←
godsystem62 posts
godsystem avatar

godsystem

@godsystemsmd

learning ai system loading... 89%

◷ Joined August 2026▧ 26 Photos and videos
godsystem avatar
godsystem@godsystemsmd·6h

Your bots don't need better prompts. They need a lane, a handoff and a rule SpaceXAI's Roman Ugarte, building a company out of Grok Bots in 72 hours: "You don't create a chat for each task. You create a bot for a unit of work" Lane is the one job a bot owns end to end, like your inbox or your outbound sales Handoff is bots messaging each other and waiting on each other's work, so nothing collides Rule is the line it can't cross, like no email leaves without your yes Anyone can chat with a bot. Almost nobody gives one a job it owns Save and watch it today Then save the guide below ↓

49:29

rari🈀 @0xwhrrari Aug 29 Article Harness Engineering: How to Build AI Agents That Don't Fall Apart Most people respond to a failing agent by changing the prompt Then they change the model Then they add a larger context window The agent still forgets decisions It still uses the wrong tool It still

godsystem avatar
⟳ godsystem retweeted
godsystem@godsystemsmd·Nov 17

"Bots persist. Chats don't." - Nick Miller, lead field engineer at SpaceXAI On stage he walked through the Grok Bot setup he runs his whole day on Every bot gets what a new hire gets: context, connectors and an outcome it's measured on Each keeps its own cloud computer, so it works with the laptop closed and remembers every correction Where it landed: 450,000 weekly users in its first month Watch this before you hire anyone to do it

27:37

rari🈀 @0xwhrrari Aug 29 Article Harness Engineering: How to Build AI Agents That Don't Fall Apart Most people respond to a failing agent by changing the prompt Then they change the model Then they add a larger context window The agent still forgets decisions It still uses the wrong tool It still

godsystem avatar
⟳ godsystem retweeted
godsystem@godsystemsmd·Nov 16

This is f*cking insane. Nobel Prize winner Demis Hassabis, the man who runs Google DeepMind, just called time on the 10-person startup team: "One person who really understands AI will outperform an entire startup team In 5 years, the people who build with AI will run whole companies with a tenth of the headcount" In this 1-hour lecture, he goes from founding DeepMind as an Apollo programme for AGI to AlphaFold, and explains why the next big companies will be tiny 10 People → 1 Person + Agents → A Company Run By Agents → A Tenth Of The Headcount The one-person company doesn't do less work, it hands the work to agents and checks what comes back The team of 10 is still hiring, the one person is already shipping Anyone can start a company, almost nobody is building the agent team that runs it Bookmark this before it disappears from your feed Then read how to build that agent team in the harness engineering guide below ↓

1:02:41

rari🈀 @0xwhrrari Aug 29 Article Harness Engineering: How to Build AI Agents That Don't Fall Apart Most people respond to a failing agent by changing the prompt Then they change the model Then they add a larger context window The agent still forgets decisions It still uses the wrong tool It still

godsystem avatar
⟳ godsystem retweeted
godsystem@godsystemsmd·Nov 15

This is f*cking insane. Nvidia founder Jensen Huang said he'd be "deeply alarmed" if a $500,000 engineer didn't burn half their salary in AI tokens: "If that $500,000 engineer did not consume at least $250,000 worth of tokens, I'm going to be deeply alarmed It's no different than one of our chip designers saying, I'm just going to use paper and pencil" In under 2.5 minutes, he explains why "too hard", "too long" and "we need more people" are thoughts that stop existing Writing Code → Writing Specs → Defining Good vs Bad → 100 Agents Per Engineer His version of the job: you don't code, you write the idea, the architecture and the spec Then you organize a team of agents and tell them what a great outcome looks like The companies that win next year won't hire more engineers, they'll give every engineer 100 agents Bookmark this before it disappears from your feed Then read how to run your own team of agents in the harness engineering guide below ↓

2:21

rari🈀 @0xwhrrari Aug 29 Article Harness Engineering: How to Build AI Agents That Don't Fall Apart Most people respond to a failing agent by changing the prompt Then they change the model Then they add a larger context window The agent still forgets decisions It still uses the wrong tool It still

godsystem avatar
⟳ godsystem retweeted
godsystem@godsystemsmd·Nov 14

This is f*cking insane. OpenAI president Greg Brockman just said they took 25% of their production engineers off every project and turned OpenAI's own models against OpenAI's own systems: "Sorry, all your projects are on hold. You are now defending. You're going to use the models to find all the holes" In 2 minutes, he explains the "defense factory": the models kept hunting until they found every critical hole they were smart enough to find Find → Triage → Fix → Deploy → Validate → New Model → Repeat His goal is that whole loop running end to end, at machine speed And the next step he describes is AI formally verifying all of software, something engineers dreamed about for decades and never pulled off The companies winning with AI aren't using it to write emails, they're running it like a factory Bookmark this, it's the playbook every company copies next year Then read how to build your own factory of agents in the harness engineering guide below ↓

2:14

rari🈀 @0xwhrrari Aug 29 Article Harness Engineering: How to Build AI Agents That Don't Fall Apart Most people respond to a failing agent by changing the prompt Then they change the model Then they add a larger context window The agent still forgets decisions It still uses the wrong tool It still

godsystem avatar
godsystem@godsystemsmd·Nov 14

This is f*cking insane. OpenAI president Greg Brockman just said they took 25% of their production engineers off every project and turned OpenAI's own models against OpenAI's own systems: "Sorry, all your projects are on hold. You are now defending. You're going to use the models to find all the holes" In 2 minutes, he explains the "defense factory": the models kept hunting until they found every critical hole they were smart enough to find Find → Triage → Fix → Deploy → Validate → New Model → Repeat His goal is that whole loop running end to end, at machine speed And the next step he describes is AI formally verifying all of software, something engineers dreamed about for decades and never pulled off The companies winning with AI aren't using it to write emails, they're running it like a factory Bookmark this, it's the playbook every company copies next year Then read how to build your own factory of agents in the harness engineering guide below ↓

2:14
godsystem avatar
⟳ godsystem retweeted
godsystem@godsystemsmd·Nov 13

This is f*cking insane. Perplexity CEO Aravind Srinivas: "There's one solopreneur, literally a one-person company, spending upwards of a million dollars a year on our product He's making $3 million a year by automating newsletters for others" In 78 seconds, he explains why the next big companies won't have employees, they'll have agents One Person → One Idea → A Team Of Agents → $3M A Year No office, no hires, no payroll, just agents doing the work of a whole team The people getting rich off AI right now aren't building models, they're running agents for other people Save this before it disappears from your feed Then read how to build your own team of agents in the harness engineering guide below ↓

1:20

rari🈀 @0xwhrrari Aug 29 Article Harness Engineering: How to Build AI Agents That Don't Fall Apart Most people respond to a failing agent by changing the prompt Then they change the model Then they add a larger context window The agent still forgets decisions It still uses the wrong tool It still

godsystem avatar
⟳ godsystem retweeted
godsystem@godsystemsmd·Nov 12

This is f*cking insane. Eric Schmidt ran Google for 10 years, and he just said the quiet part out loud about making money with AI: "If you really want to make money, it's actually easy, found an agentic AI company, I mean build an agent to do something If you build a better agent than anyone else, you can probably build one incredible company and own it yourself" In under 30 seconds, he explains why the next 1-2 years are the agentic window, and why it closes fast Prompts → Claude Code → Skills → Sub-Agents → Hooks → A Company You Own Yourself Save this before it disappears from your feed Everything below is free: Agent architecture langchain.com/blog?category_… Claude Code 101 anthropic.skilljar.com/claud… Claude Code in action anthropic.skilljar.com/claud… Prompt engineering platform.claude.com/docs/en/… Interactive prompt course github.com/anthropics/course… Claude.md + memory code.claude.com/docs/en/clau… Skills code.claude.com/docs/en/skil… Sub-agents code.claude.com/docs/en/sub-… Hooks code.claude.com/docs/en/hook… MCP code.claude.com/docs/en/mcp Routines code.claude.com/docs/en/rout… Agent SDK docs.claude.com/en/api/agent… Claude cookbooks github.com/anthropics/claude… Claude Code ultimate guide github.com/FlorianBruniaux/c… Awesome Claude Code github.com/hesreallyhim/awes… Official Claude Code docs code.claude.com/docs/en/over… All 16 of these cost $0 99% will spend the agentic window reading threads about it, the 1% who ship agents will own the companies that come out of it Bookmark it and open the first link tonight

0:29

rari🈀 @0xwhrrari Aug 29 Article Harness Engineering: How to Build AI Agents That Don't Fall Apart Most people respond to a failing agent by changing the prompt Then they change the model Then they add a larger context window The agent still forgets decisions It still uses the wrong tool It still

godsystem avatar
godsystem@godsystemsmd·Nov 12

This is f*cking insane. Eric Schmidt ran Google for 10 years, and he just said the quiet part out loud about making money with AI: "If you really want to make money, it's actually easy, found an agentic AI company, I mean build an agent to do something If you build a better agent than anyone else, you can probably build one incredible company and own it yourself" In under 30 seconds, he explains why the next 1-2 years are the agentic window, and why it closes fast Prompts → Claude Code → Skills → Sub-Agents → Hooks → A Company You Own Yourself Save this before it disappears from your feed Everything below is free: Agent architecture langchain.com/blog?category_… Claude Code 101 anthropic.skilljar.com/claud… Claude Code in action anthropic.skilljar.com/claud… Prompt engineering platform.claude.com/docs/en/… Interactive prompt course github.com/anthropics/course… Claude.md + memory code.claude.com/docs/en/clau… Skills code.claude.com/docs/en/skil… Sub-agents code.claude.com/docs/en/sub-… Hooks code.claude.com/docs/en/hook… MCP code.claude.com/docs/en/mcp Routines code.claude.com/docs/en/rout… Agent SDK docs.claude.com/en/api/agent… Claude cookbooks github.com/anthropics/claude… Claude Code ultimate guide github.com/FlorianBruniaux/c… Awesome Claude Code github.com/hesreallyhim/awes… Official Claude Code docs code.claude.com/docs/en/over… All 16 of these cost $0 99% will spend the agentic window reading threads about it, the 1% who ship agents will own the companies that come out of it Bookmark it and open the first link tonight

0:29
godsystem avatar
⟳ godsystem retweeted
godsystem@godsystemsmd·Nov 11

They would not have believed you" Then a friend told him: "Jeff, forget massage therapist. There are dog psychiatrists" He looked it up. Sure enough, you can easily hire one Farms → Factories → Internet → AI → Jobs That Don't Have Names Yet Everyone counts the jobs AI will take. Nobody can count the ones it creates, because they don't have names yet Save this before it disappears from your feed Everything below is free: AWS Skill Builder (free AI courses from Amazon) skillbuilder.aws AWS AI learning paths aws.amazon.com/ai/learn Kiro, Amazon's agentic coding IDE kiro.dev Strands Agents, Amazon's open-source agent SDK strandsagents.com The Strands harness SDK on GitHub github.com/strands-agents/ha… Strands agent examples you can copy github.com/strands-agents/sa… Bedrock AgentCore (run agents in production) aws.amazon.com/bedrock/agent… AgentCore samples github.com/awslabs/agentcore… Official AWS MCP servers github.com/awslabs/mcp Amazon Bedrock agent samples github.com/aws-samples/amazo… PartyRock (build AI apps with no code) partyrock.aws All of this costs $0 The next "dog psychiatrist" job will be invegents from these links, and in a few yearsit'll pay $200 an hour with a waitlist Bookmark it and start with the first link tonight Then read how to start building with agents in the harness engineering guide below ↓

0:40

rari🈀 @0xwhrrari Aug 29 Article Harness Engineering: How to Build AI Agents That Don't Fall Apart Most people respond to a failing agent by changing the prompt Then they change the model Then they add a larger context window The agent still forgets decisions It still uses the wrong tool It still

godsystem avatar
godsystem@godsystemsmd·Nov 11

watch it rn ↓

They would not have believed you" Then a friend told him: "Jeff, forget massage therapist. There are dog psychiatrists" He looked it up. Sure enough, you can easily hire one Farms → Factories → Internet → AI → Jobs That Don't Have Names Yet Everyone counts the jobs AI will take. Nobody can count the ones it creates, because they don't have names yet Save this before it disappears from your feed Everything below is free: AWS Skill Builder (free AI courses from Amazon) skillbuilder.aws AWS AI learning paths aws.amazon.com/ai/learn Kiro, Amazon's agentic coding IDE kiro.dev Strands Agents, Amazon's open-source agent SDK strandsagents.com The Strands harness SDK on GitHub github.com/strands-agents/ha… Strands agent examples you can copy github.com/strands-agents/sa… Bedrock AgentCore (run agents in production) aws.amazon.com/bedrock/agent… AgentCore samples github.com/awslabs/agentcore… Official AWS MCP servers github.com/awslabs/mcp Amazon Bedrock agent samples github.com/aws-samples/amazo… PartyRock (build AI apps with no code) partyrock.aws All of this costs $0 The next "dog psychiatrist" job will be invegents from these links, and in a few yearsit'll pay $200 an hour with a waitlist Bookmark it and start with the first link tonight Then read how to start building with agents in the harness engineering guide below ↓

0:40
godsystem avatar
⟳ godsystem retweeted
godsystem@godsystemsmd·Nov 10

This is f*cking insane. Y Combinator CEO Garry Tan: "When someone asks how I prompt my AI, the answer is: I don't. The skills are the prompts" In 2 minutes, he shows the setup that turns Claude Code into a full AI engineering team, built on one idea: thin harness, fat skills Prompts → Skills → A Team Of Specialists → A Startup Built By Agents One of his skills runs YC office hours: tens of thousands of hours from 16 YC partners, distilled into a file Claude follows Do a workflow once, turn it into a skill, and it runs that way forever Save this before it disappears from your feed Everything below is free: GStack (his 23 Claude Code skills: CEO, designer, eng manager, QA) github.com/garrytan/gstack GBrain (his open-source second brain) github.com/garrytan/gbrain How to make Claude Code your AI engineering team (YC) youtube.com/watch?v=wkv2ifxP… Don't Build Agents, Build Skills Instead (Anthropic) youtube.com/watch?v=CEvIs9y1… How Perplexity designs and maintains agent skills perplexity.ai/hub/blog/desig… Anthropic's official skills repo github.com/anthropics/skills Agent Skills docs platform.claude.com/docs/en/… How Anthropic equips agents with skills anthropic.com/research/equip… Agent Skills open standard agentskills.io Free Agent Skills course with Anthropic deeplearning.ai/courses/agen… All of this costs $0 Prompting is dead, skills are next, and 99% are still typing prompts from scratch every morning Bookmark it and start with the first link tonight Then read how thin harness, fat skills works in the harness engineering guide below ↓

2:16

Your agent passed every test The issue is still open A better prompt won't fix it, because the model isn't the broken part I broke down the harness, loop and graph around it: what each one owns, how they nest, and which one to fix first ↓