Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

It doesn’t matter how fast your code is generated if the requirements are unclear. Shit in == shit out. That’s why we designed Software Factory as a discipline, not as a convenience tool. Our agents scrutinize your requirements and designs so that when you pass that context to a... show more

Chamath Palihapitiya

2,425,254 subscribers

305,609 views • 8 months ago •via X (Twitter)

Education Science & Technology

Anya Rossi• Live Now

Private livecam show

41 Comments

Chamath Palihapitiya8 months ago

Measure twice, cut once…

Bryan Shankman8 months ago

I mean this with zero malice: if you’re a billionaire, why are you spending time working on an AI SaaS?

Chamath Palihapitiya8 months ago

I like trying to help make useful things when I think they could matter. I think making a factory for people to use as we go through the AI revolution feels like one of those moments. Whether it will have been a good use of time will only be known after the fact…

Bryan Shankman8 months ago

Fair enough!

🧊 ICE MAXI 🧊8 months ago

genuine question... what's stopping me from running gpt + gemini deep research on my requirements then feeding the security gaps to claude code? already paying $400/mo for these tools...why add another layer?

Chamath Palihapitiya8 months ago

Nothing is stopping you. If it works for you, keep with it.

🧊 ICE MAXI 🧊8 months ago

curious - where can i send you my decks ?

Davit8 months ago

Planning is as critical as execution with coding agents. We found same insights for spec driven development. We built our own software factory for optimizing database software.

Adi Udupa8 months ago

“Shit in, shit out” is true; but most requirements are unclear because teams avoid making tradeoffs.

Bruce Dando | Finance Transformation8 months ago

How does this challenge the requirement when often the individual providing the requirement is unclear. Does this require a prompt engineer (middle man)? Replacing workers not completely with AI but with someone who knows how to use AI?

Chamath Palihapitiya8 months ago

It allows many people to swarm the requirements phase so that knowledge quality goes up. We designed this to be used in a broad team.

Bruce Dando | Finance Transformation8 months ago

I’ve got a project to test this on. Who do I talk to.

nick8 months ago

we've been saying 'shit in, shit out' for decades but finally having an agent with the spine to say it back to the stakeholders is the real productivity hack

RajUnfiltered8 months ago

Exactly bruhhh.... Speed doesn’t fix confusion. If requirements are fuzzy, faster code just means faster bugs... and the hard part of software was never typing code, it’s thinking clearly, agreeing on intent, and locking decisions before execution. Automation without clarity just scales mistakes...

WilcosX.eth8 months ago

curious how deep the scrutiny goes

Brion Hickey8 months ago

@chamath have you guys discussed Ralph Wiggum Loop for the product ?

jason 💬8 months ago

yeah my team once wasted three weeks debugging ai-generated code that perfectly executed the wrong idea

Pinecone8 months ago

This seems cool! I spend most of my time writing and refining specs. Having at least a substantial portion of the architecture and code written and reviewed spec files is the only way I've been able to get consistently good results.

LovelaceHopeful8 months ago

I still prefer the term GiGo

Matt Beebe8 months ago

Ownership of generated code? Portability or platform lock-in? Price of tokens for planning and coding?

Erich8 months ago

Cant this be created into a claude skill?

Catalin Pop8 months ago

Any thoughts on making individual subscription? Startups?

033.eth8 months ago

is there significance to the number 8090?

Tung Nguyen8 months ago

Precisely. And this is the most underrated thing about prompting. We’ve been given ai tools with the promise that it can solve all our problems but we have limited training and know-how on using the tools optimally. Tools with built-in checks points and quality gates makes perfect sense to me. It’s always one step ahead in the pursuit of delivering actual value. The tool should know what I want to do and recognize I do not know how to arrive to the desired outcome. Ai should be curious about what the operator is trying to do. And make itself useful in solving the proposed problem. Much like how humans (some) are intrinsically curious and are problem solvers.

Doing Ain't Easy8 months ago

Good design something to look at market makers dark pools and prime lenders. Yiu know the racket. You've made enough. Liquid some assets for 2 generation and be a hero we all need. CAT System now!

arti8 months ago

Change the name to much overlap with @FactoryAI

Matt Cunningham8 months ago

You guys should build this as a wrapper around the coding cli tools like codex so that developers can use their credits to power these features.

Mike Dean8 months ago

I was hoping for an open source version tbh

Mentis 🇦🇺8 months ago

Been saying this for twenty years. No amount of tooling fixes a bad spec.

Jeryd8 months ago

Where is 8090.molt?

Life Alkemy8 months ago

When do you think all the team sitting in offshore ODCs looking and handling all the SDLC of enterprise software code shall be downsized to single digit team members? I am just thinking on the lines when this kind of 100% automation seep in SDLC liberating teams holed up in desk 24x7, yeah sort of going mainstream for enterprise code

Brett Bringardner8 months ago

I have two favorite tee shirts. MicroTrash why are you still buying garbage ? And MicroCrap why are you still buying shit ?

Eddie8 months ago

so because your requirements are vague, you blame the code? sounds like you need better specs, not a magic tool. reality check: garbage in, garbage out.

Brendan Booth8 months ago

AI didn’t eliminate the need for thinking - it made thinking the constraint. Faster code just amplifies whatever clarity (or confusion) exists upstream.

Daxesh8 months ago

Treating software as a process rather than just a tool is the correct way to think about it.

Guo Jun8 months ago

Clear thinking first, code second.

Zoe Sterling8 months ago

me explaining my feature requirements: "make it good but like, you know what I mean" claude: *generates 47 different interpretations of "good"*

Bren8 months ago

This can be done with a claude SKILL

Ali Mamak8 months ago

Speed is cheap. Clarity is rare. Requirements are the contract between imagination and reality. If you cannot explain the outcome in plain language, you are shipping guesses. Code only amplifies whatever you asked for.

Karl Wirth8 months ago

$200/person/month? SAAS? Do this with Git, $20/person/month Claude Code and @nimbalyst Claude Skills, Local, no lockin, open standards.

EJ8 months ago

codex drop just mogged.

Related Videos

AI power users are here and the case for agent-friendly software is materializing. Aaron Levie on agents outnumbering people 1000x: "If you start to imagine that we all have to build software for agents… we spend as much time now thinking about the agent interface to our tool as we do the human interface." "If you have a hundred or a thousand times more agents than people, then your software has to be built for agents… it's going to be through an API or a CLI or MCP." "What if you give a coding agent access to your SaaS tools and a coding agent access to your knowledge work, workflows and context?" "It can actually code its way or use APIs through whatever task it's trying to achieve." "That appears to be like a paradigm that is starting to compound… and I actually think it kind of makes sense as the ultimate manifestation of this stuff." Aaron Levie

AI power users are here and the case for agent-friendly software is materializing. Aaron Levie on agents outnumbering people 1000x: "If you start to imagine that we all have to build software for agents… we spend as much time now thinking about the agent interface to our tool as we do the human interface." "If you have a hundred or a thousand times more agents than people, then your software has to be built for agents… it's going to be through an API or a CLI or MCP." "What if you give a coding agent access to your SaaS tools and a coding agent access to your knowledge work, workflows and context?" "It can actually code its way or use APIs through whatever task it's trying to achieve." "That appears to be like a paradigm that is starting to compound… and I actually think it kind of makes sense as the ultimate manifestation of this stuff." Aaron Levie

a16z

42,152 views • 2 months ago

Software leverage doesn’t come from faster typing. It comes from stronger coordination. Software Factory is a multiplayer platform: shared requirements, blueprints, and work orders flow to coding agents via MCP. Alignment first. Then parallel execution. Try it:

Software leverage doesn’t come from faster typing. It comes from stronger coordination. Software Factory is a multiplayer platform: shared requirements, blueprints, and work orders flow to coding agents via MCP. Alignment first. Then parallel execution. Try it:

Chamath Palihapitiya

142,969 views • 8 months ago

No, it doesn’t. It doesn’t look to you like he did it in self-defense, you know it doesn’t. You are lying to the American people straight to our faces as if we are actually supposed to eat your shit. We are not Mike. You’re a pathetic man with no honor. That’s all there is to it.

No, it doesn’t. It doesn’t look to you like he did it in self-defense, you know it doesn’t. You are lying to the American people straight to our faces as if we are actually supposed to eat your shit. We are not Mike. You’re a pathetic man with no honor. That’s all there is to it.

MAGA Cult Slayer🦅🇺🇸

42,403 views • 9 months ago

Introducing Inspector's context-aware screenshot tool! Screenshots as context for your coding agent is a pain. I decided I was tired of it and shipped something today to help with that.

Introducing Inspector's context-aware screenshot tool! Screenshots as context for your coding agent is a pain. I decided I was tired of it and shipped something today to help with that.

mikeBuildsMore

21,965 views • 11 months ago

Bash is all you need! Which is why I'm introducing my holiday project: just-bash just-bash is a pretty complete implementation of bash in TypeScript designed to be used as a bash tool by AI agents. Because it turns out agents love exploring data via shell scripts, even beyond coding. It comes with grep, sed, awk and the 99th percentile features that an agent like Claude Code or Cursor would use. In fact, Claude Code can use it for secure bash execution. In the package - A bash-tool for AI SDK - A binary for use by yourself or your coding agents - An overlay filesystem to feed files to your agent securely - A Vercel Sandbox compatible API, so you can quickly upgrade to a real VM if you need to run binaries - An example AI agent that explores the just-bash code base using just-bash - I imported the Oils shell bash compatibility suite and just-bash passes a very good chunk What is interesting about this codebase: It was essentially entirely written by Opus 4.5. Coding agents love bash and they are good at reproducing it. They are also great at text-book recursive descent parsers and AST tweet-walk interpreters. That said, it is, like, a lot of code and I didn't read it all 😅. This is very much a hack, but it also seems to be _really_ useful. I haven't really found anything agents want to use that it doesn't support and it's fast and secure (caveats apply). It doesn't have write access to your computer and the filesystem is given a root that the agent cannot escape from. Find it at Related: Our recent blog post how we migrated our data analysis agent to bash tools and achieved incredible quality improvements The video shows the example agent investigating the just-bash code base

Bash is all you need! Which is why I'm introducing my holiday project: just-bash just-bash is a pretty complete implementation of bash in TypeScript designed to be used as a bash tool by AI agents. Because it turns out agents love exploring data via shell scripts, even beyond coding. It comes with grep, sed, awk and the 99th percentile features that an agent like Claude Code or Cursor would use. In fact, Claude Code can use it for secure bash execution. In the package - A bash-tool for AI SDK - A binary for use by yourself or your coding agents - An overlay filesystem to feed files to your agent securely - A Vercel Sandbox compatible API, so you can quickly upgrade to a real VM if you need to run binaries - An example AI agent that explores the just-bash code base using just-bash - I imported the Oils shell bash compatibility suite and just-bash passes a very good chunk What is interesting about this codebase: It was essentially entirely written by Opus 4.5. Coding agents love bash and they are good at reproducing it. They are also great at text-book recursive descent parsers and AST tweet-walk interpreters. That said, it is, like, a lot of code and I didn't read it all 😅. This is very much a hack, but it also seems to be _really_ useful. I haven't really found anything agents want to use that it doesn't support and it's fast and secure (caveats apply). It doesn't have write access to your computer and the filesystem is given a root that the agent cannot escape from. Find it at Related: Our recent blog post how we migrated our data analysis agent to bash tools and achieved incredible quality improvements The video shows the example agent investigating the just-bash code base

Malte Ubl

125,326 views • 10 months ago

YC F24's @ManicodeAI is a CLI tool that writes code for you. It uses your codebase as context and edits your files automatically. Try it out: `npm i -g manicode` Congrats on the launch, James Grugett and Brandon Chen!

YC F24's @ManicodeAI is a CLI tool that writes code for you. It uses your codebase as context and edits your files automatically. Try it out: `npm i -g manicode` Congrats on the launch, James Grugett and Brandon Chen!

Y Combinator

277,316 views • 2 years ago

What is the “dark factory” that there’s so much talk about, in relation to software factories? dex, founder of HumanLayer explains: “You can imagine a car factory where it’s all robots building the cars, and they don’t have lights. There’s no humans. You walk in there, there’s no lights, there’s not even light switches. And in software, you can have mini loops that are dark in your thing: hey, if the code review agent comes back with a problem, you loop that back to the builder agent, it fixes it and comes back, and that’s a dark factory. You don’t need a human loop for that. But the full dark factory where you don’t read any code: yeah, it’s a good way to maximize your token utilization. If your belief is that ‘my job is to extract as much intelligence out of the machine as I can’, then you go token harder. And my take is token smarter: how do I get as much value out of AI as I can without having to turn the lights off, while still maintaining control and taste and judgment and understanding the system architecture?”

What is the “dark factory” that there’s so much talk about, in relation to software factories? dex, founder of HumanLayer explains: “You can imagine a car factory where it’s all robots building the cars, and they don’t have lights. There’s no humans. You walk in there, there’s no lights, there’s not even light switches. And in software, you can have mini loops that are dark in your thing: hey, if the code review agent comes back with a problem, you loop that back to the builder agent, it fixes it and comes back, and that’s a dark factory. You don’t need a human loop for that. But the full dark factory where you don’t read any code: yeah, it’s a good way to maximize your token utilization. If your belief is that ‘my job is to extract as much intelligence out of the machine as I can’, then you go token harder. And my take is token smarter: how do I get as much value out of AI as I can without having to turn the lights off, while still maintaining control and taste and judgment and understanding the system architecture?”

The Pragmatic Engineer

37,636 views • 3 months ago

LAPD: Before we do anything here, we have to get a search warrant because there are issues regarding standing at this residence. So as long as those legal requirements are met, we can continue the investigation. Now, when I tell you that we have not identified a suspect, it doesn't mean that we're not working the case. We will not identify a suspect until that person is identified as a suspect. And we're legally seeking them. If there is a crime to be reported, and if that person is a suspect in that crime, did that.

LAPD: Before we do anything here, we have to get a search warrant because there are issues regarding standing at this residence. So as long as those legal requirements are met, we can continue the investigation. Now, when I tell you that we have not identified a suspect, it doesn't mean that we're not working the case. We will not identify a suspect until that person is identified as a suspect. And we're legally seeking them. If there is a crime to be reported, and if that person is a suspect in that crime, did that.

Acyn

607,579 views • 10 months ago

introducing Pinterest in Krea Agent. now you can bring all your pins and boards into the tool and use them as context for your agents. try it now at krea . ai / agent

introducing Pinterest in Krea Agent. now you can bring all your pins and boards into the tool and use them as context for your agents. try it now at krea . ai / agent

Krea

15,409 views • 1 month ago

New course: Spec-Driven Development with Coding Agents, built in partnership with JetBrains, and taught by Paul Everitt | @pauleveritt@fosstodon.org. Vibe coding is fast, but often produces code that doesn't match what you asked for. This short course teaches you spec-driven development: write a detailed spec defining what to build, and work with your coding agent to implement it. Many of the best developers already build this way. A spec lets you control large code changes with a few words, preserve context across agent sessions, and stay in control as your project grows in complexity. Skills you'll gain: - Write a detailed specification to define your mission, tech stack, and roadmap, giving your agent the context it needs from the start - Plan, implement, and validate features in iterative loops using a spec as your agent's guide - Apply the same repeatable workflow to both new and legacy codebases - Package your workflow into a portable agent skill that works across agents and IDEs Join and write specs that keep your coding agent on track!

New course: Spec-Driven Development with Coding Agents, built in partnership with JetBrains, and taught by Paul Everitt | @[email protected]. Vibe coding is fast, but often produces code that doesn't match what you asked for. This short course teaches you spec-driven development: write a detailed spec defining what to build, and work with your coding agent to implement it. Many of the best developers already build this way. A spec lets you control large code changes with a few words, preserve context across agent sessions, and stay in control as your project grows in complexity. Skills you'll gain: - Write a detailed specification to define your mission, tech stack, and roadmap, giving your agent the context it needs from the start - Plan, implement, and validate features in iterative loops using a spec as your agent's guide - Apply the same repeatable workflow to both new and legacy codebases - Package your workflow into a portable agent skill that works across agents and IDEs Join and write specs that keep your coding agent on track!

Andrew Ng

466,985 views • 6 months ago

The truth doesn’t care about your opinion,ladies stop thinking that men are trying to hate on you , believe it or not we care that’s why I’m posting this. We want you smelling good staying in shape and as natural as possible. We’re not gonna do the high school shit today about what about the man and vice versa. My job is to give you the information Is it to you how you process it✌🏾

The truth doesn’t care about your opinion,ladies stop thinking that men are trying to hate on you , believe it or not we care that’s why I’m posting this. We want you smelling good staying in shape and as natural as possible. We’re not gonna do the high school shit today about what about the man and vice versa. My job is to give you the information Is it to you how you process it✌🏾

TK Kirkland

11,278 views • 2 years ago

OpenAI just released Codex Security, an AI agent that scans software projects to fix vulnerabilities while ignoring harmless bugs. Testing on 1.2mn commits found 792 critical flaws and dropped false alarms by 50%. Here is how the Codex Security agent works Initially, it scans your software project to learn how all the different pieces of code connect. It uses that knowledge to build a threat model, which acts as a custom map showing exactly where your software might be exposed to hackers. The agent then hunts for potential security holes based on that specific map. When it spots a possible flaw, it does not just send you a random alert. Instead, it creates a completely safe and isolated copy of your system known as a sandbox. The agent then actively tries to break into that sandboxed copy to prove if the security weakness is a real danger. If the agent successfully exploits its own test environment, it knows the bug is a genuine threat and not a false alarm. After confirming the danger, the agent writes a custom software patch to fix the broken code. It then tests that new patch to guarantee the fix does not accidentally break any other features in your software.

OpenAI just released Codex Security, an AI agent that scans software projects to fix vulnerabilities while ignoring harmless bugs. Testing on 1.2mn commits found 792 critical flaws and dropped false alarms by 50%. Here is how the Codex Security agent works Initially, it scans your software project to learn how all the different pieces of code connect. It uses that knowledge to build a threat model, which acts as a custom map showing exactly where your software might be exposed to hackers. The agent then hunts for potential security holes based on that specific map. When it spots a possible flaw, it does not just send you a random alert. Instead, it creates a completely safe and isolated copy of your system known as a sandbox. The agent then actively tries to break into that sandboxed copy to prove if the security weakness is a real danger. If the agent successfully exploits its own test environment, it knows the bug is a genuine threat and not a false alarm. After confirming the danger, the agent writes a custom software patch to fix the broken code. It then tests that new patch to guarantee the fix does not accidentally break any other features in your software.

Rohan Paul

13,208 views • 7 months ago

Software factories should be configured as code. With Warp, your factory lives in a repo as markdown and YAML files, so your agents, automations, runners, benchmarks, and even custom model routers are all in one place. Here, I walk through a typical Warp factory config and show how it lays the foundation for self improvement! It's pretty awesome DX.

Software factories should be configured as code. With Warp, your factory lives in a repo as markdown and YAML files, so your agents, automations, runners, benchmarks, and even custom model routers are all in one place. Here, I walk through a typical Warp factory config and show how it lays the foundation for self improvement! It's pretty awesome DX.

sathvik

12,741 views • 5 days ago

Enterprises today are running single-player AI in a multiplayer business. One engineer plus one IDE plus one agent ships lines of code faster. It does not ship coordinated software. Two desks over, another engineer's agent is shipping a change that contradicts last quarter's architecture decision. Neither agent knows. Compliance assembles the audit trail from screenshots after the fact. We built Software Factory to turn that into a team. Requirements captures the business intent in plain English. Blueprints maps the architecture as it actually exists. Work Orders translates that into structured tasks agents execute through MCP in Cursor, VS Code, or Claude Code. Tests validates the deployed code against the Requirement that started it. The Knowledge Graph wires it all together. Your engineers conduct the agents. The agents share state. The audit trail is part of the build, not a quarterly forensics exercise. EY's deployment: +70% productivity, 95%+ automated test coverage, months of roadmap compressed to days.

Enterprises today are running single-player AI in a multiplayer business. One engineer plus one IDE plus one agent ships lines of code faster. It does not ship coordinated software. Two desks over, another engineer's agent is shipping a change that contradicts last quarter's architecture decision. Neither agent knows. Compliance assembles the audit trail from screenshots after the fact. We built Software Factory to turn that into a team. Requirements captures the business intent in plain English. Blueprints maps the architecture as it actually exists. Work Orders translates that into structured tasks agents execute through MCP in Cursor, VS Code, or Claude Code. Tests validates the deployed code against the Requirement that started it. The Knowledge Graph wires it all together. Your engineers conduct the agents. The agents share state. The audit trail is part of the build, not a quarterly forensics exercise. EY's deployment: +70% productivity, 95%+ automated test coverage, months of roadmap compressed to days.

8090

75,406 views • 5 months ago

What is a software factory? It is your workflow, your skills, and your domain knowledge, saved in files that the agent reads so the whole process is productized Think about the last app you built. You typed what you wanted, and the agent built it. You did not like it, so you made changes. You shipped it to Vercel, or you kept going. A factory puts structure on each of those steps. At each step you ask how to get the most from the model, and how to move faster. That is the conveyor belt. You ship quickly, and the quality stays the same. Mine is 5 md files. It works in Codex, Claude Code, and Cursor, because it does not depend on the model or the tool. Build your own.

What is a software factory? It is your workflow, your skills, and your domain knowledge, saved in files that the agent reads so the whole process is productized Think about the last app you built. You typed what you wanted, and the agent built it. You did not like it, so you made changes. You shipped it to Vercel, or you kept going. A factory puts structure on each of those steps. At each step you ask how to get the most from the model, and how to move faster. That is the conveyor belt. You ship quickly, and the quality stays the same. Mine is 5 md files. It works in Codex, Claude Code, and Cursor, because it does not depend on the model or the tool. Build your own.

The Startup Ideas Podcast (SIP) 🧃

17,326 views • 1 month ago

Introducing Sarvam Voice Agents MCP. Describe the voice agent you want, and your coding tool builds it. The MCP connects your coding tool to Sarvam, so you can create agents, run campaigns, and analyse calls from the environment you already code in. Watch Sarvam Voice Agents MCP build a voice agent, call a list of people and report back. Build your first voice agent with the MCP:

Introducing Sarvam Voice Agents MCP. Describe the voice agent you want, and your coding tool builds it. The MCP connects your coding tool to Sarvam, so you can create agents, run campaigns, and analyse calls from the environment you already code in. Watch Sarvam Voice Agents MCP build a voice agent, call a list of people and report back. Build your first voice agent with the MCP:

Sarvam

18,939 views • 12 days ago

Today we're shipping Oasis Desktop, a Mac app that lets you share local coding agents with your teammates. Oasis Desktop can spin up and share Claude Code, Codex, Pi, OpenCode, Kilo Code, Hermes Agent, and Devin CLI sessions as long as they're installed. Start a session with your team. Bring in Claude Code and say, “What is wrong with auth. The tests have been failing since Tuesday. Help us figure it out.” The agent reads the files and runs the test suite directly on your laptop. When it asks, “Why did we replace the session middleware last week?” your teammate can answer in the exact same session. Completely free - try it out today:

Today we're shipping Oasis Desktop, a Mac app that lets you share local coding agents with your teammates. Oasis Desktop can spin up and share Claude Code, Codex, Pi, OpenCode, Kilo Code, Hermes Agent, and Devin CLI sessions as long as they're installed. Start a session with your team. Bring in Claude Code and say, “What is wrong with auth. The tests have been failing since Tuesday. Help us figure it out.” The agent reads the files and runs the test suite directly on your laptop. When it asks, “Why did we replace the session middleware last week?” your teammate can answer in the exact same session. Completely free - try it out today:

Oasis

215,176 views • 2 months ago

Have you always wanted to vibe code be found it overwhelming? We got you! Welcome to the first lesson of Ramping up with Replit! 👩🏻‍💻 Today we are going to build your first app. We understand that it may feel a bit hard to start your vibe coding journey, so we are going to show you that it is as simple as Do Re Mi 🎶

Have you always wanted to vibe code be found it overwhelming? We got you! Welcome to the first lesson of Ramping up with Replit! 👩🏻‍💻 Today we are going to build your first app. We understand that it may feel a bit hard to start your vibe coding journey, so we are going to show you that it is as simple as Do Re Mi 🎶

Replit ⠕

18,927 views • 8 months ago