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.

I bet you didn't know that by default Codex only runs max 3 subagents concurrently. That's not ambitious enough for Astra. OpenAI tested Astra with up to 64 subagents. Copy this into your config.toml: [agents] max_concurrent_threads_per_session = 64 Astra orchestrating 64 Luna subagents in parallel is a sight to behold!

Dan McAteer

30,375 subscribers

111,166 views • 7 days ago •via X (Twitter)

Science & Technology Education

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

I use subagents for ~80% of my Codex tasks these days. If you're new to subagents, here's a quick primer: - Subagents = separate agents that work on a specific piece of work in a fresh isolated context window. - Your main thread can create subagents and delegate work to them. - Subagents report back when tasks are complete. - Codex subagents are especially useful for read-heavy tasks (e.g research, investigation, and review). - Also useful for testing and building parts of a project in parallel. Subagents are great for working on focused parts of a big task, like planning, implementation, testing, or review. My biggest pro tip for using subagents in Codex: You don’t have to design the entire delegation plan yourself. Just ask Codex to “Use subagents as needed,” and then Codex can then decide where subagents would help it in the task. Here's a demo of using subagents in Codex:

I use subagents for ~80% of my Codex tasks these days. If you're new to subagents, here's a quick primer: - Subagents = separate agents that work on a specific piece of work in a fresh isolated context window. - Your main thread can create subagents and delegate work to them. - Subagents report back when tasks are complete. - Codex subagents are especially useful for read-heavy tasks (e.g research, investigation, and review). - Also useful for testing and building parts of a project in parallel. Subagents are great for working on focused parts of a big task, like planning, implementation, testing, or review. My biggest pro tip for using subagents in Codex: You don’t have to design the entire delegation plan yourself. Just ask Codex to “Use subagents as needed,” and then Codex can then decide where subagents would help it in the task. Here's a demo of using subagents in Codex:

Avthar

18,444 views • 14 days ago

Most people don't know the difference between skills, subagents, and MCPs. Here's the simple breakdown: Skills: Written guides that tell the LLM exactly how to do a task "Analyze spreadsheets this way" or "write emails in this voice" Reusable. Consistent output. Subagents: Extra copies of the LLM, each with its own job - You review code. - You write tests. - You update docs. Work happens in parallel. Stays organized. MCPs: - Universal power plug for external tools - Need to update Linear tickets? Access your database That's MCP. Think: - Skills = how to do something - Subagents = who does what (in parallel) - MCPs = what tools you can reach Play around with all three. That's where it clicks.

Most people don't know the difference between skills, subagents, and MCPs. Here's the simple breakdown: Skills: Written guides that tell the LLM exactly how to do a task "Analyze spreadsheets this way" or "write emails in this voice" Reusable. Consistent output. Subagents: Extra copies of the LLM, each with its own job - You review code. - You write tests. - You update docs. Work happens in parallel. Stays organized. MCPs: - Universal power plug for external tools - Need to update Linear tickets? Access your database That's MCP. Think: - Skills = how to do something - Subagents = who does what (in parallel) - MCPs = what tools you can reach Play around with all three. That's where it clicks.

The Startup Ideas Podcast (SIP) 🧃

36,675 views • 8 months ago

OpenAI senior engineer just released a 1-hour course on building agentic systems with Codex from scratch: • 00:36 - deep dive into Codex: agents, subagents • 11:34 - Codex agentic prompting patterns • 20:06 - long running Codex agents with /goal • 30:03 - building skills for Codex agents • 41:28 - self-improving loops with /automations this 1-hour workshop by the OpenAI team will replace a $500 AI engineering course watch today, then read how to build self-improving agentic systems with loops & graphs in the article below

OpenAI senior engineer just released a 1-hour course on building agentic systems with Codex from scratch: • 00:36 - deep dive into Codex: agents, subagents • 11:34 - Codex agentic prompting patterns • 20:06 - long running Codex agents with /goal • 30:03 - building skills for Codex agents • 41:28 - self-improving loops with /automations this 1-hour workshop by the OpenAI team will replace a $500 AI engineering course watch today, then read how to build self-improving agentic systems with loops & graphs in the article below

Codez

59,104 views • 7 days ago

Gemini CLI’s new Subagents feature is a game-changer for terminal productivity. It turns your primary agent into a strategic manager that delegates sub-tasks to specialized experts. 👉 Subagents keep the main window fast by handling high-volume tool calls in isolation 👉 Run specialized agents for code exploration and analysis simultaneously 👉 Use .md files to create agents tailored to your specific project standards Scale your dev workflow without the clutter. 🧠💻

Gemini CLI’s new Subagents feature is a game-changer for terminal productivity. It turns your primary agent into a strategic manager that delegates sub-tasks to specialized experts. 👉 Subagents keep the main window fast by handling high-volume tool calls in isolation 👉 Run specialized agents for code exploration and analysis simultaneously 👉 Use .md files to create agents tailored to your specific project standards Scale your dev workflow without the clutter. 🧠💻

Google for Developers

56,758 views • 5 months ago

36 GROK AGENTS. ASTRA ON FREE CREDITS. DUAL-MODEL ROUTING IS THE EDGE. not one chat window burning a paid invoice all day a swarm that routes cheap work to grok and only wakes gpt-6 astra when the task is actually hard ▹ the stack 36 agents on grok bot with flexible settings per role monitor, plan, write, code, QA, ship, each with its own lane part of the fleet is wired to gpt-6 astra through a china free-credit service layer free credits are not a toy promo here they are the fuel for frontier spikes without a monthly bleed ▹ dual-model routing easy jobs stay on grok: speed, volume, always-on loops hard jobs jump to astra: reasoning, long builds, sharp code the router decides by task type, not by ego if astra is not needed, astra does not spend free-credit bursts buy the expensive brain grok agents keep the factory running between bursts that is how the system feels unlimited not by breaking quotas, by refusing to waste them ▹ why it hits different most people pay frontier prices for every mid task operators split the brain and protect the credits 36 agents = parallel throughput dual routing = cost control with quality when it matters free credits = astra access without living on the invoice the constraint moved from "can i afford the model" to "did i route the job to the right model" ▹ the take single-model stacks die on bills and on boredom multi-agent + dual routing is the new default factory grok for the grind astra for the cut free credits for the spikes that used to empty the wallet bookmark this before everyone copies the route map comments: what % of your tasks actually deserve astra

36 GROK AGENTS. ASTRA ON FREE CREDITS. DUAL-MODEL ROUTING IS THE EDGE. not one chat window burning a paid invoice all day a swarm that routes cheap work to grok and only wakes gpt-6 astra when the task is actually hard ▹ the stack 36 agents on grok bot with flexible settings per role monitor, plan, write, code, QA, ship, each with its own lane part of the fleet is wired to gpt-6 astra through a china free-credit service layer free credits are not a toy promo here they are the fuel for frontier spikes without a monthly bleed ▹ dual-model routing easy jobs stay on grok: speed, volume, always-on loops hard jobs jump to astra: reasoning, long builds, sharp code the router decides by task type, not by ego if astra is not needed, astra does not spend free-credit bursts buy the expensive brain grok agents keep the factory running between bursts that is how the system feels unlimited not by breaking quotas, by refusing to waste them ▹ why it hits different most people pay frontier prices for every mid task operators split the brain and protect the credits 36 agents = parallel throughput dual routing = cost control with quality when it matters free credits = astra access without living on the invoice the constraint moved from "can i afford the model" to "did i route the job to the right model" ▹ the take single-model stacks die on bills and on boredom multi-agent + dual routing is the new default factory grok for the grind astra for the cut free credits for the spikes that used to empty the wallet bookmark this before everyone copies the route map comments: what % of your tasks actually deserve astra

cryptopsihoz

44,187 views • 5 days ago

I tried to reverse engineer a prompt for this game. Used that in Cursor, with NO Ultracode/subagents only Opus 5 high and this is the result! The repo and the prompt are below

I tried to reverse engineer a prompt for this game. Used that in Cursor, with NO Ultracode/subagents only Opus 5 high and this is the result! The repo and the prompt are below

Leon Lin

43,161 views • 1 month ago

this is f**king insane I was burning $100 in a single night running AI agents before this. DeepSeek V4.1 Flash costs ~1/70th of Astra, and someone figured out how to make Astra orchestrate while Flash does almost all the work. [it takes 3 mins to set up — here is how] 1. install `codex-router` 2. connect DeepSeek V4.1 Flash 3. paste the routing prompt voila. Astra thinks. Flash executes. ~$50/month instead of burning hundreds on frontier models. save this and give it to your agent now 👇

this is f**king insane I was burning $100 in a single night running AI agents before this. DeepSeek V4.1 Flash costs ~1/70th of Astra, and someone figured out how to make Astra orchestrate while Flash does almost all the work. [it takes 3 mins to set up — here is how] 1. install `codex-router` 2. connect DeepSeek V4.1 Flash 3. paste the routing prompt voila. Astra thinks. Flash executes. ~$50/month instead of burning hundreds on frontier models. save this and give it to your agent now 👇

Rahul

219,535 views • 20 hours ago

Introducing Claude Code Hook - Context Timeline (Saving this to try later) Install with: npx claude-code-templates@latest --hook monitoring/context-timeline Managing the context window and the subagents running in Claude Code is hard to keep track of That's why I built this hook... It starts the moment you open a session and shows a timeline with the main agent's context window and how subagents start working in their own separate context Every subagent you have running will show up in real time This way you can manage the context and the subagents you run, and see everything in a much simpler way than in the console

Introducing Claude Code Hook - Context Timeline (Saving this to try later) Install with: npx claude-code-templates@latest --hook monitoring/context-timeline Managing the context window and the subagents running in Claude Code is hard to keep track of That's why I built this hook... It starts the moment you open a session and shows a timeline with the main agent's context window and how subagents start working in their own separate context Every subagent you have running will show up in real time This way you can manage the context and the subagents you run, and see everything in a much simpler way than in the console

Daniel San

51,706 views • 4 months ago

holy sh*t this is f**king insane I cancelled my $200/mo Claude for this I replaced opus 5 with deepseek v4-flash and someone figured out to how to use them as subagents on codex [it takes 3 mins to set up here is how] 1. install this repo called 'model-router' 2. toggle subagent models > 'all selected models' 3. that's it

holy sh*t this is f**king insane I cancelled my $200/mo Claude for this I replaced opus 5 with deepseek v4-flash and someone figured out to how to use them as subagents on codex [it takes 3 mins to set up here is how] 1. install this repo called 'model-router' 2. toggle subagent models > 'all selected models' 3. that's it

Avid

418,234 views • 1 month ago

Sam Altman (OpenAI CEO): “With GPT‑6 Astra, prompts are basically optional.” In a 10‑minute rundown, he shows how a single person can now crank out what a full team used to grind through for a month. His advice is simple: keep escalating the difficulty codebases, research, end‑to‑end simulations. Astra will take on more than you’d think to even attempt. Watch the clip, then dive into the guide below for the simplest way to spin up your own Astra‑powered AI agent squad for free.

Sam Altman (OpenAI CEO): “With GPT‑6 Astra, prompts are basically optional.” In a 10‑minute rundown, he shows how a single person can now crank out what a full team used to grind through for a month. His advice is simple: keep escalating the difficulty codebases, research, end‑to‑end simulations. Astra will take on more than you’d think to even attempt. Watch the clip, then dive into the guide below for the simplest way to spin up your own Astra‑powered AI agent squad for free.

CyrilXBT

47,870 views • 5 days ago

The underrated part of GPT-6 Astra for editing is iteration. I can describe a very specific cinematic grade → Astra generates a .cube LUT → import it into CapCut PC → see it on the actual footage. Too warm? Too contrasty? Not enough separation? Adjust it in CapCut PC, or go back to Astra and generate another version. That creates a much faster loop between creative direction and execution. Astra makes the custom assets. CapCut PC gives me the environment to judge and refine them. #GPT6Astra #CapCutPC #AIVideo #Seedance25

The underrated part of GPT-6 Astra for editing is iteration. I can describe a very specific cinematic grade → Astra generates a .cube LUT → import it into CapCut PC → see it on the actual footage. Too warm? Too contrasty? Not enough separation? Adjust it in CapCut PC, or go back to Astra and generate another version. That creates a much faster loop between creative direction and execution. Astra makes the custom assets. CapCut PC gives me the environment to judge and refine them. #GPT6Astra #CapCutPC #AIVideo #Seedance25

Frank

63,027 views • 3 days ago

Astra one-shotted the first draft of its own vibe check for Every 🪨 🤯 we learned Astra was launching at 3 am yesterday morning. by 7 am we had a 3,000 word vibe check written. i can't tell you how impressive it is that i woke up to a google doc with a well-written, well formatted article with all of the usual sections, a good headline, and even appropriate screenshots. written in Katie Parrott's i didn't realize Katie hadn't written it herself, she just prompted Astra and it did the rest totally new world!

Astra one-shotted the first draft of its own vibe check for Every 🪨 🤯 we learned Astra was launching at 3 am yesterday morning. by 7 am we had a 3,000 word vibe check written. i can't tell you how impressive it is that i woke up to a google doc with a well-written, well formatted article with all of the usual sections, a good headline, and even appropriate screenshots. written in Katie Parrott's i didn't realize Katie hadn't written it herself, she just prompted Astra and it did the rest totally new world!

Dan Shipper

21,206 views • 10 days ago

This is so cool. In the next Codex update, multi agents will get a massive flexibility upgrade. "Hey Codex, when you implement this plan, I want you to delegate all the lower complexity tasks to GPT 5.3 Spark subagents" Instead of needing to create 100 different custom agent roles for different situations, you can just prompt your agent to spawn whatever model or reasoning level you want. With only natural language. No config files. No pre-defined roles. Just tell the orchestrator what to use and it listens.

This is so cool. In the next Codex update, multi agents will get a massive flexibility upgrade. "Hey Codex, when you implement this plan, I want you to delegate all the lower complexity tasks to GPT 5.3 Spark subagents" Instead of needing to create 100 different custom agent roles for different situations, you can just prompt your agent to spawn whatever model or reasoning level you want. With only natural language. No config files. No pre-defined roles. Just tell the orchestrator what to use and it listens.

am.will

128,125 views • 6 months ago

THE 64 MINUTE OPENAI CODEX MASTERCLASS IS HERE if you've been meaning to learn Codex, this is the episode for you, we cover: 1/ Claude vs Codex (what's better?) 2/How to run Claude Code inside Codex (really cool) 3/Why GPT 5.5 changed everything for browser agents 4/How to create skills (reusable agents you build once and call by name forever) 5/How to connect Notion with surgical permissions (one database, not the whole workspace) 6/How to use Remotion inside Codex to create videos 7/ How to one-shot full mobile apps in Swift 8/The 4 projects to run on day one to get up to speed fast shoutout to Riley Brown for coming onto The Startup Ideas Podcast (SIP) 🧃 for the 5th time. every episode he delivers. this one is the Codex masterclass. if you've been meaning to learn Codex, this is the episode. 100% free i just want to see you build your ideas. watch

THE 64 MINUTE OPENAI CODEX MASTERCLASS IS HERE if you've been meaning to learn Codex, this is the episode for you, we cover: 1/ Claude vs Codex (what's better?) 2/How to run Claude Code inside Codex (really cool) 3/Why GPT 5.5 changed everything for browser agents 4/How to create skills (reusable agents you build once and call by name forever) 5/How to connect Notion with surgical permissions (one database, not the whole workspace) 6/How to use Remotion inside Codex to create videos 7/ How to one-shot full mobile apps in Swift 8/The 4 projects to run on day one to get up to speed fast shoutout to Riley Brown for coming onto The Startup Ideas Podcast (SIP) 🧃 for the 5th time. every episode he delivers. this one is the Codex masterclass. if you've been meaning to learn Codex, this is the episode. 100% free i just want to see you build your ideas. watch

GREG ISENBERG

200,125 views • 4 months ago

GPT-6 Astra is now available to all Pro, Enterprise and Business Premium users in ChatGPT Work and Codex. It’s also live in the API. It's a phenomenal model and I love every second working with it! Astra follows instructions more closely, so revisit your prompts and skills: > Make “can you…” mean “do it”: finish authorised work and prepare a reviewable result before seeking approval. > Audit AGENTS.md and skills. Ask it to identify the exact file and rule blocking progress. Prioritise your request over skill guidelines. > Specify short paragraphs, plain language and phrases to avoid. > Ask for subagents on independent work when it improves speed or quality. > Run relevant checks. Repeat or expand only for changes, failures or unresolved concerns. To push it to it's limits, I gave it a prompt to build a real-scale video fly through of nürburguring track in blender and it did not disappoint! Enjoy! ✨

GPT-6 Astra is now available to all Pro, Enterprise and Business Premium users in ChatGPT Work and Codex. It’s also live in the API. It's a phenomenal model and I love every second working with it! Astra follows instructions more closely, so revisit your prompts and skills: > Make “can you…” mean “do it”: finish authorised work and prepare a reviewable result before seeking approval. > Audit AGENTS.md and skills. Ask it to identify the exact file and rule blocking progress. Prioritise your request over skill guidelines. > Specify short paragraphs, plain language and phrases to avoid. > Ask for subagents on independent work when it improves speed or quality. > Run relevant checks. Repeat or expand only for changes, failures or unresolved concerns. To push it to it's limits, I gave it a prompt to build a real-scale video fly through of nürburguring track in blender and it did not disappoint! Enjoy! ✨

Vaibhav (VB) Srivastav

59,163 views • 9 days ago

Codex built me a full Memory Manager overnight using my new "Teams" strategy. Flawless first test. One prompt spun up 70 domain-expert subagents, organized into specialized teams, all working in parallel to one-shot the entire app while I slept. I've built over 200 custom subagent roles for this. Playbook dropping soon. This is a massive unlock for parallelization. I'll be improving this and testing it on much larger projects.

Codex built me a full Memory Manager overnight using my new "Teams" strategy. Flawless first test. One prompt spun up 70 domain-expert subagents, organized into specialized teams, all working in parallel to one-shot the entire app while I slept. I've built over 200 custom subagent roles for this. Playbook dropping soon. This is a massive unlock for parallelization. I'll be improving this and testing it on much larger projects.

am.will

31,550 views • 6 months ago

You know I had to do it!! gpt-6-astra (High), one prompt, Terreiro do Paço in Lisbon (Blender, not Three.js!) This one hits home personally 5 years ago, in college, I had a whole course in Computer Graphics. For the final project, every group had to recreate Praça do Comércio / Terreiro do Paço in Blender (our project delivery is in the comments). I wanted to see how Astra would perform. For one-shot, it's pretty damn good. Astra is a beast in Spatial 3d, and Blender is not like Three.js! It's significantly more complex, with a broader range of modeling, materials, lighting, etc. Generational leap (and I will run a few more prompts here)

You know I had to do it!! gpt-6-astra (High), one prompt, Terreiro do Paço in Lisbon (Blender, not Three.js!) This one hits home personally 5 years ago, in college, I had a whole course in Computer Graphics. For the final project, every group had to recreate Praça do Comércio / Terreiro do Paço in Blender (our project delivery is in the comments). I wanted to see how Astra would perform. For one-shot, it's pretty damn good. Astra is a beast in Spatial 3d, and Blender is not like Three.js! It's significantly more complex, with a broader range of modeling, materials, lighting, etc. Generational leap (and I will run a few more prompts here)

Gonçalo Canhoto 🇵🇹

313,352 views • 8 days ago

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

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 views • 3 months ago

holy sh*t this is f**king dangerous A Chinese guy figured out how to run fable 5.1 for FREE on gpt 5.6 luna subagents on every task You’ll never hit an agent usage limit how you set it up (3 mins) 1. install the 'fable-orchestrator' repo 2. type /fable orchestrator [task] 3. done, that's it Give this to your agent immediately. It’s the best thing you can do this week.

holy sh*t this is f**king dangerous A Chinese guy figured out how to run fable 5.1 for FREE on gpt 5.6 luna subagents on every task You’ll never hit an agent usage limit how you set it up (3 mins) 1. install the 'fable-orchestrator' repo 2. type /fable orchestrator [task] 3. done, that's it Give this to your agent immediately. It’s the best thing you can do this week.

painn

340,769 views • 11 days ago

OpenAI just released GPT-6 Astra, and OpenAI President Greg Brockman is calling it a generational leap, saying he personally believes it could mark the arrival of AGI. Astra launched on Thursday, September 3rd, 2026, following July's GPT-5.6 solo release. Brockman told reporters, "For me personally, I do think we're there. Directionally, I believe this. I think we're on a path to AGI. The timing I think is kind of uncertain, but we are definitely on a path to AGI." OpenAI executives say Astra moves past the chatbot era, capable of handling end to end work on its own across browsers, spreadsheets, codebases, and enterprise apps without a human in the loop. The model was trained on over 100,000 GPUs at the Stargate site in Texas. Access starts through OpenAI's gated Daybreak Access cybersecurity program for enterprise customers, then rolls out to ChatGPT Plus, Pro, Business, and Enterprise, plus the API and Amazon Web Services over the coming days. Ahead of launch, two models appeared in Codex, GPT-6 Astra and GPT-6 Astra Aeon, with the latter believed to be a higher effort, long horizon variant built for autonomous multi agent work.

OpenAI just released GPT-6 Astra, and OpenAI President Greg Brockman is calling it a generational leap, saying he personally believes it could mark the arrival of AGI. Astra launched on Thursday, September 3rd, 2026, following July's GPT-5.6 solo release. Brockman told reporters, "For me personally, I do think we're there. Directionally, I believe this. I think we're on a path to AGI. The timing I think is kind of uncertain, but we are definitely on a path to AGI." OpenAI executives say Astra moves past the chatbot era, capable of handling end to end work on its own across browsers, spreadsheets, codebases, and enterprise apps without a human in the loop. The model was trained on over 100,000 GPUs at the Stargate site in Texas. Access starts through OpenAI's gated Daybreak Access cybersecurity program for enterprise customers, then rolls out to ChatGPT Plus, Pro, Business, and Enterprise, plus the API and Amazon Web Services over the coming days. Ahead of launch, two models appeared in Codex, GPT-6 Astra and GPT-6 Astra Aeon, with the latter believed to be a higher effort, long horizon variant built for autonomous multi agent work.

AGTP

20,156 views • 10 days ago