problems faster. If your AI answers are generic, your prompt probably has no job, no context, no constraints, and no standard for what “good” means.

kocer

25,573 次观看 • 2 个月前

PhD Students – How to automatically identify 90% of the issues in your research paper before you submit it to a journal? This is possible through manual or automated paper review. First, let’s understand the following. 𝐖𝐡𝐚𝐭 𝐢𝐬 𝐚 𝐩𝐚𝐩𝐞𝐫 𝐫𝐞𝐯𝐢𝐞𝐰? Paper review is a process in which subject matter experts evaluate your paper based on the following criteria: 1. Significance – Is this research important? 2. Novelty – Is this research new? 3. Methodology – Is this research carried out in the correct way? 4. Verifiability – Can other researchers verify this research? 5. Presentation – Is the research presented in the right way? 𝐖𝐡𝐲 𝐭𝐨 𝐡𝐚𝐯𝐞 𝐲𝐨𝐮𝐫 𝐩𝐚𝐩𝐞𝐫 𝐫𝐞𝐯𝐢𝐞𝐰𝐞𝐝 𝐛𝐞𝐟𝐨𝐫𝐞 𝐬𝐮𝐛𝐦𝐢𝐬𝐬𝐢𝐨𝐧? ➟ Identify the critical issues in your paper ➟ Fix those issues to increase the chances of your paper acceptance 𝐇𝐨𝐰 𝐭𝐨 𝐚𝐮𝐭𝐨𝐦𝐚𝐭𝐞 “𝐬𝐞𝐥𝐟-𝐫𝐞𝐯𝐢𝐞𝐰” 𝐨𝐟 𝐲𝐨𝐮𝐫 𝐩𝐚𝐩𝐞𝐫? Paperpal just launched an amazing feature – AI Review. Using this feature, you can get instant self-feedback. This feature will help you in the following ways. ➝ Check for gaps in your logic ➝ Get feedback on the structure and flow of your writing ➝ Review your research questions ➝ Identify opportunities to strengthen your paper ➝ Increase the chances of your paper acceptance Here is a step-by-step process for using AI Review feature. Step 1: Go to and login. Step 2: Open an existing document or make a new document Step 3: Go to the right-side bar and click on checks | AI Review. Step 4: For this feature to work there should be more than 150 words. Step 5: Copy and paste your paper. Step 6: Now go to the right side and check the prompts Step 7: With these prompts, you will evaluate your paper. Step 8: You will find various prompts e.g., suggest writing feedback, check flow and structure etc. Step 9: You can select a prompt from the existing prompts or write your custom prompt and execute Step 10: Paperpal will generate feedback as per the prompt. Step 11: Read through the feedback and save it for further use. Use other specific prompts for tailored feedback. Step 12: This way you can evaluate various aspects of your paper yourself. This is a very customized and efficient way of automatically reviewing your paper. You can also go one step further to work on the feedback and improve your paper based on suggestions. Please note that AI Review feature does not replace human or expert reviewers in any way. This feature only aims to provide you with quick self-feedback. Try the AI Review feature of Paperpal. Paperpal link:

PhD Students – How to automatically identify 90% of the issues in your research paper before you submit it to a journal? This is possible through manual or automated paper review. First, let’s understand the following. 𝐖𝐡𝐚𝐭 𝐢𝐬 𝐚 𝐩𝐚𝐩𝐞𝐫 𝐫𝐞𝐯𝐢𝐞𝐰? Paper review is a process in which subject matter experts evaluate your paper based on the following criteria: 1. Significance – Is this research important? 2. Novelty – Is this research new? 3. Methodology – Is this research carried out in the correct way? 4. Verifiability – Can other researchers verify this research? 5. Presentation – Is the research presented in the right way? 𝐖𝐡𝐲 𝐭𝐨 𝐡𝐚𝐯𝐞 𝐲𝐨𝐮𝐫 𝐩𝐚𝐩𝐞𝐫 𝐫𝐞𝐯𝐢𝐞𝐰𝐞𝐝 𝐛𝐞𝐟𝐨𝐫𝐞 𝐬𝐮𝐛𝐦𝐢𝐬𝐬𝐢𝐨𝐧? ➟ Identify the critical issues in your paper ➟ Fix those issues to increase the chances of your paper acceptance 𝐇𝐨𝐰 𝐭𝐨 𝐚𝐮𝐭𝐨𝐦𝐚𝐭𝐞 “𝐬𝐞𝐥𝐟-𝐫𝐞𝐯𝐢𝐞𝐰” 𝐨𝐟 𝐲𝐨𝐮𝐫 𝐩𝐚𝐩𝐞𝐫? Paperpal just launched an amazing feature – AI Review. Using this feature, you can get instant self-feedback. This feature will help you in the following ways. ➝ Check for gaps in your logic ➝ Get feedback on the structure and flow of your writing ➝ Review your research questions ➝ Identify opportunities to strengthen your paper ➝ Increase the chances of your paper acceptance Here is a step-by-step process for using AI Review feature. Step 1: Go to and login. Step 2: Open an existing document or make a new document Step 3: Go to the right-side bar and click on checks | AI Review. Step 4: For this feature to work there should be more than 150 words. Step 5: Copy and paste your paper. Step 6: Now go to the right side and check the prompts Step 7: With these prompts, you will evaluate your paper. Step 8: You will find various prompts e.g., suggest writing feedback, check flow and structure etc. Step 9: You can select a prompt from the existing prompts or write your custom prompt and execute Step 10: Paperpal will generate feedback as per the prompt. Step 11: Read through the feedback and save it for further use. Use other specific prompts for tailored feedback. Step 12: This way you can evaluate various aspects of your paper yourself. This is a very customized and efficient way of automatically reviewing your paper. You can also go one step further to work on the feedback and improve your paper based on suggestions. Please note that AI Review feature does not replace human or expert reviewers in any way. This feature only aims to provide you with quick self-feedback. Try the AI Review feature of Paperpal. Paperpal link:

Faheem Ullah

15,270 次观看 • 1 年前

ChatGPT casually dropped an APP STORE 🤯 It can now: - browse the web (RIP Bing waitlist, cutoff) - write and run Python (RIP replit?) - access org info (RIP docsearch startups) - add third party plugins from OpenTable, Wolfram, Instacart, Zapier, etc) - developer SDK in preview

ChatGPT casually dropped an APP STORE 🤯 It can now: - browse the web (RIP Bing waitlist, cutoff) - write and run Python (RIP replit?) - access org info (RIP docsearch startups) - add third party plugins from OpenTable, Wolfram, Instacart, Zapier, etc) - developer SDK in preview

swyx

1,859,493 次观看 • 3 年前

new banner pic :) what an incredible year for AI Engineer. 3 conferences, 7.5m views, and hundreds of jobs/ideas shared. Could not have done without Benjamin Dunphy, Lia McBride, and the rest of our Tiny Team. special thanks to Alex Lieberman and Jed Borovik for stepping up as the AIE CODE emcees!! Having track hosts and emcees from the community is what makes AIE a true industry conference, by AI Engineers for AI Engineers. Googlers introducing Anthropics, Consultants introducing competitors. Everyone just wants to win on merit and hire/market/ learn/take the pulse of the industry. I hope to always keep it that way.

new banner pic :) what an incredible year for AI Engineer. 3 conferences, 7.5m views, and hundreds of jobs/ideas shared. Could not have done without Benjamin Dunphy, Lia McBride, and the rest of our Tiny Team. special thanks to Alex Lieberman and Jed Borovik for stepping up as the AIE CODE emcees!! Having track hosts and emcees from the community is what makes AIE a true industry conference, by AI Engineers for AI Engineers. Googlers introducing Anthropics, Consultants introducing competitors. Everyone just wants to win on merit and hire/market/ learn/take the pulse of the industry. I hope to always keep it that way.

swyx

11,613 次观看 • 10 个月前

Skills are the quickest way to 10x the quality and consistency of what you get from Claude Code. And you don't need to be a developer to use them. Anthropic just published how they use hundreds of skills internally every day. Most skill tutorials are made for developers — if you're in marketing, sales, content ops, or GTM, you probably watched those and moved on. But skills are just as important for non-developers. A skill is just a reusable prompt with clear instructions for a specific task. Instead of prompting Claude the same way over and over, you build it once and invoke it every time. I have a skill for writing on LinkedIn. A different one for YouTube outlines. Another for X. Each platform has different rules, different voice, different structure — so each one gets its own skill. If you're doing something repeatedly, it's time to make a skill. The biggest mistake most people make: building skills as a single .md file. A single file dumps everything into context whether Claude needs it or not. Wastes tokens. Gets worse results. Skills should be folders. Here's the structure that works: skill.md — the orchestrator. Tells Claude which files to read and when. It doesn't contain rules itself — it's the playbook. instructions/ — separate files for voice, structure, scope. Claude only loads the one it needs for the current step. examples/ — good AND bad. Good examples show what success looks like. Bad examples show patterns to avoid — AI writing tells, weak hooks, generic CTAs. Most people skip bad examples. Don't. eval/ — a checklist that scores every output before you see it. "Does it have a clear hook?" "Is it free of AI buzzwords?" Pass or fail on each item. templates/ — output formatting so you get consistent structure every time. The three types of skills that matter most for non-developers: 1. Business automation. Writing a newsletter. Checking reports and drafting follow-ups. Running programmatic ad campaigns. Any workflow you repeat — build a skill for it. 2. Content templates. Landing page copy, meta ads, email sequences, SEO briefs. Each one has specific requirements. Each one gets its own skill. 3. Thinking partners. This is the one people miss. Skills don't have to produce output. They can help you think — an advisory board that reviews your work from your ICP's perspective, a coach that pressure-tests your strategy, an ideation partner that researches competitors before suggesting your next move. If you already have skills as .md files, here's the exact prompt to restructure them in the Anthropic approved format: "I want to restructure my Claude Code skill file. Right now my skill is a single .md file and I want to break it into a folder system following Anthropic's best practices. Read my current skill file, then restructure it into a folder with: a skill.md orchestrator, an instructions/ folder with separate files for each concern (voice, structure, scope), an examples/ folder with good and bad examples, an eval/ folder with a quality checklist, and a templates/ folder for output formatting. Keep all my existing rules and intent — just reorganize them into the modular structure." Paste that into Claude Code pointed at the folder where your skill lives. It handles the rest. A few caveats: 1. Don't add too many skills. Every skill adds context Claude has to process. 50 skills loaded means everything slows down. Start with 3-5 covering your most repeated workflows. 2. Vet skills before downloading. If you grab a skill from the internet, read what's inside first. Skills can include shell commands and scripts. Check what you're running. 3. Share what works. Build a skill that performs well, put it in a shared GitHub repo. Your marketing org gets shared skills for copywriting, SEO, ad copy — new hires invoke the skill instead of learning every playbook from scratch. Onboarding time drops dramatically. 4. Keep your skills updated. When you see output you love, add it as a good example. When you see a pattern you hate, add it as a bad example. The skill gets sharper every time. I made a full video walking through all of this — including a live build of two skills from scratch (no terminal, no code), the exact prompt I use to restructure old skills, and 5 pro tips from Anthropic's internal playbook. Share this with your non-developer friends that want to do more with AI; or bookmark it to come back to at a later time.

Skills are the quickest way to 10x the quality and consistency of what you get from Claude Code. And you don't need to be a developer to use them. Anthropic just published how they use hundreds of skills internally every day. Most skill tutorials are made for developers — if you're in marketing, sales, content ops, or GTM, you probably watched those and moved on. But skills are just as important for non-developers. A skill is just a reusable prompt with clear instructions for a specific task. Instead of prompting Claude the same way over and over, you build it once and invoke it every time. I have a skill for writing on LinkedIn. A different one for YouTube outlines. Another for X. Each platform has different rules, different voice, different structure — so each one gets its own skill. If you're doing something repeatedly, it's time to make a skill. The biggest mistake most people make: building skills as a single .md file. A single file dumps everything into context whether Claude needs it or not. Wastes tokens. Gets worse results. Skills should be folders. Here's the structure that works: skill.md — the orchestrator. Tells Claude which files to read and when. It doesn't contain rules itself — it's the playbook. instructions/ — separate files for voice, structure, scope. Claude only loads the one it needs for the current step. examples/ — good AND bad. Good examples show what success looks like. Bad examples show patterns to avoid — AI writing tells, weak hooks, generic CTAs. Most people skip bad examples. Don't. eval/ — a checklist that scores every output before you see it. "Does it have a clear hook?" "Is it free of AI buzzwords?" Pass or fail on each item. templates/ — output formatting so you get consistent structure every time. The three types of skills that matter most for non-developers: 1. Business automation. Writing a newsletter. Checking reports and drafting follow-ups. Running programmatic ad campaigns. Any workflow you repeat — build a skill for it. 2. Content templates. Landing page copy, meta ads, email sequences, SEO briefs. Each one has specific requirements. Each one gets its own skill. 3. Thinking partners. This is the one people miss. Skills don't have to produce output. They can help you think — an advisory board that reviews your work from your ICP's perspective, a coach that pressure-tests your strategy, an ideation partner that researches competitors before suggesting your next move. If you already have skills as .md files, here's the exact prompt to restructure them in the Anthropic approved format: "I want to restructure my Claude Code skill file. Right now my skill is a single .md file and I want to break it into a folder system following Anthropic's best practices. Read my current skill file, then restructure it into a folder with: a skill.md orchestrator, an instructions/ folder with separate files for each concern (voice, structure, scope), an examples/ folder with good and bad examples, an eval/ folder with a quality checklist, and a templates/ folder for output formatting. Keep all my existing rules and intent — just reorganize them into the modular structure." Paste that into Claude Code pointed at the folder where your skill lives. It handles the rest. A few caveats: 1. Don't add too many skills. Every skill adds context Claude has to process. 50 skills loaded means everything slows down. Start with 3-5 covering your most repeated workflows. 2. Vet skills before downloading. If you grab a skill from the internet, read what's inside first. Skills can include shell commands and scripts. Check what you're running. 3. Share what works. Build a skill that performs well, put it in a shared GitHub repo. Your marketing org gets shared skills for copywriting, SEO, ad copy — new hires invoke the skill instead of learning every playbook from scratch. Onboarding time drops dramatically. 4. Keep your skills updated. When you see output you love, add it as a good example. When you see a pattern you hate, add it as a bad example. The skill gets sharper every time. I made a full video walking through all of this — including a live build of two skills from scratch (no terminal, no code), the exact prompt I use to restructure old skills, and 5 pro tips from Anthropic's internal playbook. Share this with your non-developer friends that want to do more with AI; or bookmark it to come back to at a later time.

JJ Englert

29,322 次观看 • 5 个月前

🤯 with OpenAI Whisper + TTS + GPT4V, you can just permanently share your screen with an AI and talk to it to get help on anything on screen. a generally intelligent assistant for your digital life. with a bit more engineering you could probably feed it extra docs and context based on the currently active tab/app.

🤯 with OpenAI Whisper + TTS + GPT4V, you can just permanently share your screen with an AI and talk to it to get help on anything on screen. a generally intelligent assistant for your digital life. with a bit more engineering you could probably feed it extra docs and context based on the currently active tab/app.

swyx

1,051,687 次观看 • 2 年前

Claymotion ads are crushing it on Meta right now. Built a free claude skill to make them 👇 If you've been scrolling Meta lately, you've seen them — stop-motion clay characters, tactile textures, weirdly satisfying to watch. CTRs are 2-3x the feed average. Almost nobody is running them. The problem: they look impossible to make unless you have a studio. They're not. You just need the right prompts. So I packaged the prompt system as a Claude Code skill. It's free. Here's what it does: Paste your product URL. Out comes a full claymotion ad plan: 1/ Shot-by-shot storyboard 5-7 shots with the narrative arc. Setup → product reveal → payoff → CTA. 2/ Image prompt per shot Exact prompt you paste into Midjourney, Nano Banana, or any image gen. Camera angle, lighting, clay texture specs, character details — dialed in for consistency across shots. 3/ Video prompt per shot The animation prompt you paste into Kling, Veo, Seedance, or Sora. Motion direction, pacing, transitions — so the shots actually flow. 4/ VO script per shot Voiceover copy written for rhythm. Timed to the shot length. Hook, body, CTA — all on brand. 5/ Music + sfx direction Tone notes for the track. Specific sfx cues per shot (squish, pop, whoosh) You take the outputs. Paste them into your image + video generators. Stitch the shots. Record the VO. A full claymotion ad in under an hour, at the cost of a few API credits. Instead of $3,000 and 3 weeks with an animation studio. Why claymotion works right now: → Pattern break — nothing else in the feed looks like it → Tactile feel — clay reads as "real" even when AI-generated → High dwell time — people watch the whole thing → Cheap to test — 5-10 variations per product is now feasible Comment "Clay" and I'll send you: → The Claude Code skill (free) → A starter prompt pack → 3 example storyboards so you can see the output (must be connected)

Ahad Shams | AI Ads Guy

17,021 次观看 • 5 个月前

$🚨🌟 Watch Part 1 Of The EstateX Vlog Series! 🌟🚨 The wait is over, EstateX Family! We’re beyond excited to announce the release of the very first part of our brand-new EstateX Vlog Series, starring none other than Bart de Bruijn & Steve! 🎥🤴🤴 In this 5-part series, you’ll get a front-row seat to the world of real estate #tokenization, #blockchain, and the exciting EstateX journey. The first episode takes you behind the scenes in Dubai as Bart and Steve tour the luxurious Neptune floating villa—an exclusive property that can soon be fractionalized on the #EstateX platform! 🏝️ What to expect in Part 1: • A deep dive into a truly unique real estate opportunity—the Neptune yacht villa 🌊 • Insights from Bart & Steve on how EstateX is revolutionizing luxury real estate investments 🏘️ • A whole lot of fun, laughs, and inspiration along the way! 😄 💡 Why should you watch? This series isn’t just about cool properties—it’s about how EstateX is changing the game for real estate investing. You’ll see firsthand how we diversify our portfolio, offering unparalleled opportunities, and how we’re getting closer to our goal of making real estate investing as simple as online shopping. As you can see, we’re starting slowly our marketing campaign with KOLs posting and our promised high quality marketing content. 🗓️ More to come! This is just the beginning—4 more parts of this vlog series are coming your way! Stay tuned for exclusive content, behind-the-scenes moments, and sneak peeks that will keep you on the edge of your seat. Trust us, you won’t want to miss a single episode! 🎬 Watch Part 1 now, and stay tuned for more updates from the EstateX Vlog Series. Onwards and upwards, EstateX Fam! 🚀💫 #RWA$

🚨🌟 Watch Part 1 Of The EstateX Vlog Series! 🌟🚨 The wait is over, EstateX Family! We’re beyond excited to announce the release of the very first part of our brand-new EstateX Vlog Series, starring none other than Bart de Bruijn & Steve! 🎥🤴🤴 In this 5-part series, you’ll get a front-row seat to the world of real estate #tokenization, #blockchain, and the exciting EstateX journey. The first episode takes you behind the scenes in Dubai as Bart and Steve tour the luxurious Neptune floating villa—an exclusive property that can soon be fractionalized on the #EstateX platform! 🏝️ What to expect in Part 1: • A deep dive into a truly unique real estate opportunity—the Neptune yacht villa 🌊 • Insights from Bart & Steve on how EstateX is revolutionizing luxury real estate investments 🏘️ • A whole lot of fun, laughs, and inspiration along the way! 😄 💡 Why should you watch? This series isn’t just about cool properties—it’s about how EstateX is changing the game for real estate investing. You’ll see firsthand how we diversify our portfolio, offering unparalleled opportunities, and how we’re getting closer to our goal of making real estate investing as simple as online shopping. As you can see, we’re starting slowly our marketing campaign with KOLs posting and our promised high quality marketing content. 🗓️ More to come! This is just the beginning—4 more parts of this vlog series are coming your way! Stay tuned for exclusive content, behind-the-scenes moments, and sneak peeks that will keep you on the edge of your seat. Trust us, you won’t want to miss a single episode! 🎬 Watch Part 1 now, and stay tuned for more updates from the EstateX Vlog Series. Onwards and upwards, EstateX Fam! 🚀💫 #RWA

EstateX

108,612 次观看 • 1 年前

🤖🔬 Can AI actually do science end-to-end? 🧠📈 And how would we know when it matches, or surpasses, humans? ⚡🧪 AI is rapidly automating scientific discovery, but benchmarking full-cycle discovery, from 💡 ideation → 🧑‍💻 execution → 📊 conclusions, remains unsolved: 🧐🧐🧐 ❌🛠️ Open-ended discovery → manual validation (costly, unscalable) ❌📏 Metric-driven benchmarks (e.g., MLE-Bench) → convenient but narrow (is higher accuracy really enough?) ❌🤖⚖️ LLM-as-judge → useful, but fundamentally risky if used alone 🔥🚀 Introducing FIRE-Bench🔥: Fullcycle Insight Rediscovery Evaluation 👉🌐 📚✨ A benchmark that turns fresh, human-verified insights from recent 🏆 NeurIPS / ICLR / ICML papers into masked, end-to-end discovery challenges 🧩 🌍🔐 Constrained open-ended discovery–backed by ground truth. 📌 Key takeaways: 1⃣ 📖🧱 Reference-based evaluation still matters: constrained LLM judging helps, but human-grounded references remain essential until agents can consistently match human conclusions 2⃣ 🏆🧠 Expert-validated ground truth: all tasks come from recent NeurIPS / ICLR / ICML papers, with contamination carefully controlled 3⃣ 🔁🎭 Rediscovery, not reproduction: original 🧪 methods, 📊 experiments, 💻 implementations, and 📈 analyses are fully masked to create real discovery challenges 🔑 Key empirical findings: 💡 The "Science Gap" is Real: Even the best setup (Claude Code + Sonnet-4) caps out at an F1 score of 46.7. On hard tasks, agents struggle to break 30 💡 Success is a "Lottery": Performance has incredibly high variance. Reliability is a major unsolved issue. 💡 Coding is no longer the bottleneck; high-level reasoning and analysis are: ~74% of errors stem from flawed planning, not coding ⚙️ How it works: 🔹 Research-Problem Trees: We parse papers into trees (from broad roots to concrete leaves). This allows us to select intermediate nodes that perfectly balance open-ended exploration with verifiable ground truth. 🔹 Claim-Level Evaluation: We match AI conclusions against human conclusions using granular claim decomposition (F1 score). 🔹 Creativity Check: We score false positives to see if agents are finding novel truths (Spoiler🚨: they aren’t creative yet). 🔹 New Diagnostic Taxonomy: failures traced across four stages: 🧠 Planning → 🛠️ Implementation → ▶️ Execution → 🧾 Conclusion 🔹 Additional Analyses: cost efficiency, contamination checks, and more. 👀 The Future: 🚀 Live-FIRE-Bench: a live, continuously updated FIRE-Bench to track real-time progress on the latest research (Newest LLMs should be benchmarked with the newest research) 🚀 Stronger scaffolding (search + planning + coding) 🧠🧰 and converting FIRE-Bench into interactive environments for training research agents 🚀 Toward real creativity: We want better systems that can produce genuinely novel conclusions toward creativity 🎨⏳ 🚀 Better systems 🧠✨ and better benchmarks 📏 must co-evolve 🔄 over time 📜🎥 Paper, video, demo, and research trees: 👉🌐 #AI 🤖 #MachineLearning 📚 #AI4Science 🔬 #LLMs 🧠 #Research 🧪 #AgenticAI 🚀 #FireBench 🔥

🤖🔬 Can AI actually do science end-to-end? 🧠📈 And how would we know when it matches, or surpasses, humans? ⚡🧪 AI is rapidly automating scientific discovery, but benchmarking full-cycle discovery, from 💡 ideation → 🧑‍💻 execution → 📊 conclusions, remains unsolved: 🧐🧐🧐 ❌🛠️ Open-ended discovery → manual validation (costly, unscalable) ❌📏 Metric-driven benchmarks (e.g., MLE-Bench) → convenient but narrow (is higher accuracy really enough?) ❌🤖⚖️ LLM-as-judge → useful, but fundamentally risky if used alone 🔥🚀 Introducing FIRE-Bench🔥: Fullcycle Insight Rediscovery Evaluation 👉🌐 📚✨ A benchmark that turns fresh, human-verified insights from recent 🏆 NeurIPS / ICLR / ICML papers into masked, end-to-end discovery challenges 🧩 🌍🔐 Constrained open-ended discovery–backed by ground truth. 📌 Key takeaways: 1⃣ 📖🧱 Reference-based evaluation still matters: constrained LLM judging helps, but human-grounded references remain essential until agents can consistently match human conclusions 2⃣ 🏆🧠 Expert-validated ground truth: all tasks come from recent NeurIPS / ICLR / ICML papers, with contamination carefully controlled 3⃣ 🔁🎭 Rediscovery, not reproduction: original 🧪 methods, 📊 experiments, 💻 implementations, and 📈 analyses are fully masked to create real discovery challenges 🔑 Key empirical findings: 💡 The "Science Gap" is Real: Even the best setup (Claude Code + Sonnet-4) caps out at an F1 score of 46.7. On hard tasks, agents struggle to break 30 💡 Success is a "Lottery": Performance has incredibly high variance. Reliability is a major unsolved issue. 💡 Coding is no longer the bottleneck; high-level reasoning and analysis are: ~74% of errors stem from flawed planning, not coding ⚙️ How it works: 🔹 Research-Problem Trees: We parse papers into trees (from broad roots to concrete leaves). This allows us to select intermediate nodes that perfectly balance open-ended exploration with verifiable ground truth. 🔹 Claim-Level Evaluation: We match AI conclusions against human conclusions using granular claim decomposition (F1 score). 🔹 Creativity Check: We score false positives to see if agents are finding novel truths (Spoiler🚨: they aren’t creative yet). 🔹 New Diagnostic Taxonomy: failures traced across four stages: 🧠 Planning → 🛠️ Implementation → ▶️ Execution → 🧾 Conclusion 🔹 Additional Analyses: cost efficiency, contamination checks, and more. 👀 The Future: 🚀 Live-FIRE-Bench: a live, continuously updated FIRE-Bench to track real-time progress on the latest research (Newest LLMs should be benchmarked with the newest research) 🚀 Stronger scaffolding (search + planning + coding) 🧠🧰 and converting FIRE-Bench into interactive environments for training research agents 🚀 Toward real creativity: We want better systems that can produce genuinely novel conclusions toward creativity 🎨⏳ 🚀 Better systems 🧠✨ and better benchmarks 📏 must co-evolve 🔄 over time 📜🎥 Paper, video, demo, and research trees: 👉🌐 #AI 🤖 #MachineLearning 📚 #AI4Science 🔬 #LLMs 🧠 #Research 🧪 #AgenticAI 🚀 #FireBench 🔥

Zhen Wang

18,565 次观看 • 7 个月前