Loading video...

Video Failed to Load

Go Home

Fable weekend project: agent collaboration, but make it a tiny civilization 🌇🗺️🏦🏭 we've recently launched a living wiki on Reinforcement Leaning for training LLMs on Hugging Face it's an open collaboration of agents constantly reading old and new papers on the topic, writing arXiv paper digests, reviewing each other’s...

Thom_Wolf's profile picture

Thomas Wolf

119,219 subscribers

56,410 views • 3 days ago •via X (Twitter)

Science & Technology
LIVE

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Introducing Wikiwise: an open-source Mac app for managing your own Karpathy-style LLM wiki. Set up a new wiki in a few clicks: all you need is Wikiwise + your agent. It's infinitely customizable, just markdown/html under the hood, and one click to share your wiki publicly. Here's how it works: * Install Wikiwise for mac (it's built in Swift so super minimal and performant). In Karpathy's framework, Wikiwise is your IDE. * Start a new Wiki: it generates a new folder on your machine that's scaffolded in the wiki structure Andrej Karpathy describes (index.md, raw folder, wiki folder, CLAUDE.md/AGENTS.md, although it tries to be as un-opinionated as possible). * Then just point your agent (Codex, Claude Code, Cursor, etc) at the folder and tell it what to import -- files on your machine, connect to your Readwise account, or urls from the web. * Your agent creates wthe wiki for you: Your agent will know how to ingest your raw sources (via the AGENTS.md) and will immediately start writing+linking wiki pages for you. * Go crazy on customization! The rendered wiki pages live as static html/css in your folder too so just tell your agent to change stuff, and if you need any more customization Wikiwise is fully open source :) * Ask questions about your research with your agent, ask it to bring in new sources, write new documents, etc. * (optionally) Hit the Publish button to share your wiki with friends/colleagues at a custom URL === I tried to walk the line on a couple constraints with Wikiwise: 1. I wanted it to be easy to spin up new wikis, especially without chaining together a bunch of different apps. It takes me a few minutes to spin up a new wiki on a topic -- I already have five! 2. Infinitely Customizable: one great aspect of building a wiki as Karpathy described is that you can modify any aspect of your wiki with your agent. Every new wiki styling+structure is self-contained in the local folder, which allows you to preserve this. Wikiwise is just an IDE that makes the setup easier and includes a nice un-opinionated starting state. 3. Minimal: Wikiwise is built mostly in Swift, and the DMG you install to download it is only 2.6MB (!) 4. Easy Publishing: my colleague Eleanor Konik has been building her own LLM wikis for months, but has always really struggled to actually share them with her book club. There are tools to do it, but figuring out hosting is always a huge headache. This seemed like an ideal usecase for a tool like Wikiwise to solve. The process of building wikiwise was also pretty interesting -- I "bootstrapped" the app in a way by first building my own wiki based on Karpathy's tweet and other notes I had, and slowly formed the shape of the project in collaboration with my LLM. This was all done in 3 days over the latest Readwise company hackathon we had. Truly an incredible time to be alive. Anyways, curious what you think! Links in next tweet.
3:10

Introducing Wikiwise: an open-source Mac app for managing your own Karpathy-style LLM wiki. Set up a new wiki in a few clicks: all you need is Wikiwise + your agent. It's infinitely customizable, just markdown/html under the hood, and one click to share your wiki publicly. Here's how it works: * Install Wikiwise for mac (it's built in Swift so super minimal and performant). In Karpathy's framework, Wikiwise is your IDE. * Start a new Wiki: it generates a new folder on your machine that's scaffolded in the wiki structure Andrej Karpathy describes (index.md, raw folder, wiki folder, CLAUDE.md/AGENTS.md, although it tries to be as un-opinionated as possible). * Then just point your agent (Codex, Claude Code, Cursor, etc) at the folder and tell it what to import -- files on your machine, connect to your Readwise account, or urls from the web. * Your agent creates wthe wiki for you: Your agent will know how to ingest your raw sources (via the AGENTS.md) and will immediately start writing+linking wiki pages for you. * Go crazy on customization! The rendered wiki pages live as static html/css in your folder too so just tell your agent to change stuff, and if you need any more customization Wikiwise is fully open source :) * Ask questions about your research with your agent, ask it to bring in new sources, write new documents, etc. * (optionally) Hit the Publish button to share your wiki with friends/colleagues at a custom URL === I tried to walk the line on a couple constraints with Wikiwise: 1. I wanted it to be easy to spin up new wikis, especially without chaining together a bunch of different apps. It takes me a few minutes to spin up a new wiki on a topic -- I already have five! 2. Infinitely Customizable: one great aspect of building a wiki as Karpathy described is that you can modify any aspect of your wiki with your agent. Every new wiki styling+structure is self-contained in the local folder, which allows you to preserve this. Wikiwise is just an IDE that makes the setup easier and includes a nice un-opinionated starting state. 3. Minimal: Wikiwise is built mostly in Swift, and the DMG you install to download it is only 2.6MB (!) 4. Easy Publishing: my colleague Eleanor Konik has been building her own LLM wikis for months, but has always really struggled to actually share them with her book club. There are tools to do it, but figuring out hosting is always a huge headache. This seemed like an ideal usecase for a tool like Wikiwise to solve. The process of building wikiwise was also pretty interesting -- I "bootstrapped" the app in a way by first building my own wiki based on Karpathy's tweet and other notes I had, and slowly formed the shape of the project in collaboration with my LLM. This was all done in 3 days over the latest Readwise company hackathon we had. Truly an incredible time to be alive. Anyways, curious what you think! Links in next tweet.

Tristan

95,483 views • 2 months ago

Demo alert: What if your company wiki was automatically written and updated by LLMs? Jim Fisher built a demo of this for Granola using all the meetings we've shared internally. Observations so far: it's incredibly up to date (no more stale docs!) and it requires zero effort. Not having a human author feels... different. When it's missing context on something (ie when something happens but we don't talk about it in a meeting) I want to add it manually to the wiki, but then that'll disappear the next time the page gets regenerated... Generally, makes me bullish on workflows where humans feed tons of unstructured data (easiest of which to create is conversations) to an AI, and then AI organizes + structures it on the fly.
1:00

Demo alert: What if your company wiki was automatically written and updated by LLMs? Jim Fisher built a demo of this for Granola using all the meetings we've shared internally. Observations so far: it's incredibly up to date (no more stale docs!) and it requires zero effort. Not having a human author feels... different. When it's missing context on something (ie when something happens but we don't talk about it in a meeting) I want to add it manually to the wiki, but then that'll disappear the next time the page gets regenerated... Generally, makes me bullish on workflows where humans feed tons of unstructured data (easiest of which to create is conversations) to an AI, and then AI organizes + structures it on the fly.

Chris Pedregal

15,213 views • 1 year ago

HERMES AGENT SHIPS WITH A BUNDLED SKILL FOR ANDREJ KARPATHY'S LLM WIKI PATTERN. A SELF-IMPROVING KNOWLEDGE BASE THAT GROWS EVERY TIME YOU FEED IT. mentioned this briefly in the overnight workflow article. here is the full breakdown. what it is: a self-improving knowledge base built as interlinked markdown files. unlike RAG (which rediscovers knowledge from scratch every query), the wiki compiles knowledge once and keeps it current. cross-references stay linked. contradictions get flagged automatically. synthesis reflects everything ingested so far. why this matters for Hermes memory: Hermes built-in memory knows YOU. it remembers your conversations, your preferences, your business context across sessions. but it doesn't know your inbox. or your meeting transcripts. or that article you saved last week. or the expert framework you want it to learn. the LLM Wiki solves that. THE DIVISION OF LABOR human curates sources and directs analysis. agent summarizes, cross-references, files, and maintains consistency. you drop in articles, transcripts, notes. Hermes indexes them, links related concepts, flags contradictions, updates affected pages. your knowledge base grows itself. SETUP IS ONE COMMAND the skill ships with Hermes. enable it. set WIKI_PATH in ~/.hermes/.env: WIKI_PATH=/Users/you/wiki defaults to ~/wiki if unset. then drop anything into it: "index this article into my wiki: [paste URL or text]" Hermes reads it, builds a source page, updates related entries, flags contradictions. THE OBSIDIAN ANGLE set OBSIDIAN_VAULT_PATH to the same directory. now your wiki is visible in Obsidian's graph view. nodes, links, backlinks. all built by Hermes. for headless servers: install obsidian-headless. syncs vaults without a GUI. agent writes from the server, you read on your laptop. THE COMPOUND EFFECT Hermes knows you. the wiki knows your world. combine them and the agent answers questions using BOTH contexts at once. month 1: you explain things twice. month 3: the agent references the wiki on its own. answers get sharper because the knowledge base got sharper. AUTOMATIONS THAT FEED THE WIKI set cron jobs to ingest automatically: "every day at 9am, check Granola for new meetings. add any new transcripts to my wiki under meeting notes." "every morning, scan my Gmail starred items. add anything worth keeping to the wiki." "every week, check arXiv for new papers in [your niche]. summarize and file." your wiki grows while you sleep. Hermes never forgets what gets indexed. THE LIMITATION TO KNOW unlike Hermes memory (which is conversational and lives across sessions), the wiki is a separate knowledge layer. Hermes won't pull from the wiki automatically unless you reference it or save it as a skill. best setup: build an LLM Wiki personality that tells Hermes to consult the wiki when answering strategy questions or domain-specific queries. full HERMES AGENT OVERNIGHT WORKFLOW👇
14:28

HERMES AGENT SHIPS WITH A BUNDLED SKILL FOR ANDREJ KARPATHY'S LLM WIKI PATTERN. A SELF-IMPROVING KNOWLEDGE BASE THAT GROWS EVERY TIME YOU FEED IT. mentioned this briefly in the overnight workflow article. here is the full breakdown. what it is: a self-improving knowledge base built as interlinked markdown files. unlike RAG (which rediscovers knowledge from scratch every query), the wiki compiles knowledge once and keeps it current. cross-references stay linked. contradictions get flagged automatically. synthesis reflects everything ingested so far. why this matters for Hermes memory: Hermes built-in memory knows YOU. it remembers your conversations, your preferences, your business context across sessions. but it doesn't know your inbox. or your meeting transcripts. or that article you saved last week. or the expert framework you want it to learn. the LLM Wiki solves that. THE DIVISION OF LABOR human curates sources and directs analysis. agent summarizes, cross-references, files, and maintains consistency. you drop in articles, transcripts, notes. Hermes indexes them, links related concepts, flags contradictions, updates affected pages. your knowledge base grows itself. SETUP IS ONE COMMAND the skill ships with Hermes. enable it. set WIKI_PATH in ~/.hermes/.env: WIKI_PATH=/Users/you/wiki defaults to ~/wiki if unset. then drop anything into it: "index this article into my wiki: [paste URL or text]" Hermes reads it, builds a source page, updates related entries, flags contradictions. THE OBSIDIAN ANGLE set OBSIDIAN_VAULT_PATH to the same directory. now your wiki is visible in Obsidian's graph view. nodes, links, backlinks. all built by Hermes. for headless servers: install obsidian-headless. syncs vaults without a GUI. agent writes from the server, you read on your laptop. THE COMPOUND EFFECT Hermes knows you. the wiki knows your world. combine them and the agent answers questions using BOTH contexts at once. month 1: you explain things twice. month 3: the agent references the wiki on its own. answers get sharper because the knowledge base got sharper. AUTOMATIONS THAT FEED THE WIKI set cron jobs to ingest automatically: "every day at 9am, check Granola for new meetings. add any new transcripts to my wiki under meeting notes." "every morning, scan my Gmail starred items. add anything worth keeping to the wiki." "every week, check arXiv for new papers in [your niche]. summarize and file." your wiki grows while you sleep. Hermes never forgets what gets indexed. THE LIMITATION TO KNOW unlike Hermes memory (which is conversational and lives across sessions), the wiki is a separate knowledge layer. Hermes won't pull from the wiki automatically unless you reference it or save it as a skill. best setup: build an LLM Wiki personality that tells Hermes to consult the wiki when answering strategy questions or domain-specific queries. full HERMES AGENT OVERNIGHT WORKFLOW👇

YanXbt

30,248 views • 24 days ago

The DeepSeek-R1 paper is a gem! Highly encourage everyone to read it. It's clear that LLM reasoning capabilities can be learned in different ways. RL, if applied correctly and at scale, can lead to some really powerful and interesting scaling and emergent properties. There is more to RL than meets the eye! Here is my breakdown of the paper along with a few tests: The multi-state training might not make sense initially but they provide clues on optimizations that we can continue to tap into. Data quality is still very important for enhancing the usability of the LLM. Unlike other reasoning LLMs, DeepSeek-R1's training recipe and weights are open so we can build on top of it. This opens up exciting research opportunities. About the attached clip: the previous preview model wasn't able to solve this task. DeepSeek-R1 can solve this and many other tasks that o1 can solve. It's a very good model for coding and math.
1:57

The DeepSeek-R1 paper is a gem! Highly encourage everyone to read it. It's clear that LLM reasoning capabilities can be learned in different ways. RL, if applied correctly and at scale, can lead to some really powerful and interesting scaling and emergent properties. There is more to RL than meets the eye! Here is my breakdown of the paper along with a few tests: The multi-state training might not make sense initially but they provide clues on optimizations that we can continue to tap into. Data quality is still very important for enhancing the usability of the LLM. Unlike other reasoning LLMs, DeepSeek-R1's training recipe and weights are open so we can build on top of it. This opens up exciting research opportunities. About the attached clip: the previous preview model wasn't able to solve this task. DeepSeek-R1 can solve this and many other tasks that o1 can solve. It's a very good model for coding and math.

elvis

140,692 views • 1 year ago

🚨 RL for LLMs is finally accessible. Introducing OpenTinker: The first community-driven, open-source framework designed to democratize Reinforcement Learning for LLMs. Inspired by Thinking Machines's amazing Tinker, we realize the biggest bottleneck in agentic LLM research isn’t the math—it’s the setup. Current RL pipelines are messy. Configuring VeRL for every single experiment is a productivity killer. OpenTinker fixed it. 🛠 How OpenTinker Works: Decoupled Design of Server and Client - Setup Once, Run Forever: Configure the OpenTinker backend on your GPU cluster once. - Develop Locally: Define your RL environments directly on your laptop. - Train on the Cloud: Simply point your local client to the backend. The cluster handles the compute; you handle the science. 📉 The 10x Development Efficiency Thanks to our elegant architectural decomposition, OpenTinker reduces the time to develop a new RL training pipeline by at least an order of magnitude. ⚡ Turn Idle GPU Compute into Gold Small labs often have underutilized hardware. OpenTinker turns your idle GPUs into an internal/external API service for - RL Training - SFT - Inference 🎯 Who needs OpenTinker? - Researchers tired of infrastructure hell. - Labs needing to standardize workflows. - Teams wanting to maximize hardware ROI. Thanks my amazing PhD student Siqi Zhu for leading the project. We are building the future of open RL infra. Be the first to build with us. 👇 Start Building with OpenTinker Now 🚀 Repo: 🌐 Blog: If you believe RL should be accessible to everyone, give us a star, repost this 🔄 post, and let us know what agents you plan to build!
0:42

🚨 RL for LLMs is finally accessible. Introducing OpenTinker: The first community-driven, open-source framework designed to democratize Reinforcement Learning for LLMs. Inspired by Thinking Machines's amazing Tinker, we realize the biggest bottleneck in agentic LLM research isn’t the math—it’s the setup. Current RL pipelines are messy. Configuring VeRL for every single experiment is a productivity killer. OpenTinker fixed it. 🛠 How OpenTinker Works: Decoupled Design of Server and Client - Setup Once, Run Forever: Configure the OpenTinker backend on your GPU cluster once. - Develop Locally: Define your RL environments directly on your laptop. - Train on the Cloud: Simply point your local client to the backend. The cluster handles the compute; you handle the science. 📉 The 10x Development Efficiency Thanks to our elegant architectural decomposition, OpenTinker reduces the time to develop a new RL training pipeline by at least an order of magnitude. ⚡ Turn Idle GPU Compute into Gold Small labs often have underutilized hardware. OpenTinker turns your idle GPUs into an internal/external API service for - RL Training - SFT - Inference 🎯 Who needs OpenTinker? - Researchers tired of infrastructure hell. - Labs needing to standardize workflows. - Teams wanting to maximize hardware ROI. Thanks my amazing PhD student Siqi Zhu for leading the project. We are building the future of open RL infra. Be the first to build with us. 👇 Start Building with OpenTinker Now 🚀 Repo: 🌐 Blog: If you believe RL should be accessible to everyone, give us a star, repost this 🔄 post, and let us know what agents you plan to build!

Jiaxuan You

58,144 views • 6 months ago

LLM Wikis are being slept on. I argue that creating knowledge bases with LLMs or coding agents is one of the most valuable applications of AI today. It's about being intentional in building and scaling your intelligence stack. To showcase this, I wanted to share an LLM Wiki I have built over the last couple of months. It's called PaperWiki, and I use it across all my research workflows, along with my research agents. In fact, I also use it to curate papers I share with my communities, newsletter, and on X. The PaperWiki is updated regularly with automations, so I basically have agents on a loop maintaining it. All the entries are ingested from different sources and stored in a vault (Obsidian) and further indexed using qmd. And then further presented via an HTML artifact. So all of it is easily accessible to all my agents and easily searchable through full-text search and rich semantic search. The structure of the wiki has proven significantly useful to start interesting and exciting cutting-edge research projects with my research agents (from building tiny and more efficient gpt/difussion llms to building out SoTA harnesses and memory systems). It turns out that agents love markdown files and can more easily navigate the papers given the rich metadata structure of the wiki. I am just getting started on this, but it's clear to me that we should all be experimenting with LLM Wikis. Here's why: Building LLM knowledge bases gets you into the habit of leveraging AI outputs in all kinds of creative ways. It's the good kind of tokenmaxxing we should all be pushing for. LLM Wikis can be maintained automatically in a loop. I use an automation that updates the wiki every day based on papers I curate. The curation is another automation I run in a loop (with a bit of human in the loop), so I get to build on all my previous knowledge and expertise, and all of it compounds the deeper the integration/layers. One interesting result of this process is that I feel like I can better spot high-quality papers and remove noise more easily. Social media could never solve that. And most paper aggregators use metrics I simply don't trust. I like that agents can help with the noise vs. signal problem. This is important for research. Lots of people consider agents to produce mostly slop. But it doesn't have to be that way. Careful curations, prompts, automations, verifiers, and human-in-the-loop can produce some astonishing results. And you really don't need frontier models for this. I use a combination of frontier models (opus-4.8) and open-weight models (deepseek-v4-flash) to maintain this. An exciting future work (we are working on this DAIR.AI) is to tune specialized models on top of this to allow LLMs to quickly understand cutting-edge research ideas and can better conceptualize research strategies that further accelerate scientific research agents. I plan to open-source a bunch of this work, including the artifact, but this is currently work in progress, and I was excited to share some thoughts as I continue working on it. Sharing more as I go. Stay tuned!
0:27

LLM Wikis are being slept on. I argue that creating knowledge bases with LLMs or coding agents is one of the most valuable applications of AI today. It's about being intentional in building and scaling your intelligence stack. To showcase this, I wanted to share an LLM Wiki I have built over the last couple of months. It's called PaperWiki, and I use it across all my research workflows, along with my research agents. In fact, I also use it to curate papers I share with my communities, newsletter, and on X. The PaperWiki is updated regularly with automations, so I basically have agents on a loop maintaining it. All the entries are ingested from different sources and stored in a vault (Obsidian) and further indexed using qmd. And then further presented via an HTML artifact. So all of it is easily accessible to all my agents and easily searchable through full-text search and rich semantic search. The structure of the wiki has proven significantly useful to start interesting and exciting cutting-edge research projects with my research agents (from building tiny and more efficient gpt/difussion llms to building out SoTA harnesses and memory systems). It turns out that agents love markdown files and can more easily navigate the papers given the rich metadata structure of the wiki. I am just getting started on this, but it's clear to me that we should all be experimenting with LLM Wikis. Here's why: Building LLM knowledge bases gets you into the habit of leveraging AI outputs in all kinds of creative ways. It's the good kind of tokenmaxxing we should all be pushing for. LLM Wikis can be maintained automatically in a loop. I use an automation that updates the wiki every day based on papers I curate. The curation is another automation I run in a loop (with a bit of human in the loop), so I get to build on all my previous knowledge and expertise, and all of it compounds the deeper the integration/layers. One interesting result of this process is that I feel like I can better spot high-quality papers and remove noise more easily. Social media could never solve that. And most paper aggregators use metrics I simply don't trust. I like that agents can help with the noise vs. signal problem. This is important for research. Lots of people consider agents to produce mostly slop. But it doesn't have to be that way. Careful curations, prompts, automations, verifiers, and human-in-the-loop can produce some astonishing results. And you really don't need frontier models for this. I use a combination of frontier models (opus-4.8) and open-weight models (deepseek-v4-flash) to maintain this. An exciting future work (we are working on this DAIR.AI) is to tune specialized models on top of this to allow LLMs to quickly understand cutting-edge research ideas and can better conceptualize research strategies that further accelerate scientific research agents. I plan to open-source a bunch of this work, including the artifact, but this is currently work in progress, and I was excited to share some thoughts as I continue working on it. Sharing more as I go. Stay tuned!

elvis

51,419 views • 7 days ago

Prompt engineering is where people go when they refuse to write files Karpathy's LLM Wiki gist is painfully simple: > raw sources stay untouched > the agent writes the wiki > `CLAUDE.md` or `AGENTS.md` tells it the rules > `index.md` maps the pages > `log.md` records what changed That is why a good answer should not die in chat history If the model gives you a useful comparison, summary, decision, bug explanation, or research thread, file it back into the project Use: > `wiki/topics/[topic].md` > `wiki/entities/[person-or-company].md` > `wiki/sources/[source-name].md` > `wiki/analysis/[question].md` Make the agent keep the wiki from rotting: > backlinks > summaries > contradictions > stale claims > orphan pages > missing source links For coding projects, same idea: > `CLAUDE.md`: who you are, project rules, files not to touch > `index.md`: current files and features > `log.md`: date, change, reason > saved pages: decisions you do not want to re-explain Start by saving one good answer today
11:03

Prompt engineering is where people go when they refuse to write files Karpathy's LLM Wiki gist is painfully simple: > raw sources stay untouched > the agent writes the wiki > `CLAUDE.md` or `AGENTS.md` tells it the rules > `index.md` maps the pages > `log.md` records what changed That is why a good answer should not die in chat history If the model gives you a useful comparison, summary, decision, bug explanation, or research thread, file it back into the project Use: > `wiki/topics/[topic].md` > `wiki/entities/[person-or-company].md` > `wiki/sources/[source-name].md` > `wiki/analysis/[question].md` Make the agent keep the wiki from rotting: > backlinks > summaries > contradictions > stale claims > orphan pages > missing source links For coding projects, same idea: > `CLAUDE.md`: who you are, project rules, files not to touch > `index.md`: current files and features > `log.md`: date, change, reason > saved pages: decisions you do not want to re-explain Start by saving one good answer today

Alpha Batcher

14,041 views • 1 month ago

Here's how I'm running automated content engine in 2 files 1 markdown file = my wiki 1 html file = my dashboard that's the whole stack. [ the architecture, in plain words ]: LLM wiki = a single markdown file holding my audience DNA, 15 tracked creators, every viral topic from the last 30 days HTML artifact = a single page that reads that markdown file AND can trigger my agents the artifact and the agent talk to each other directly the wiki is the shared brain [ what I actually see when I open it at 9am ]: > 5 trending topics ranked by my audience-DNA fit > 3 KOL posts worth quoting today > last week's saved tweets (so I can ride waves that are still warm) > buttons: [draft tweet] [draft QT] [schedule] [log idea] 1. I click "draft tweet" on a topic 2. the artifact pings my agent 3. agent reads the wiki, drafts in MY voice, returns it to the artifact 4. I edit, schedule, done 15 minutes from morning coffee to 3 scheduled posts [ how to build the same in one evening ]: > step 1: dump your domain knowledge into ONE markdown file (audience profile, KOL list, content rules, voice guide, anything an agent would need to do YOUR job) > step 2: ask claude to build an html artifact that reads from that file ("here's my wiki, build me a dashboard with these views") > step 3: add buttons for the actions you do daily (draft, schedule, log, score, search — your workflow, not mine) > step 4: wire each button to call your agent via tool calls (so the artifact and the agent talk directly) the moment your artifact reads your wiki AND triggers your agents.. most SaaS tools you currently pay for quietly become unnecessary dashboards I used to pay $50/month for now sit in a single html file I can rebuild in 20 minutes every "I'll build a SaaS for this" idea you had last year is a 200-line file you write in an afternoon if you want to get the same content engine, just reply "CONTENT" and will send you in DMs later we're going from buying software to owning it.
0:45

Here's how I'm running automated content engine in 2 files 1 markdown file = my wiki 1 html file = my dashboard that's the whole stack. [ the architecture, in plain words ]: LLM wiki = a single markdown file holding my audience DNA, 15 tracked creators, every viral topic from the last 30 days HTML artifact = a single page that reads that markdown file AND can trigger my agents the artifact and the agent talk to each other directly the wiki is the shared brain [ what I actually see when I open it at 9am ]: > 5 trending topics ranked by my audience-DNA fit > 3 KOL posts worth quoting today > last week's saved tweets (so I can ride waves that are still warm) > buttons: [draft tweet] [draft QT] [schedule] [log idea] 1. I click "draft tweet" on a topic 2. the artifact pings my agent 3. agent reads the wiki, drafts in MY voice, returns it to the artifact 4. I edit, schedule, done 15 minutes from morning coffee to 3 scheduled posts [ how to build the same in one evening ]: > step 1: dump your domain knowledge into ONE markdown file (audience profile, KOL list, content rules, voice guide, anything an agent would need to do YOUR job) > step 2: ask claude to build an html artifact that reads from that file ("here's my wiki, build me a dashboard with these views") > step 3: add buttons for the actions you do daily (draft, schedule, log, score, search — your workflow, not mine) > step 4: wire each button to call your agent via tool calls (so the artifact and the agent talk directly) the moment your artifact reads your wiki AND triggers your agents.. most SaaS tools you currently pay for quietly become unnecessary dashboards I used to pay $50/month for now sit in a single html file I can rebuild in 20 minutes every "I'll build a SaaS for this" idea you had last year is a 200-line file you write in an afternoon if you want to get the same content engine, just reply "CONTENT" and will send you in DMs later we're going from buying software to owning it.

Ronin

49,920 views • 2 months ago

now my clippy is an expert at whatever I want to learn + it has a chat window so i can easily follow along the conversation 1. It uses Andrej Karpathy's llm wiki so that it has specialized knowledge 2. it has a chat window so that i can follow along the conversation and copy-paste suggestions from clicky I can also type in any questions or comments if I don't feel like pushing to talk I tested this on the new Claude Managed Agents (scraped anthropics docs for it and put it into the wiki) and it works great! this is going to be an amazing way to learn new things!
1:53

now my clippy is an expert at whatever I want to learn + it has a chat window so i can easily follow along the conversation 1. It uses Andrej Karpathy's llm wiki so that it has specialized knowledge 2. it has a chat window so that i can follow along the conversation and copy-paste suggestions from clicky I can also type in any questions or comments if I don't feel like pushing to talk I tested this on the new Claude Managed Agents (scraped anthropics docs for it and put it into the wiki) and it works great! this is going to be an amazing way to learn new things!

Dan Peguine

26,516 views • 3 months ago

Super excited to launch a new AI course! 🚀 Fine-Tuning & Reinforcement Learning for LLMs: Intro to Post-Training A collaboration between AMD 🤝 Andrew Ng’s DeepLearning.AI to give every developer the tools & compute to work with the same post-training techniques, used across today’s leading AI labs. 🎓 Learn for free → 🧵
0:30

Super excited to launch a new AI course! 🚀 Fine-Tuning & Reinforcement Learning for LLMs: Intro to Post-Training A collaboration between AMD 🤝 Andrew Ng’s DeepLearning.AI to give every developer the tools & compute to work with the same post-training techniques, used across today’s leading AI labs. 🎓 Learn for free → 🧵

Sharon Zhou

20,386 views • 8 months ago

In case you missed it, we recently launched "Post-training of LLMs," a short course where you'll: ✅ Understand when and why to use post-training methods like Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), and Online Reinforcement Learning. ✅ Learn the concepts underlying the three post-training methods of SFT, DPO, and Online RL, their common use-cases, and how to curate high-quality data to effectively train a model using each method. ✅ Download a pre-trained model and implement post-training pipelines to turn a base model into an instruct model, change the identity of a chat assistant, and improve a model’s math capabilities. Learn more and enroll for free:
1:31

In case you missed it, we recently launched "Post-training of LLMs," a short course where you'll: ✅ Understand when and why to use post-training methods like Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), and Online Reinforcement Learning. ✅ Learn the concepts underlying the three post-training methods of SFT, DPO, and Online RL, their common use-cases, and how to curate high-quality data to effectively train a model using each method. ✅ Download a pre-trained model and implement post-training pipelines to turn a base model into an instruct model, change the identity of a chat assistant, and improve a model’s math capabilities. Learn more and enroll for free:

DeepLearning.AI

16,771 views • 11 months ago

"One of the very confusing things about the models right now: how to reconcile the fact that they are doing so well on evals. And you look at the evals and you go, 'Those are pretty hard evals.' But the economic impact seems to be dramatically behind. There is [a possible] explanation. Back when people were doing pre-training, the question of what data to train on was answered, because that answer was everything. So you don't have to think if it's going to be this data or that data. When people do RL training, they say, 'Okay, we want to have this kind of RL training for this thing and that kind of RL training for that thing.' You say, 'Hey, I would love our model to do really well when we release it. I want the evals to look great. What would be RL training that could help on this task?' If you combine this with generalization of the models actually being inadequate, that has the potential to explain a lot of what we are seeing, this disconnect between eval performance and actual real-world performance"
2:21

"One of the very confusing things about the models right now: how to reconcile the fact that they are doing so well on evals. And you look at the evals and you go, 'Those are pretty hard evals.' But the economic impact seems to be dramatically behind. There is [a possible] explanation. Back when people were doing pre-training, the question of what data to train on was answered, because that answer was everything. So you don't have to think if it's going to be this data or that data. When people do RL training, they say, 'Okay, we want to have this kind of RL training for this thing and that kind of RL training for that thing.' You say, 'Hey, I would love our model to do really well when we release it. I want the evals to look great. What would be RL training that could help on this task?' If you combine this with generalization of the models actually being inadequate, that has the potential to explain a lot of what we are seeing, this disconnect between eval performance and actual real-world performance"

Dwarkesh Patel

502,060 views • 7 months ago

The Independent #witchhatatelier wiki is Witch Hat's best and biggest wiki - and it's 100% add free! Tell your fellow fans to use the Indie wiki instead of Fandom! We need your support.
0:24

The Independent #witchhatatelier wiki is Witch Hat's best and biggest wiki - and it's 100% add free! Tell your fellow fans to use the Indie wiki instead of Fandom! We need your support.

Witch Hat Atelier Wiki

23,698 views • 2 months ago

This is Farzapedia. I had an LLM take 2,500 entries from my diary, Apple Notes, and some iMessage convos to create a personal Wikipedia for me. It made 400 detailed articles for my friends, my startups, research areas, and even my favorite animes and their impact on me complete with backlinks. But, this Wiki was not built for me! I built it for my agent! The structure of the wiki files and how it's all backlinked is very easily crawlable by any agent + makes it a truly useful knowledge base. I can spin up Claude Code on the wiki and starting at index.md (a catalog of all my articles) the agent does a really good job at drilling into the specific pages on my wiki it needs context on when I have a query. For example, when trying to cook up a new landing page I may ask: "I'm trying to design this landing page for a new idea I have. Please look into the images and films that inspired me recently and give me ideas for new copy and aesthetics". In my diary I kept track of everything from: learnings, people, inspo, interesting links, images. So the agent reads my wiki and pulls up my "Philosophy" articles from notes on a Studio Ghibli documentary, "Competitor" articles with YC companies whose landing pages I screenshotted, and pics of 1970s Beatles merch I saved years ago. And it delivers a great answer. I built a similar system to this a year ago with RAG but it was ass. A knowledge base that lets an agent find what it needs via a file system it actually understands just works better. The most magical thing now is as I add new things to my wiki (articles, images of inspo, meeting notes) the system will likely update 2-3 different articles where it feels that context belongs, or, just creates a new article. It's like this super genius librarian for your brain that's always filing stuff for your perfectly and also let's you easily query the knowledge for tasks useful to you (ex. design, product, writing, etc) and it never gets tired. I might spend next week productizing this, if that's of interest to you DM me + tell me your usecase!
2:24

This is Farzapedia. I had an LLM take 2,500 entries from my diary, Apple Notes, and some iMessage convos to create a personal Wikipedia for me. It made 400 detailed articles for my friends, my startups, research areas, and even my favorite animes and their impact on me complete with backlinks. But, this Wiki was not built for me! I built it for my agent! The structure of the wiki files and how it's all backlinked is very easily crawlable by any agent + makes it a truly useful knowledge base. I can spin up Claude Code on the wiki and starting at index.md (a catalog of all my articles) the agent does a really good job at drilling into the specific pages on my wiki it needs context on when I have a query. For example, when trying to cook up a new landing page I may ask: "I'm trying to design this landing page for a new idea I have. Please look into the images and films that inspired me recently and give me ideas for new copy and aesthetics". In my diary I kept track of everything from: learnings, people, inspo, interesting links, images. So the agent reads my wiki and pulls up my "Philosophy" articles from notes on a Studio Ghibli documentary, "Competitor" articles with YC companies whose landing pages I screenshotted, and pics of 1970s Beatles merch I saved years ago. And it delivers a great answer. I built a similar system to this a year ago with RAG but it was ass. A knowledge base that lets an agent find what it needs via a file system it actually understands just works better. The most magical thing now is as I add new things to my wiki (articles, images of inspo, meeting notes) the system will likely update 2-3 different articles where it feels that context belongs, or, just creates a new article. It's like this super genius librarian for your brain that's always filing stuff for your perfectly and also let's you easily query the knowledge for tasks useful to you (ex. design, product, writing, etc) and it never gets tired. I might spend next week productizing this, if that's of interest to you DM me + tell me your usecase!

Farza 🇵🇰🇺🇸

2,060,446 views • 3 months ago

We developed an RL method for fine-tuning our models for precise tasks in just a few hours or even minutes. Instead of training the whole model, we add an “RL token” output to π-0.6, our latest model, which is used by a tiny actor and critic to learn quickly with RL.
2:15

We developed an RL method for fine-tuning our models for precise tasks in just a few hours or even minutes. Instead of training the whole model, we add an “RL token” output to π-0.6, our latest model, which is used by a tiny actor and critic to learn quickly with RL.

Physical Intelligence

431,973 views • 3 months ago

New Course: Post-training of LLMs Learn to post-train and customize an LLM in this short course, taught by Banghua Zhu, Assistant Professor at the University of Washington University of Washington, and co-founder of @NexusflowX. Training an LLM to follow instructions or answer questions has two key stages: pre-training and post-training. In pre-training, it learns to predict the next word or token from large amounts of unlabeled text. In post-training, it learns useful behaviors such as following instructions, tool use, and reasoning. Post-training transforms a general-purpose token predictor—trained on trillions of unlabeled text tokens—into an assistant that follows instructions and performs specific tasks. Because it is much cheaper than pre-training, it is practical for many more teams to incorporate post-training methods into their workflows than pre-training. In this course, you’ll learn three common post-training methods—Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), and Online Reinforcement Learning (RL)—and how to use each one effectively. With SFT, you train the model on pairs of input and ideal output responses. With DPO, you provide both a preferred (chosen) and a less preferred (rejected) response and train the model to favor the preferred output. With RL, the model generates an output, receives a reward score based on human or automated feedback, and updates the model to improve performance. You’ll learn the basic concepts, common use cases, and principles for curating high-quality data for effective training. Through hands-on labs, you’ll download a pre-trained model from Hugging Face and post-train it using SFT, DPO, and RL to see how each technique shapes model behavior. In detail, you’ll: - Understand what post-training is, when to use it, and how it differs from pre-training. - Build an SFT pipeline to turn a base model into an instruct model. - Explore how DPO reshapes behavior by minimizing contrastive loss—penalizing poor responses and reinforcing preferred ones. - Implement a DPO pipeline to change the identity of a chat assistant. - Learn online RL methods such as Proximal Policy Optimization (PPO) and Group Relative Policy Optimization (GRPO), and how to design reward functions. - Train a model with GRPO to improve its math capabilities using a verifiable reward. Post-training is one of the most rapidly developing areas of LLM training. Whether you’re building a high-accuracy context-specific assistant, fine-tuning a model's tone, or improving task-specific accuracy, this course will give you experience with the most important techniques shaping how LLMs are post-trained today. Please sign up here:
2:52

New Course: Post-training of LLMs Learn to post-train and customize an LLM in this short course, taught by Banghua Zhu, Assistant Professor at the University of Washington University of Washington, and co-founder of @NexusflowX. Training an LLM to follow instructions or answer questions has two key stages: pre-training and post-training. In pre-training, it learns to predict the next word or token from large amounts of unlabeled text. In post-training, it learns useful behaviors such as following instructions, tool use, and reasoning. Post-training transforms a general-purpose token predictor—trained on trillions of unlabeled text tokens—into an assistant that follows instructions and performs specific tasks. Because it is much cheaper than pre-training, it is practical for many more teams to incorporate post-training methods into their workflows than pre-training. In this course, you’ll learn three common post-training methods—Supervised Fine-Tuning (SFT), Direct Preference Optimization (DPO), and Online Reinforcement Learning (RL)—and how to use each one effectively. With SFT, you train the model on pairs of input and ideal output responses. With DPO, you provide both a preferred (chosen) and a less preferred (rejected) response and train the model to favor the preferred output. With RL, the model generates an output, receives a reward score based on human or automated feedback, and updates the model to improve performance. You’ll learn the basic concepts, common use cases, and principles for curating high-quality data for effective training. Through hands-on labs, you’ll download a pre-trained model from Hugging Face and post-train it using SFT, DPO, and RL to see how each technique shapes model behavior. In detail, you’ll: - Understand what post-training is, when to use it, and how it differs from pre-training. - Build an SFT pipeline to turn a base model into an instruct model. - Explore how DPO reshapes behavior by minimizing contrastive loss—penalizing poor responses and reinforcing preferred ones. - Implement a DPO pipeline to change the identity of a chat assistant. - Learn online RL methods such as Proximal Policy Optimization (PPO) and Group Relative Policy Optimization (GRPO), and how to design reward functions. - Train a model with GRPO to improve its math capabilities using a verifiable reward. Post-training is one of the most rapidly developing areas of LLM training. Whether you’re building a high-accuracy context-specific assistant, fine-tuning a model's tone, or improving task-specific accuracy, this course will give you experience with the most important techniques shaping how LLMs are post-trained today. Please sign up here:

Andrew Ng

125,146 views • 1 year ago

I like to think of evals as something active, not passive -- it's a North Star that steers LLMs toward higher intelligence. Evals should drive your RL/SFT/post-training decisions. Internal evals at frontier labs make a huge difference -- and you can see it in how models behave differently (GPT seems better at one-shot tasks, Claude at multi-turn). If you want to learn more about building evals that actually improve your model in post-training, check out our AMD x DeeplearningAI course "Fine-tuning & RL for LLMs: Intro to Post-training" (content is free):
0:30

I like to think of evals as something active, not passive -- it's a North Star that steers LLMs toward higher intelligence. Evals should drive your RL/SFT/post-training decisions. Internal evals at frontier labs make a huge difference -- and you can see it in how models behave differently (GPT seems better at one-shot tasks, Claude at multi-turn). If you want to learn more about building evals that actually improve your model in post-training, check out our AMD x DeeplearningAI course "Fine-tuning & RL for LLMs: Intro to Post-training" (content is free):

Sharon Zhou

16,613 views • 5 months ago

What if you could train AI agents on a laptop as easily as on a GPU cluster? Researchers from UIUC's U Lab, led by Prof. Jiaxuan You, just open-sourced OpenTinker. It's a new "Reinforcement-Learning-as-a-Service" (RLaaS) system that decouples the complex training pipeline into simple, distributed services with friendly APIs. The result? It breaks down the major engineering barriers to RL, outperforming traditional frameworks in accessibility and ease of deployment, finally making agent training viable for more developers and teams. Project: Code: U Lab: Our report: 📬 #PapersAccepted by Jiqizhixin
0:42

What if you could train AI agents on a laptop as easily as on a GPU cluster? Researchers from UIUC's U Lab, led by Prof. Jiaxuan You, just open-sourced OpenTinker. It's a new "Reinforcement-Learning-as-a-Service" (RLaaS) system that decouples the complex training pipeline into simple, distributed services with friendly APIs. The result? It breaks down the major engineering barriers to RL, outperforming traditional frameworks in accessibility and ease of deployment, finally making agent training viable for more developers and teams. Project: Code: U Lab: Our report: 📬 #PapersAccepted by Jiqizhixin

机器之心 JIQIZHIXIN

15,893 views • 6 months ago

We’re launching Mundi, the first open-source web GIS built for AI. After years of training geospatial AI models, we decided now is the moment to build the GIS software of the next decade. After the AI Vectorizer, Georeferencer, and Kue (our LLM agent inside QGIS), we realized that GIS is at an inflection point. But existing mapping software was neither powerful enough to allow for synchronous collaboration with AI agents, nor open source to allow the community to build on top of it. If we want to get the most out of LLMs, we need to create our own GIS that is purposefully designed for AI—Mundi. With Mundi, you collaborate with Kue, our AI GIS agent, to edit your map, rather than physically interacting with tools as you would in a legacy GIS.
5:05

We’re launching Mundi, the first open-source web GIS built for AI. After years of training geospatial AI models, we decided now is the moment to build the GIS software of the next decade. After the AI Vectorizer, Georeferencer, and Kue (our LLM agent inside QGIS), we realized that GIS is at an inflection point. But existing mapping software was neither powerful enough to allow for synchronous collaboration with AI agents, nor open source to allow the community to build on top of it. If we want to get the most out of LLMs, we need to create our own GIS that is purposefully designed for AI—Mundi. With Mundi, you collaborate with Kue, our AI GIS agent, to edit your map, rather than physically interacting with tools as you would in a legacy GIS.

Brendan Ashworth

41,011 views • 1 year ago

Postman is going all-in on AI! Their AI Agent Builder is live (and it's pretty powerful!) Postman's API Network has 100,000+ APIs and they just released a suite of tools for building AI Agents that can automatically discover and connect to any of these APIs. Think about it for a second: You can now build an AI agent that searches the network, discovers an API, processes its documentation, connects to it, and evaluates whether everything worked at the end of the process. This is a huge deal. Every person I know uses the Postman client to work with APIs. Now, they will have access to the AI Agent Builder, a complete suite for designing, building, testing, and deploying agents. I've built a few simple agents myself, and it's tough. The ability to discover and use an API automatically is really exciting (and a huge time saver if it works as promised!) In collaboration with the Postman team, who is sponsoring this post, I recorded a video with my thoughts about the Postman AI Agent Builder.
3:38

Postman is going all-in on AI! Their AI Agent Builder is live (and it's pretty powerful!) Postman's API Network has 100,000+ APIs and they just released a suite of tools for building AI Agents that can automatically discover and connect to any of these APIs. Think about it for a second: You can now build an AI agent that searches the network, discovers an API, processes its documentation, connects to it, and evaluates whether everything worked at the end of the process. This is a huge deal. Every person I know uses the Postman client to work with APIs. Now, they will have access to the AI Agent Builder, a complete suite for designing, building, testing, and deploying agents. I've built a few simple agents myself, and it's tough. The ability to discover and use an API automatically is really exciting (and a huge time saver if it works as promised!) In collaboration with the Postman team, who is sponsoring this post, I recorded a video with my thoughts about the Postman AI Agent Builder.

Santiago

230,571 views • 1 year ago