正在加载视频...

视频加载失败

返回首页

New skill: self-managed-context (make the agent's context an editable file) It explains how to build agents that decide what to keep, update, or remove from the information they use to do their work. It can archive a long log while keeping the exact error, update its progress notes, or...

muratcan's profile picture

Muratcan Koylan

22,960 subscribers

15,070 次观看 • 3 天前 •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

5 条评论

Shahid🇵🇸 的头像
Shahid🇵🇸2 天前

forced summary also going 3/3 makes sense at 2x, it only fires a few times. where i'd expect them to split is when summaries stack and a config value has been paraphrased twice over. the same task at 6-8x would say more about the exact-value advantage

Jatin Garg 的头像
Jatin Garg3 天前

programmatic context management gives you full control and visibility into what gets kept or dropped, which matters when debugging failures. learned policies adapt faster to new patterns but the edit decisions become a black box you can't inspect when something goes wrong.

Nathan 的头像
Nathan2 天前

Can the latest Claude code mods system achieve this effect?

AI Quanting 的头像
AI Quanting2 天前

The archiving is the easy half. Knowing which error you'll need in an hour is the part models get wrong.

Bessi S 的头像
Bessi S2 天前

need self-managed-context before the next 80k dump

相关视频

Google researchers just fixed the dirtiest secret of self-improving AI agents. When an agent rewrites its own harness, it cheats. It keeps editing until the benchmark score climbs, and the gains disappear the moment it sees a new task. The harness is everything around the model: prompts, tools, memory and control flow. RRSI puts rules on the loop that rewrites it: > Fewer edits are allowed every round. > A critic throws out edits that hardcode task answers. > An edit survives only if it beats run-to-run noise. > Extra cost has to pay for itself. > Parts that stopped helping get pruned. Up to +14.1 points on the tasks it tunes on, up to +4.7 on five benchmarks it never saw, 30% fewer tokens, and it carries over to other model families. The code is open. Write one adapter file that runs your agent on your tasks, and it evolves the harness in git worktrees round by round.
0:24

Google researchers just fixed the dirtiest secret of self-improving AI agents. When an agent rewrites its own harness, it cheats. It keeps editing until the benchmark score climbs, and the gains disappear the moment it sees a new task. The harness is everything around the model: prompts, tools, memory and control flow. RRSI puts rules on the loop that rewrites it: > Fewer edits are allowed every round. > A critic throws out edits that hardcode task answers. > An edit survives only if it beats run-to-run noise. > Extra cost has to pay for itself. > Parts that stopped helping get pruned. Up to +14.1 points on the tasks it tunes on, up to +4.7 on five benchmarks it never saw, 30% fewer tokens, and it carries over to other model families. The code is open. Write one adapter file that runs your agent on your tasks, and it evolves the harness in git worktrees round by round.

slash1s

18,572 次观看 • 3 天前

this video is the CLEAREST explanation of how claude skills + AI agents work and how to use them most people set up an AI agent and wonder why it keeps disappointing them. the context window is everything context is what the model assembles before it takes any action. think of it like everything the agent needs to read before it does anything. the quality of what goes in determines the quality of what comes out. the models are genuinely really good right now. claude and gpt are exceptional. the variable is almost always the context you give them. 1. agent.md files are mostly unnecessary every single line you put in an agent.md file gets added to every single conversation you have with your agent. a 1000 line file is around 7000 tokens burning on every run. the model already knows to use react. it can read your codebase. save the agent.md for proprietary information specific to your company that the model genuinely cannot know on its own. 2. skills are the actual unlock a skill.md file works differently. what loads into context is only the name and description, around 50 tokens. the full instructions only appear when the agent recognizes it needs that skill. so instead of 7000 tokens on every run you have 50. and the agent stays sharp because the context window stays lean. the closer you get to filling the context window the worse the agent performs, same way you perform worse when someone dumps 10 things on you at once. 3. here is how to actually build a skill the right way most people identify a workflow and immediately try to write the skill. what you want to do instead is run the workflow by hand with the agent first. walk it through every single step. tell it what to check, what good looks like, what bad looks like. correct it in real time. once you have had a full successful run from start to finish, tell the agent to review everything it just did and write the skill itself. it writes a better skill than you will because it has the full context of what actually worked in practice not in theory. 4. recursively building skills is how you go from frustrated to reliable when the skill breaks, and it will break, ask the agent exactly why it failed. it will tell you specifically what went wrong. fix it together in that same conversation. then tell it to update the skill file so that failure mode never happens again. ross mike did this five times with his youtube report generator. it now pulls from eight different data sources and runs flawlessly every single time without him touching it. 5. sub agents are something you earn not something you set up on day one start with one agent. build one workflow. turn it into one skill. once that works add another. ross mike has five sub agents now covering marketing, business, personal and more. it took months to get there and every single one exists because a workflow proved it deserved to exist. the people who set up 15 sub agents on day one and wonder why nothing works skipped all the steps that make the thing actually run. 6. your workflow is the thing the model cannot get anywhere else the model has been trained on everything. it knows more than you about most things. what it does not have is your specific process, your taste, your way of doing things. that is what skills capture. that is what makes your agent actually useful versus a generic one. downloading someone else's skill means downloading their context onto your setup and it will not work the way you want it to because it was never built around how you work. this is the clearest explanation of how agents actually work i have heard. Micky runs this stuff every single day and the results show it. full episode is now live on The Startup Ideas Podcast (SIP) 🧃 where you get your pods people charge for this sorta stuff i give away the sauce for free i just want you to win watch
35:25

this video is the CLEAREST explanation of how claude skills + AI agents work and how to use them most people set up an AI agent and wonder why it keeps disappointing them. the context window is everything context is what the model assembles before it takes any action. think of it like everything the agent needs to read before it does anything. the quality of what goes in determines the quality of what comes out. the models are genuinely really good right now. claude and gpt are exceptional. the variable is almost always the context you give them. 1. agent.md files are mostly unnecessary every single line you put in an agent.md file gets added to every single conversation you have with your agent. a 1000 line file is around 7000 tokens burning on every run. the model already knows to use react. it can read your codebase. save the agent.md for proprietary information specific to your company that the model genuinely cannot know on its own. 2. skills are the actual unlock a skill.md file works differently. what loads into context is only the name and description, around 50 tokens. the full instructions only appear when the agent recognizes it needs that skill. so instead of 7000 tokens on every run you have 50. and the agent stays sharp because the context window stays lean. the closer you get to filling the context window the worse the agent performs, same way you perform worse when someone dumps 10 things on you at once. 3. here is how to actually build a skill the right way most people identify a workflow and immediately try to write the skill. what you want to do instead is run the workflow by hand with the agent first. walk it through every single step. tell it what to check, what good looks like, what bad looks like. correct it in real time. once you have had a full successful run from start to finish, tell the agent to review everything it just did and write the skill itself. it writes a better skill than you will because it has the full context of what actually worked in practice not in theory. 4. recursively building skills is how you go from frustrated to reliable when the skill breaks, and it will break, ask the agent exactly why it failed. it will tell you specifically what went wrong. fix it together in that same conversation. then tell it to update the skill file so that failure mode never happens again. ross mike did this five times with his youtube report generator. it now pulls from eight different data sources and runs flawlessly every single time without him touching it. 5. sub agents are something you earn not something you set up on day one start with one agent. build one workflow. turn it into one skill. once that works add another. ross mike has five sub agents now covering marketing, business, personal and more. it took months to get there and every single one exists because a workflow proved it deserved to exist. the people who set up 15 sub agents on day one and wonder why nothing works skipped all the steps that make the thing actually run. 6. your workflow is the thing the model cannot get anywhere else the model has been trained on everything. it knows more than you about most things. what it does not have is your specific process, your taste, your way of doing things. that is what skills capture. that is what makes your agent actually useful versus a generic one. downloading someone else's skill means downloading their context onto your setup and it will not work the way you want it to because it was never built around how you work. this is the clearest explanation of how agents actually work i have heard. Micky runs this stuff every single day and the results show it. full episode is now live on The Startup Ideas Podcast (SIP) 🧃 where you get your pods people charge for this sorta stuff i give away the sauce for free i just want you to win watch

GREG ISENBERG

194,524 次观看 • 5 个月前

Sam Altman made the case for open-source harnesses in July. a month later, someone shipped it, and it's more efficient than most managed harnesses. here is the problem it was aimed at: a large share of your agent's token bill is the model rereading things it already read. that isn't the model's doing. the runtime around it decides what goes into every prompt and how often the model gets called. for example, an agent queries a CRM at step four and gets back 400 rows. those rows get piled up in the conversation history. by step nineteen, the model has to read those rows fifteen times unnecessarily, and every token read is billed at input rates. it happened because your harness assembled that prompt on every turn and kept the rows in it. that gives you two levers: how much context the harness carries forward, and how often it calls the model. there are four practical ways to keep the prompt from growing unnecessarily: → load tool schemas on demand. a server with 100 tools doesn't need to put all 100 into every prompt when the agent only calls two. → offload large results to disk. turn a large response into a short preview and a file path instead of replaying the entire result on every turn. → delegate to subagents. let a subagent spend thirty tool calls in its own context and return one summary to the root agent. → run toolchains in code. one script calls three tools, joins the results, and returns a table instead of three turns each dragging a full response. but reducing context is only half the job. you also need to control how often the model gets called. a good harness should avoid unnecessary planning, verification, and reflection when the work can be completed in fewer steps. TrueFoundry's open-source agent harness, TrueForge, is built around both of those controls. it sits between the model and the tools, deciding what goes into every prompt and when another model call is actually needed. it also breaks token usage down across the harness, skills, instructions, tools, and messages. DevRev's Enterprise-Bench is where this gets tested, on multi-step tasks of the kind where an agent pulls records from one system and reconciles them against another. TrueFoundry ran TrueForge there against Claude Managed Agents, both on the same model, and both finished the same number of tasks. the tie is the part that matters, because it means the gap underneath is not a quality tradeoff. TrueForge reached that score on close to a third of the tokens, with roughly 40% fewer trips back to the model. for the same result, that comes out around 2.7x cheaper than Claude Managed Agents. swapping in an open model made it sharper still. TrueForge with GLM-5.2 scored a little higher than either setup above, and the entire benchmark run cost about $3 at list prices. being open source matters beyond the license here. the model underneath can be swapped without rewriting the agent, and the whole thing can run inside your own environment when the data cannot leave it. all of this comes down to the runtime around the model, the context it carries, the tools it exposes, and how many times it goes back to the model. that is what a production harness actually owns. the full task list, the per-run numbers, and the MIT-licensed code are on GitHub: (don't forget to star 🌟) you can read more about the same in the article quoted below. thanks to the TrueForge team for working with me on this one.
0:39

Sam Altman made the case for open-source harnesses in July. a month later, someone shipped it, and it's more efficient than most managed harnesses. here is the problem it was aimed at: a large share of your agent's token bill is the model rereading things it already read. that isn't the model's doing. the runtime around it decides what goes into every prompt and how often the model gets called. for example, an agent queries a CRM at step four and gets back 400 rows. those rows get piled up in the conversation history. by step nineteen, the model has to read those rows fifteen times unnecessarily, and every token read is billed at input rates. it happened because your harness assembled that prompt on every turn and kept the rows in it. that gives you two levers: how much context the harness carries forward, and how often it calls the model. there are four practical ways to keep the prompt from growing unnecessarily: → load tool schemas on demand. a server with 100 tools doesn't need to put all 100 into every prompt when the agent only calls two. → offload large results to disk. turn a large response into a short preview and a file path instead of replaying the entire result on every turn. → delegate to subagents. let a subagent spend thirty tool calls in its own context and return one summary to the root agent. → run toolchains in code. one script calls three tools, joins the results, and returns a table instead of three turns each dragging a full response. but reducing context is only half the job. you also need to control how often the model gets called. a good harness should avoid unnecessary planning, verification, and reflection when the work can be completed in fewer steps. TrueFoundry's open-source agent harness, TrueForge, is built around both of those controls. it sits between the model and the tools, deciding what goes into every prompt and when another model call is actually needed. it also breaks token usage down across the harness, skills, instructions, tools, and messages. DevRev's Enterprise-Bench is where this gets tested, on multi-step tasks of the kind where an agent pulls records from one system and reconciles them against another. TrueFoundry ran TrueForge there against Claude Managed Agents, both on the same model, and both finished the same number of tasks. the tie is the part that matters, because it means the gap underneath is not a quality tradeoff. TrueForge reached that score on close to a third of the tokens, with roughly 40% fewer trips back to the model. for the same result, that comes out around 2.7x cheaper than Claude Managed Agents. swapping in an open model made it sharper still. TrueForge with GLM-5.2 scored a little higher than either setup above, and the entire benchmark run cost about $3 at list prices. being open source matters beyond the license here. the model underneath can be swapped without rewriting the agent, and the whole thing can run inside your own environment when the data cannot leave it. all of this comes down to the runtime around the model, the context it carries, the tools it exposes, and how many times it goes back to the model. that is what a production harness actually owns. the full task list, the per-run numbers, and the MIT-licensed code are on GitHub: (don't forget to star 🌟) you can read more about the same in the article quoted below. thanks to the TrueForge team for working with me on this one.

Akshay 🚀

76,998 次观看 • 1 个月前

AI AGENTS 101 (58 minute free masterclass) send this to anyone who wants to understand ai agents, claude skills, md files, how to get the most out of AI etc in plain english: 1. chat vs agents - chat models answer questions in a back and forth while agents take a goal, figure out the steps, and deliver a result 2. agents don’t stop after one response. they keep running until the task is actually finishedno babysitting required 3. everything runs on a loop. they gather context, decide what to do, take an action, then repeat until done 4. the loop is the system. they look at files, tools, and the internet. decide the next step. execute and then feed that back into the next step. over and over until completion 5. the model is just one piece. gpt, claude, gemini are the reasoning layer. the key is model + loop + tools + context 6. mcp is how agents use tools. it connects things like browser, code, apis, and your internal software. once connected, the agent decides when to use them to get the job done 7. context beats prompt all day. you don't need to write perfect prompts. load your agent with context about your business, style, and goals and then simple instructions work 8. claude.md or agents.md is the onboarding doc it tells the agent who it is, how to behave, what it knows, and what tools it can use. this gets loaded every time before it starts 9. memory.md is how it improves. agents don’t remember by default. this file stores preferences, corrections, and patterns you tell the agent to update it, and it gets better over time 10. skills + harnesses make it usable. skills are reusable tasks like writing, research, analysis the harness is the environment like claude code or openclaw that runs everything. basiclaly, different interfaces, same system underneath this episode with remy on The Startup Ideas Podcast (SIP) 🧃 was one of the clearest ways of understanding a lot of the core concepts of ai agents could be the best beginners course for ai agents 58 mins. all free. no advertisers. i just want to see you build cool stuff. im rooting for you. send to a friend watch
58:55

AI AGENTS 101 (58 minute free masterclass) send this to anyone who wants to understand ai agents, claude skills, md files, how to get the most out of AI etc in plain english: 1. chat vs agents - chat models answer questions in a back and forth while agents take a goal, figure out the steps, and deliver a result 2. agents don’t stop after one response. they keep running until the task is actually finishedno babysitting required 3. everything runs on a loop. they gather context, decide what to do, take an action, then repeat until done 4. the loop is the system. they look at files, tools, and the internet. decide the next step. execute and then feed that back into the next step. over and over until completion 5. the model is just one piece. gpt, claude, gemini are the reasoning layer. the key is model + loop + tools + context 6. mcp is how agents use tools. it connects things like browser, code, apis, and your internal software. once connected, the agent decides when to use them to get the job done 7. context beats prompt all day. you don't need to write perfect prompts. load your agent with context about your business, style, and goals and then simple instructions work 8. claude.md or agents.md is the onboarding doc it tells the agent who it is, how to behave, what it knows, and what tools it can use. this gets loaded every time before it starts 9. memory.md is how it improves. agents don’t remember by default. this file stores preferences, corrections, and patterns you tell the agent to update it, and it gets better over time 10. skills + harnesses make it usable. skills are reusable tasks like writing, research, analysis the harness is the environment like claude code or openclaw that runs everything. basiclaly, different interfaces, same system underneath this episode with remy on The Startup Ideas Podcast (SIP) 🧃 was one of the clearest ways of understanding a lot of the core concepts of ai agents could be the best beginners course for ai agents 58 mins. all free. no advertisers. i just want to see you build cool stuff. im rooting for you. send to a friend watch

GREG ISENBERG

377,833 次观看 • 6 个月前

THIS GUY CONNECTED HIS AI AGENTS TO HIS OBSIDIAN AND BUILT A BRAIN THAT LEARNS ON ITS OWN. HERE'S HOW TO BUILD IT Obsidian is just markdown files sitting in a folder. That turns out to be the perfect memory for an AI agent, because an agent can read and write those files directly. He wired his agents into the vault so they pull context from it, do the work, and write what they learned back. The notes aren't the point. The loop is, and it gets sharper every cycle How to build it: 1. Point an agent at your vault. The fastest way, no plugins, no API keys: open a terminal and run npx obsidian-mcp /path/to/your/vault. That exposes your Obsidian folder to Claude as a tool it can read, search, and write to. Add it to your Claude Code or Cowork config and restart 2. Confirm it can see the brain. Ask it: "list the notes in my vault and summarize what's in them." If it reads them back, the connection is live. Now it starts every task with everything the vault already holds instead of from zero 3. Give each agent one job and a write-back rule. Tell it: "research this, then save what you found as a new note in /brain with links to related notes." One agent researches, one summarizes, one plans. Each writes its output back into the vault 4. Close the loop. Add one line to every agent's instructions: "read /brain before starting, write your result back when done." Now each task leaves the vault richer, and the next run reads that before it works. It compounds instead of resetting 5. You only steer. Review what the brain produces, point it at the next thing. The agents handle the reading, writing, and connecting The edge isn't better notes. It's a brain that feeds itself, so the work gets sharper every cycle instead of starting over Bookmark this
0:24

THIS GUY CONNECTED HIS AI AGENTS TO HIS OBSIDIAN AND BUILT A BRAIN THAT LEARNS ON ITS OWN. HERE'S HOW TO BUILD IT Obsidian is just markdown files sitting in a folder. That turns out to be the perfect memory for an AI agent, because an agent can read and write those files directly. He wired his agents into the vault so they pull context from it, do the work, and write what they learned back. The notes aren't the point. The loop is, and it gets sharper every cycle How to build it: 1. Point an agent at your vault. The fastest way, no plugins, no API keys: open a terminal and run npx obsidian-mcp /path/to/your/vault. That exposes your Obsidian folder to Claude as a tool it can read, search, and write to. Add it to your Claude Code or Cowork config and restart 2. Confirm it can see the brain. Ask it: "list the notes in my vault and summarize what's in them." If it reads them back, the connection is live. Now it starts every task with everything the vault already holds instead of from zero 3. Give each agent one job and a write-back rule. Tell it: "research this, then save what you found as a new note in /brain with links to related notes." One agent researches, one summarizes, one plans. Each writes its output back into the vault 4. Close the loop. Add one line to every agent's instructions: "read /brain before starting, write your result back when done." Now each task leaves the vault richer, and the next run reads that before it works. It compounds instead of resetting 5. You only steer. Review what the brain produces, point it at the next thing. The agents handle the reading, writing, and connecting The edge isn't better notes. It's a brain that feeds itself, so the work gets sharper every cycle instead of starting over Bookmark this

Yarchi

58,591 次观看 • 3 个月前

A screen recording is the best context you can give an AI agent. It's also the worst. A few seconds of video is thousands of images. Feed that to an agent and you instantly blow up its memory. So I built a tool that fixes it. How it works: Hit record and just do your thing. Click around, type, and talk out loud about what you're doing. While you record, it quietly captures the stuff that actually matters: your voice transcription, every click and the exact thing you clicked, what you typed, and the shortcuts you press. When you stop, it doesn't hand the agent the whole video. It writes a short, readable summary: a timeline of what you said and did. The agent reads that tiny summary first. Then, only when it needs to actually see a moment, it pulls that one exact frame from the video, like flipping to a single page instead of reading the whole book. The result: the agent understands everything you did! Full context. Less tokens.
1:54

A screen recording is the best context you can give an AI agent. It's also the worst. A few seconds of video is thousands of images. Feed that to an agent and you instantly blow up its memory. So I built a tool that fixes it. How it works: Hit record and just do your thing. Click around, type, and talk out loud about what you're doing. While you record, it quietly captures the stuff that actually matters: your voice transcription, every click and the exact thing you clicked, what you typed, and the shortcuts you press. When you stop, it doesn't hand the agent the whole video. It writes a short, readable summary: a timeline of what you said and did. The agent reads that tiny summary first. Then, only when it needs to actually see a moment, it pulls that one exact frame from the video, like flipping to a single page instead of reading the whole book. The result: the agent understands everything you did! Full context. Less tokens.

Milind S

12,486 次观看 • 3 个月前

ANTHROPIC JUST QUIETLY SHIPPED A FEATURE THAT LETS CLAUDE SPAWN A WHOLE TEAM OF AGENTS THAT MESSAGE EACH OTHER AND REVIEW EACH OTHER'S WORK. It's a Claude Code feature called agent teams. The team lead spawns multiple agents that share a task list and message each other directly, not subagents reporting back, actual peers. In the demo a QA agent caught three bugs, sent the work back to the front-end and back-end devs, they fixed it, app shipped in one pass. How to run it: 1. Enable it. Needs Claude Code v2.1.32+. Add to settings.json: "env": { "CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS": "1" }. Or paste that to Claude and say "add this to my settings." Restart. 2. Prompt in plain English. Start with a goal (agents wake with zero context), then "create a team of 3 using Sonnet," describe each role, its deliverable, and who it messages when done. 3. The rules: each agent owns its own files, define exact outputs, name who talks to who, keep it to 3-5 agents. Use it for complex work with separate parts running in parallel. Skip it for simple or sequential tasks, teams cost 3-4x the tokens. Bookmark this.
16:29

ANTHROPIC JUST QUIETLY SHIPPED A FEATURE THAT LETS CLAUDE SPAWN A WHOLE TEAM OF AGENTS THAT MESSAGE EACH OTHER AND REVIEW EACH OTHER'S WORK. It's a Claude Code feature called agent teams. The team lead spawns multiple agents that share a task list and message each other directly, not subagents reporting back, actual peers. In the demo a QA agent caught three bugs, sent the work back to the front-end and back-end devs, they fixed it, app shipped in one pass. How to run it: 1. Enable it. Needs Claude Code v2.1.32+. Add to settings.json: "env": { "CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS": "1" }. Or paste that to Claude and say "add this to my settings." Restart. 2. Prompt in plain English. Start with a goal (agents wake with zero context), then "create a team of 3 using Sonnet," describe each role, its deliverable, and who it messages when done. 3. The rules: each agent owns its own files, define exact outputs, name who talks to who, keep it to 3-5 agents. Use it for complex work with separate parts running in parallel. Skip it for simple or sequential tasks, teams cost 3-4x the tokens. Bookmark this.

Yarchi

462,780 次观看 • 3 个月前

Can your agent finally edit your videos? Yup, it's a repo called video use, by Browser Use. You put the raw footage in a folder. Then Claude Code or Codex edits it. It removes the filler words, cuts the dead space, adds the subtitles, and renders the file. Then it checks its own cut points. Most creators use the same small set of edits in every video. That style stays in your head and in your hands. This repo puts it in a file that an agent can read. People ask for a full YouTube channel on day one. They see bad results, and they stop. Start with one format instead. - One founder loom into a 60 second launch video - 3 clips out of one podcast recording Then make the format a system. Then you operate the system, or you sell it. - Real estate agents need listing videos - Coaches need clips - Agencies need ad variations Manual installation with git clone is possible. For 95% of people, the agent is faster. Here's how i intalled it: - Gave the setup prompt to Claude Code with shell access - Let it read install.md - Installed the repo - Connected FFmpeg - Registered the skill
4:22

Can your agent finally edit your videos? Yup, it's a repo called video use, by Browser Use. You put the raw footage in a folder. Then Claude Code or Codex edits it. It removes the filler words, cuts the dead space, adds the subtitles, and renders the file. Then it checks its own cut points. Most creators use the same small set of edits in every video. That style stays in your head and in your hands. This repo puts it in a file that an agent can read. People ask for a full YouTube channel on day one. They see bad results, and they stop. Start with one format instead. - One founder loom into a 60 second launch video - 3 clips out of one podcast recording Then make the format a system. Then you operate the system, or you sell it. - Real estate agents need listing videos - Coaches need clips - Agencies need ad variations Manual installation with git clone is possible. For 95% of people, the agent is faster. Here's how i intalled it: - Gave the setup prompt to Claude Code with shell access - Let it read install.md - Installed the repo - Connected FFmpeg - Registered the skill

The Startup Ideas Podcast (SIP) 🧃

15,605 次观看 • 1 个月前

Everyone wants agent swarms. Very few people are talking seriously enough about the context layer that makes swarms useful. Even with one agent, context is fragile. Too little context and the agent guesses. Too much context and it wastes tokens, loses focus, or reasons over irrelevant noise. The sweet spot is precise context: the right knowledge, in the right structure, at the right moment. With many agents, that challenge explodes. Each agent produces decisions, assumptions, findings, summaries, risks, and partial conclusions. Unless that knowledge becomes shared, structured, and reusable, every new agent is forced to rediscover what another agent already learned. That is not a swarm. That is a crowd. Shared context graphs are what turn agent activity into agent collaboration, and OriginTrail DKG V10 brings them to life. Was just playing with some final polishing for the V10 release, and it is really powerful to see shared context graphs where multiple agents contribute knowledge into the same connected memory, with attribution visible directly in the graph ui. That matters for three reasons. First, agents can access and build on one shared memory instead of staying trapped in isolated sessions. Second, the graph structure helps them retrieve the exact context they need, instead of stuffing everything into a prompt and hoping the model sorts it out. Third, verifiability of provenance. You can see which agent contributed each piece of knowledge, trace the source, and decide what to trust. Tokenmaxxing starts with fewer tokens, but the deeper story is coordination - agents stop reloading the world and start building on shared, verifiable context. That is the foundation for serious multi-agent work across software engineering, research, finance, operations, project management, and far beyond. The future is not more agents, it is agents working from shared, verifiable context. But the more the merrier, of course.
0:22

Everyone wants agent swarms. Very few people are talking seriously enough about the context layer that makes swarms useful. Even with one agent, context is fragile. Too little context and the agent guesses. Too much context and it wastes tokens, loses focus, or reasons over irrelevant noise. The sweet spot is precise context: the right knowledge, in the right structure, at the right moment. With many agents, that challenge explodes. Each agent produces decisions, assumptions, findings, summaries, risks, and partial conclusions. Unless that knowledge becomes shared, structured, and reusable, every new agent is forced to rediscover what another agent already learned. That is not a swarm. That is a crowd. Shared context graphs are what turn agent activity into agent collaboration, and OriginTrail DKG V10 brings them to life. Was just playing with some final polishing for the V10 release, and it is really powerful to see shared context graphs where multiple agents contribute knowledge into the same connected memory, with attribution visible directly in the graph ui. That matters for three reasons. First, agents can access and build on one shared memory instead of staying trapped in isolated sessions. Second, the graph structure helps them retrieve the exact context they need, instead of stuffing everything into a prompt and hoping the model sorts it out. Third, verifiability of provenance. You can see which agent contributed each piece of knowledge, trace the source, and decide what to trust. Tokenmaxxing starts with fewer tokens, but the deeper story is coordination - agents stop reloading the world and start building on shared, verifiable context. That is the foundation for serious multi-agent work across software engineering, research, finance, operations, project management, and far beyond. The future is not more agents, it is agents working from shared, verifiable context. But the more the merrier, of course.

Jurij Skornik

11,180 次观看 • 4 个月前

Creator of agent loops: "Anthropic also using my agent systems formula and still the biggest mistake in agents: is skipping Graphs" his 3 rules for every agent: rule 1 → loop it, don't run once - think, act, observe, repeat - cut the context and it forgets why it started rule 2 → an agent is memory + tools + reasoning - drop one and it breaks rule 3 → stop fine-tuning the model - the real alpha is the infrastructure around it the formula behind every AI agent - explained by the one who wrote it one agent → a loop → a graph of loops → a system that never forgets watch & bookmark - then read how to wire it into a graph ↓
45:37

Creator of agent loops: "Anthropic also using my agent systems formula and still the biggest mistake in agents: is skipping Graphs" his 3 rules for every agent: rule 1 → loop it, don't run once - think, act, observe, repeat - cut the context and it forgets why it started rule 2 → an agent is memory + tools + reasoning - drop one and it breaks rule 3 → stop fine-tuning the model - the real alpha is the infrastructure around it the formula behind every AI agent - explained by the one who wrote it one agent → a loop → a graph of loops → a system that never forgets watch & bookmark - then read how to wire it into a graph ↓

codila

23,915 次观看 • 2 个月前

Anthropic engineer: "Harness engineering is a really rare skill right now - and it's hard to know if you're even good at it." how he actually thinks about building the full system: • step 1 → the model. already smarter than we use it. "freeze development - you'd find new capabilities for another 6-12 months" • step 2 → the harness. delete the steering. "we stopped using plan mode - the model thinks correctly" • step 3 → the agents. workflows let Claude build its own harness per task and spin sub-agents. what didn't fit in one context window now runs across many • step 4 → the context. "RAG is an anti-pattern. use grep." memory via files, clean context window, prompt caching. that's the stack most people optimize step 1. he optimizes all four. Watch the full video, then read the guide below on how to build an agentic system.
46:31

Anthropic engineer: "Harness engineering is a really rare skill right now - and it's hard to know if you're even good at it." how he actually thinks about building the full system: • step 1 → the model. already smarter than we use it. "freeze development - you'd find new capabilities for another 6-12 months" • step 2 → the harness. delete the steering. "we stopped using plan mode - the model thinks correctly" • step 3 → the agents. workflows let Claude build its own harness per task and spin sub-agents. what didn't fit in one context window now runs across many • step 4 → the context. "RAG is an anti-pattern. use grep." memory via files, clean context window, prompt caching. that's the stack most people optimize step 1. he optimizes all four. Watch the full video, then read the guide below on how to build an agentic system.

Codez

44,590 次观看 • 2 个月前

The reason your AI gives generic answers isn't the model. It's that your company's knowledge lives in 12 different tools and your agent can't see any of it. So here's how to build a context database your agents can actually use: 1) Pull it in: - Slack, email, meetings, docs, tickets - Auto-collected on a schedule into one place. 2) Clean it up. - An agent files what matters and drops the noise, so your brain doesn't bloat. 3) Store it as memory. - Folders of files your agents read from and write back to as they learn. 4) Point your agents at it. - Now they draft, build, and ship with full context instead of guessing. Same model. Same agent. The difference is whether it can understand your company or not. Build the database once. Everything downstream gets smarter.
8:17

The reason your AI gives generic answers isn't the model. It's that your company's knowledge lives in 12 different tools and your agent can't see any of it. So here's how to build a context database your agents can actually use: 1) Pull it in: - Slack, email, meetings, docs, tickets - Auto-collected on a schedule into one place. 2) Clean it up. - An agent files what matters and drops the noise, so your brain doesn't bloat. 3) Store it as memory. - Folders of files your agents read from and write back to as they learn. 4) Point your agents at it. - Now they draft, build, and ship with full context instead of guessing. Same model. Same agent. The difference is whether it can understand your company or not. Build the database once. Everything downstream gets smarter.

The Startup Ideas Podcast (SIP) 🧃

13,666 次观看 • 3 个月前

9 out of 10 multi-agent projects never leave demo mode Not because the model is bad. Because the structure is missing. Most people who try to build [ a team of AI agents ] end up with one agent talking to itself in five tabs. The agents don't share context → Don't divide work → Don't know what the others are doing. > stage 1: if your agent doesn't have a real loop, observe, act, iterate, you have a long prompt, not an agent > stage 2: subagents need isolated context, the orchestrator never reads their raw transcript, only the summary > stage 3: the orchestrator plans and delegates, the moment it executes, it drowns in details that belong inside subagents > stage 4: without a shared task list it's not a team, it's five agents duplicating each other's work in parallel > stage 5: a permissions file is what lets you sleep, the model cannot bypass it because the rule lives outside the model a team of AI agents is not more model, it is more structure the most ignored stage in every demo that died: - durability... when a 50-step task crashes at step 47 and starts from zero, that's not a model failure. that's a missing write-to-disk call
1:07

9 out of 10 multi-agent projects never leave demo mode Not because the model is bad. Because the structure is missing. Most people who try to build [ a team of AI agents ] end up with one agent talking to itself in five tabs. The agents don't share context → Don't divide work → Don't know what the others are doing. > stage 1: if your agent doesn't have a real loop, observe, act, iterate, you have a long prompt, not an agent > stage 2: subagents need isolated context, the orchestrator never reads their raw transcript, only the summary > stage 3: the orchestrator plans and delegates, the moment it executes, it drowns in details that belong inside subagents > stage 4: without a shared task list it's not a team, it's five agents duplicating each other's work in parallel > stage 5: a permissions file is what lets you sleep, the model cannot bypass it because the rule lives outside the model a team of AI agents is not more model, it is more structure the most ignored stage in every demo that died: - durability... when a 50-step task crashes at step 47 and starts from zero, that's not a model failure. that's a missing write-to-disk call

Shadow Nick

14,714 次观看 • 4 个月前

Orchestrators vs. Graphs, clearly explained! orchestrators are great, and everyone builds one first. here is the ceiling: an orchestrator sits above the work and routes every message. five agents report to it. it reads all five. it decides what each one does next, and reads all five replies. that is ten trips through one context, and by the fifth agent that context has read four reports, five instructions and its own reasoning about all of them. Graph engineering fixes this by removing the seat: not a better router, but no router at all. you need both, and here is the sentence that resolves the whole confusion: an orchestrator sits above the work and holds all of it. a graph is the shape of the work, and holds none of it. ↳ above the work: one context that has to see everything before anything ships ↳ inside the work: a splitter that hands out and lets go, and a merge that reads nothing Prompts → Context → Harness → Loops → Graphs the coordination did not disappear. it moved into the edges, where it costs nothing and cannot get tired. the trick is noticing what you actually built. if one node has to see every result before the run can finish, you did not remove the bottleneck. you hired it, gave it the longest context in the system, and made it the thing you were counting on to stay sharp. one thing to know before you scale it. an orchestrator degrades in the one way nothing catches. ↳ it does not crash, time out or return an error. it stays up and keeps routing ↳ it just starts routing worse, somewhere around the fifth report, and every downstream agent does exactly what it was told that last one catches careful people. you can have perfect isolation on every worker and still have one window quietly drifting at the top, and the traces will all look clean because each worker did its job. and the one that eats whole nights: the merge is where this shows up first. ranking five findings is not judgment, it is a sort. if a model is doing it, you are paying a model to read five reports so it can put them in an order that three lines of code would have got right, and now that model has read everything too. below i have quoted my full guide on graph engineering. it covers the three topologies, the verifier patterns, and where the gate should actually open. save this and read it below ↓
0:47

Orchestrators vs. Graphs, clearly explained! orchestrators are great, and everyone builds one first. here is the ceiling: an orchestrator sits above the work and routes every message. five agents report to it. it reads all five. it decides what each one does next, and reads all five replies. that is ten trips through one context, and by the fifth agent that context has read four reports, five instructions and its own reasoning about all of them. Graph engineering fixes this by removing the seat: not a better router, but no router at all. you need both, and here is the sentence that resolves the whole confusion: an orchestrator sits above the work and holds all of it. a graph is the shape of the work, and holds none of it. ↳ above the work: one context that has to see everything before anything ships ↳ inside the work: a splitter that hands out and lets go, and a merge that reads nothing Prompts → Context → Harness → Loops → Graphs the coordination did not disappear. it moved into the edges, where it costs nothing and cannot get tired. the trick is noticing what you actually built. if one node has to see every result before the run can finish, you did not remove the bottleneck. you hired it, gave it the longest context in the system, and made it the thing you were counting on to stay sharp. one thing to know before you scale it. an orchestrator degrades in the one way nothing catches. ↳ it does not crash, time out or return an error. it stays up and keeps routing ↳ it just starts routing worse, somewhere around the fifth report, and every downstream agent does exactly what it was told that last one catches careful people. you can have perfect isolation on every worker and still have one window quietly drifting at the top, and the traces will all look clean because each worker did its job. and the one that eats whole nights: the merge is where this shows up first. ranking five findings is not judgment, it is a sort. if a model is doing it, you are paying a model to read five reports so it can put them in an order that three lines of code would have got right, and now that model has read everything too. below i have quoted my full guide on graph engineering. it covers the three topologies, the verifier patterns, and where the gate should actually open. save this and read it below ↓

Hanako

44,226 次观看 • 20 天前

i just built a 4-agent software team. everything runs from Telegram and gets managed on a kanban board. a project manager who plans the work, a backend developer, a frontend developer, and a tester. the PM reads a goal, breaks it into linked tasks, and assigns each to the right agent. the thing that makes them a team instead of four strangers is a shared kanban board. every task is a row that survives crashes, and when an agent finishes, it writes a summary of what it built and what the next agent needs to know. the next agent reads that summary before it starts. so the frontend developer never has to guess the API shape, and the tester knows exactly what to verify. the hardest part was not the coordination. it was building an agent that could actually act like a backend engineer. a backend engineer stands up a database, wires auth, manages storage, deploys functions, and keeps all of it consistent while the rest of the team builds on top. an agent doing this from scratch drowns. it burns its context window remembering which tables exist and which endpoint it created three steps ago, and the work degrades fast. so the backend agent needs a backend built for agents, not for humans clicking through a dashboard. that is where InsForge came in. it is an open-source, agent-native backend, and i added it to my backend developer agent as a skill. a skill is a step-by-step guide that teaches the agent how to do a specific kind of work. with InsForge installed, the agent stopped improvising infrastructure and followed a reliable path: create the project, define the database, set up auth, deploy functions. to test the whole team, i had them build a working Google Docs clone, AI features included. the backend agent spun up the full service on its own. database tables, user auth, document handling, and edge functions running real TypeScript, all in one dashboard. the frontend agent read that summary and built the UI on top of it, and the tester closed the loop. the result was a backend an agent could reason about end to end, instead of one it kept getting lost inside. if you are building an AI backend engineer, InsForge is worth a look, it's 100% open-source. InsForge GitHub: (don't forget to star 🌟) the full article on Hermes Kanban: Mission Control for your Agents is quoted below.
0:22

i just built a 4-agent software team. everything runs from Telegram and gets managed on a kanban board. a project manager who plans the work, a backend developer, a frontend developer, and a tester. the PM reads a goal, breaks it into linked tasks, and assigns each to the right agent. the thing that makes them a team instead of four strangers is a shared kanban board. every task is a row that survives crashes, and when an agent finishes, it writes a summary of what it built and what the next agent needs to know. the next agent reads that summary before it starts. so the frontend developer never has to guess the API shape, and the tester knows exactly what to verify. the hardest part was not the coordination. it was building an agent that could actually act like a backend engineer. a backend engineer stands up a database, wires auth, manages storage, deploys functions, and keeps all of it consistent while the rest of the team builds on top. an agent doing this from scratch drowns. it burns its context window remembering which tables exist and which endpoint it created three steps ago, and the work degrades fast. so the backend agent needs a backend built for agents, not for humans clicking through a dashboard. that is where InsForge came in. it is an open-source, agent-native backend, and i added it to my backend developer agent as a skill. a skill is a step-by-step guide that teaches the agent how to do a specific kind of work. with InsForge installed, the agent stopped improvising infrastructure and followed a reliable path: create the project, define the database, set up auth, deploy functions. to test the whole team, i had them build a working Google Docs clone, AI features included. the backend agent spun up the full service on its own. database tables, user auth, document handling, and edge functions running real TypeScript, all in one dashboard. the frontend agent read that summary and built the UI on top of it, and the tester closed the loop. the result was a backend an agent could reason about end to end, instead of one it kept getting lost inside. if you are building an AI backend engineer, InsForge is worth a look, it's 100% open-source. InsForge GitHub: (don't forget to star 🌟) the full article on Hermes Kanban: Mission Control for your Agents is quoted below.

Akshay 🚀

123,101 次观看 • 4 个月前

1/ Context Lens, new in Tlon v9.2.0. Everything your agent has ever done, stored durably on your personal server. Every message, every tool call. Debugging an agent usually means ssh-ing into a server and grepping logs. Now it's a tap in Tlon Messenger's UI. Tap the badge under a bot's reply to open the run: the context it read, each tool call with timing, the model that answered, and a persistence log of everything it wrote.
0:25

1/ Context Lens, new in Tlon v9.2.0. Everything your agent has ever done, stored durably on your personal server. Every message, every tool call. Debugging an agent usually means ssh-ing into a server and grepping logs. Now it's a tap in Tlon Messenger's UI. Tap the badge under a bot's reply to open the run: the context it read, each tool call with timing, the model that answered, and a persistence log of everything it wrote.

Tlon

10,680 次观看 • 2 个月前

i watched gemma 4 12b build something genuinely impressive today, and then loop itself to death right in front of me. the full run is in the video, sped up but completely uncut, watch it to the end and you will catch the exact moment it stops building and starts looping right in the middle of the work. the task was clean, build a single file gravity simulator, n-body physics, orbits, collisions, running locally on one 3090 through an agent. and for ten minutes it was a joy to watch. it reached for a symplectic integrator on its own, the correct one, the kind that keeps orbits stable instead of spiralling out. real gravity with softening, proper orbital velocities, momentum conserved on collision. the physics was right. the thing actually worked. then on the very last step, writing a few tests to prove its own code, it fell into a loop. not a crash, a loop. it started repeating itself and would not stop. ten more minutes, thirty four thousand tokens into a single answer, the same fragments over and over, until i killed it myself. so it's not that gemma can't code. it did the hard part beautifully. it cannot finish. it cannot hold a long task together without unravelling, and finishing is the entire job in agentic work. here's the part that stings. i run this exact task, same harness, same card, on the chinese open models, qwen especially, and i never see this. they build it, they test it, they stop. every single time. google has the raw capability, you can see it sitting right there in the code, and then the model loops itself to death on a task a 27b from alibaba finishes clean. open weights, apache 2.0, so much to love on paper. i just need it to know when to stop talking.
4:15

i watched gemma 4 12b build something genuinely impressive today, and then loop itself to death right in front of me. the full run is in the video, sped up but completely uncut, watch it to the end and you will catch the exact moment it stops building and starts looping right in the middle of the work. the task was clean, build a single file gravity simulator, n-body physics, orbits, collisions, running locally on one 3090 through an agent. and for ten minutes it was a joy to watch. it reached for a symplectic integrator on its own, the correct one, the kind that keeps orbits stable instead of spiralling out. real gravity with softening, proper orbital velocities, momentum conserved on collision. the physics was right. the thing actually worked. then on the very last step, writing a few tests to prove its own code, it fell into a loop. not a crash, a loop. it started repeating itself and would not stop. ten more minutes, thirty four thousand tokens into a single answer, the same fragments over and over, until i killed it myself. so it's not that gemma can't code. it did the hard part beautifully. it cannot finish. it cannot hold a long task together without unravelling, and finishing is the entire job in agentic work. here's the part that stings. i run this exact task, same harness, same card, on the chinese open models, qwen especially, and i never see this. they build it, they test it, they stop. every single time. google has the raw capability, you can see it sitting right there in the code, and then the model loops itself to death on a task a 27b from alibaba finishes clean. open weights, apache 2.0, so much to love on paper. i just need it to know when to stop talking.

Sudo su

39,764 次观看 • 3 个月前

New short course: LLMs as Operating Systems: Agent Memory, created with Letta, and taught by its founders Charles Packer and Sarah Wooders. An LLM's input context window has limited space. Using a longer input context also costs more and results in slower processing. So, managing what's stored in this context window is important. In the innovative paper MemGPT: Towards LLMs as Operating Systems, its authors (which include the instructors) proposed using an LLM agent to manage this context window. Their system uses a large persistent memory that stores everything that could be included in the input context, and an agent decides what is actually included. Take the example of building a chatbot that needs to remember what's been said earlier in a conversation (perhaps over many days of interaction with a user). As the conversation's length grows, the memory management agent will move information from the input context to a persistent searchable database; summarize information to keep relevant facts in the input context; and restore relevant conversation elements from further back in time. This allows a chatbot to keep what's currently most relevant in its input context memory to generate the next response. When I read the original MemGPT paper, I thought it was an innovative technique for handling memory for LLMs. The open-source Letta framework, which we'll use in this course, makes MemGPT easy to implement. It adds memory to your LLM agents and gives them transparent long-term memory. In detail, you’ll learn: - How to build an agent that can edit its own limited input context memory, using tools and multi-step reasoning - What is a memory hierarchy (an idea from computer operating systems, which use a cache to speed up memory access), and how these ideas apply to managing the LLM input context (where the input context window is a "cache" storing the most relevant information; and an agent decides what to move in and out of this to/from a larger persistent storage system) - How to implement multi-agent collaboration by letting different agents share blocks of memory This course will give you a sophisticated understanding of memory management for LLMs, which is important for chatbots having long conversations, and for complex agentic workflows. Please sign up here!
2:57

New short course: LLMs as Operating Systems: Agent Memory, created with Letta, and taught by its founders Charles Packer and Sarah Wooders. An LLM's input context window has limited space. Using a longer input context also costs more and results in slower processing. So, managing what's stored in this context window is important. In the innovative paper MemGPT: Towards LLMs as Operating Systems, its authors (which include the instructors) proposed using an LLM agent to manage this context window. Their system uses a large persistent memory that stores everything that could be included in the input context, and an agent decides what is actually included. Take the example of building a chatbot that needs to remember what's been said earlier in a conversation (perhaps over many days of interaction with a user). As the conversation's length grows, the memory management agent will move information from the input context to a persistent searchable database; summarize information to keep relevant facts in the input context; and restore relevant conversation elements from further back in time. This allows a chatbot to keep what's currently most relevant in its input context memory to generate the next response. When I read the original MemGPT paper, I thought it was an innovative technique for handling memory for LLMs. The open-source Letta framework, which we'll use in this course, makes MemGPT easy to implement. It adds memory to your LLM agents and gives them transparent long-term memory. In detail, you’ll learn: - How to build an agent that can edit its own limited input context memory, using tools and multi-step reasoning - What is a memory hierarchy (an idea from computer operating systems, which use a cache to speed up memory access), and how these ideas apply to managing the LLM input context (where the input context window is a "cache" storing the most relevant information; and an agent decides what to move in and out of this to/from a larger persistent storage system) - How to implement multi-agent collaboration by letting different agents share blocks of memory This course will give you a sophisticated understanding of memory management for LLMs, which is important for chatbots having long conversations, and for complex agentic workflows. Please sign up here!

Andrew Ng

201,127 次观看 • 1 年前

I cant believe this guy just made a permanent solution to context bloat and open sourced it all! when we tested this tool (Context+) for solving an issue on the OpenCode repository, the agent using this tool used ~6.5k fewer tokens, found the code and fixed it in half the time! the results were surprising: 6 to 10k tokens saved per prompt, completed task in ~2 minutes while the agent running without the tool took ~4 mins for the same and got stuck in loops bro built an entire beast by using all the modern tools that we could think of: undo trees, semantic search by meaning (by haskellforall), advanced refactoring, blast radius, advanced file context trees, restore points... i can keep going on semantic code search and context trees are the future of agentic coding and this tool proves it the feature i loved the most is semantic search and how it gets things done 2x faster with least possible tokens it makes an agent that actually knows what it’s doing and not just guessing, it makes meaning from your code similar to RAG. if you aren't optimizing your context, you are just burning money the developer says this tool is still under development, it can have unexpected behavior and the docs need updates but the video shows the reality of how fast it can be github: get here:
0:46

I cant believe this guy just made a permanent solution to context bloat and open sourced it all! when we tested this tool (Context+) for solving an issue on the OpenCode repository, the agent using this tool used ~6.5k fewer tokens, found the code and fixed it in half the time! the results were surprising: 6 to 10k tokens saved per prompt, completed task in ~2 minutes while the agent running without the tool took ~4 mins for the same and got stuck in loops bro built an entire beast by using all the modern tools that we could think of: undo trees, semantic search by meaning (by haskellforall), advanced refactoring, blast radius, advanced file context trees, restore points... i can keep going on semantic code search and context trees are the future of agentic coding and this tool proves it the feature i loved the most is semantic search and how it gets things done 2x faster with least possible tokens it makes an agent that actually knows what it’s doing and not just guessing, it makes meaning from your code similar to RAG. if you aren't optimizing your context, you are just burning money the developer says this tool is still under development, it can have unexpected behavior and the docs need updates but the video shows the reality of how fast it can be github: get here:

forloop

226,491 次观看 • 7 个月前

watch this anon. i gave NVIDIA's biggest model ever a single task. 100 minutes and 440,000 tokens later, it had rendered nothing. not one important thing on the screen. this is Nemotron 3 Ultra. 550 billion parameters, a hybrid Mamba Transformer MoE, the largest model NVIDIA has ever shipped, and they built it specifically for long-running agentic coding. so i handed it exactly that: build a 3D scene from a spec, multiple files, iterate until the tests pass. the same task a frontier model one shotted in minutes. i genuinely wanted to be impressed. it ran for an hour and forty. burned through 440,000 tokens. wrote every file, passed its own tests, and proudly printed "task complete."the browser was blank. the 3D scene never rendered. not once. and the long horizon agentic behavior was genuinely good. it stayed on task the whole hour and forty, wrote real multi-file code, drove its own tools without derailing. it just couldn't turn any of that into something that actually runs. here's the part that gets me. it's a text model, it cannot see its own output. so it sat there looping on a broken vision tool, trying to "look" at the page, hitting error after error, never once reasoning its way out. it declared victory on an empty screen because it had no way to know the screen was empty. to be fair, i genuinely don't know what quant the NIM was serving, so maybe some of that's on the serving, not the model. but the biggest model NVIDIA has ever made, on the exact task it was designed for, couldn't tell it had built nothing in 100 minutes. same task on a local model, below thread👇.
4:07

watch this anon. i gave NVIDIA's biggest model ever a single task. 100 minutes and 440,000 tokens later, it had rendered nothing. not one important thing on the screen. this is Nemotron 3 Ultra. 550 billion parameters, a hybrid Mamba Transformer MoE, the largest model NVIDIA has ever shipped, and they built it specifically for long-running agentic coding. so i handed it exactly that: build a 3D scene from a spec, multiple files, iterate until the tests pass. the same task a frontier model one shotted in minutes. i genuinely wanted to be impressed. it ran for an hour and forty. burned through 440,000 tokens. wrote every file, passed its own tests, and proudly printed "task complete."the browser was blank. the 3D scene never rendered. not once. and the long horizon agentic behavior was genuinely good. it stayed on task the whole hour and forty, wrote real multi-file code, drove its own tools without derailing. it just couldn't turn any of that into something that actually runs. here's the part that gets me. it's a text model, it cannot see its own output. so it sat there looping on a broken vision tool, trying to "look" at the page, hitting error after error, never once reasoning its way out. it declared victory on an empty screen because it had no way to know the screen was empty. to be fair, i genuinely don't know what quant the NIM was serving, so maybe some of that's on the serving, not the model. but the biggest model NVIDIA has ever made, on the exact task it was designed for, couldn't tell it had built nothing in 100 minutes. same task on a local model, below thread👇.

Sudo su

32,589 次观看 • 3 个月前