Dan McAteer
@daniel_mac8 • 31,756 subscribers
I build AI agents for a living and write about it. Writing at https://t.co/ROXTvUfk75 · Writing @latentspacepod · Agentic Ops @ AnswerRocket
Shorts
![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!](https://image.24vids.com/tw-2096634766883696666/amplify_video_thumb/2096634632758538240/img/5HZAy33oD2Kbwg2a.jpg)
Videos
Codex users: here's an amazing way to take advantage of the efficiency and capability of GPT-5.6 Luna Max. It's called 'sol-advisor'. 1. GPT-5.6 Sol High as orchestrator 2. GPT-5.6 Luna Max as implementer for routine tasks 3. GPT-5.6 Terra Max as implementer for complex tasks 4. Fresh GPT-5.6 Sol instance as reviewer A free and open-source plugin for Codex. Installation instructions below 👇 I dare you to install it, and tell me if you hit your weekly usage limits using 'sol-advisor'. I doubt you can.
Dan McAteer694,937 Aufrufe • vor 2 Monaten
This is the best way to use Claude Fable in Claude Code without immediately hitting your limits. 1. Model set to Fable 5 2. Reasoning on Max 3. Instruct Claude to run a dynamic workflow where: 3a. Fable is the orchestrator 3b. Opus does the reasoning heavy phases Fable is so overpowered that you don't need its intelligence for every step. Let it orchestrate Opus or even Sonnet.
Dan McAteer912,533 Aufrufe • vor 4 Monaten
Fable 5.1 orchestrates GPT-5.6 Luna agents. Via the 'fable-advisor' plugin for Claude Code. 1. Fable 5.1 orchestrates 2. GPT-5.6 Luna implements 3. Fable 5.1 reviews Think of it like in-context distillation: give Luna the intelligence of Fable 5.1 at a fraction the cost.
Dan McAteer155,042 Aufrufe • vor 1 Monat
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 McAteer111,166 Aufrufe • vor 1 Monat
Grok 4.5 is a "Wow!" moment. 🤯 > Opus 4.8 & GPT-5.5 level on benchmarks > Costs less than Sonnet 5 Is that real? Hard to believe. Added Grok 4.5 as an implementer agent in fable-advisor. It replaces Sonnet 5. Use Fable as the orchestrator and a combo of: 1. Grok 4.5 2. Opus 4.8 3. GPT-5.5 As implementer agents. Uses the Grok Build CLI so if you log in with an X premium account you don't have to pay API fees.
Dan McAteer191,832 Aufrufe • vor 3 Monaten
You should know that you should install the 'you-should-know' plugin in Claude Code *TODAY*. It surfaces important information from complex tasks that you might miss. Another example of the evolving "Attention Interface". Turn it on: /plugin enable cc-plugin-you-should-know@builtin
Dan McAteer28,102 Aufrufe • vor 14 Tagen
This is one of the coolest open-source AI agent projects I've seen in a while: 'Understand Anything' It's a plugin for Claude Code, Codex, OpenCode etc. that analyzes your codebase and turns it into a knowledge base that you can interact with. It explains the codebase to you, rather than showing you the structure. It seems like it's designed for code but I opened my Obsidian vault of podcast highlights in Claude Code, then ran /understand. The result is a knowledge graph that I can search of highlights from 888 podcast episodes and 144K lines of markdown text.
Dan McAteer175,169 Aufrufe • vor 4 Monaten
A GPT-5.4 rumor was that the model can persist state. Jeff Dean mentioned this on his Latent.Space appearance, so clearly something the AI Labs are thinking about. I'd bet there's a good chance they've discovered how to effectively integrate State-Space Models with Transformers at scale. SSMs are designed to carry a hidden state forward with every computational step, and scale linearly rather than quadratically like Transformers. Also backs up the rumor that GPT-5.4 will have a 2 million token context window. Persistent state would essentially mean that AI models go from being Guy Pearce's character in "Memento" to Dustin Hoffman's character in "Rain Man" overnight. It would give AI models true long-term memory. It would be a massive breakthrough for real.
Dan McAteer212,788 Aufrufe • vor 7 Monaten
I pointed the new Claude Managed Agents dynamic workflows at 18k lines of code. > 21 proven bugs. > 5 minutes. > $8.36, on my Claude Max API credits. A lead agent writes a plan, runs it across many agents in phases, then merges the results. Try it in Claude Code: /claude-api managed-agents-onboard bug-hunter It burns tokens fast. It's worth it.
Dan McAteer10,875 Aufrufe • vor 9 Tagen
The best way to take advantage of ChatGPT Voice. I am completely and utterly voice-pilled. Voice interaction with AI is going to change the way humans interact with computers. Here's what you can do: 1. Create a remote connection to an always-on desktop PC through the ChatGPT Desktop app (I have a Linux box in my basement) 2. Customize the ChatGPT iOS widget and add a remote voice connection 3. Tap the widget. It will open a voice chat in ChatGPT Work/Codex, on your desktop, with access to your files and all your apps. It's like an always on Jarvis that has its own computer. Amazing stuff. Try it out!
Dan McAteer46,466 Aufrufe • vor 1 Monat
karpathy on Dwarkesh in Oct. '25: > AI agents don't work > They are not intelligent enough > They are not multi-modal enough > They can't do computer use karpathy on No Priors in March '26: > AI agents fail but it's a skill issue for the humans > You didn't give them the right instructions AI agents absolutely hit an inflection point in December of 2025 and there's no looking back.
Dan McAteer136,855 Aufrufe • vor 7 Monaten
this is getting weird 🤯 used Operator to access Google AI Studio used Gemini 2.0 to create instructions to build a website instructed Operator to use the instructions to create a portfolio website on replit THIS IS NOT A TOOL TO BOOK FLIGHTS!!! 🚨 THIS IS AN INTELLIGENT AGENT THAT CAN DO COGNITIVE WORK FOR HUMANS IN THE INFORMATION LAYER OF HUMANITY THAT WE CALL THE WEB.
Dan McAteer338,926 Aufrufe • vor 1 Jahr
If you do nothing else today, do yourself this one huge favor: turn on the Codex Chrome extension. It's surprisingly fast, but the speed is not the point. The point is, if you're a knowledge worker, you can have Codex perform routine browser tasks for you. It frees you to do higher value work. I've used it every day at work. There's no reason a human should have to spend their time manually manipulating a browser for routine tasks. Your job is to be a human being and create. Let Codex be the computer.
Dan McAteer66,904 Aufrufe • vor 4 Monaten
Claude Fable 5 is best used as an orchestrator that does the heavy reasoning and delegates token intensive but simpler tasks to cheaper models. 'fable-advisor' is your free and open-source plugin that does exactly that. New in fable-advisor: > A directive to Fable to minimize its own token usage (works surprisingly well) > GPT-5.5 as an implementer sub-agent via the Codex CLI (uses your ChatGPT sub so you don't pay extra) How to use it: 1. claude plugin marketplace add DannyMac180/fable-advisor && claude plugin install fable-advisor 2. /model fable
Dan McAteer56,732 Aufrufe • vor 3 Monaten
holy moly @windsurf from Codeium is the best AI coding agent experience i've had 💨 🏄♂️ I scaffolded an entire Next.js app, including frontend and backend - db, auth, frontend framework etc. by simply talking to "Cascade" their AI coding agent they've figured something out something with the idea of - "Flows": "AI flows ensure that the AI always understands the context of your work, adapting instantly as you transition between tasks, so it can assist without needing to be brought up to speed or interrupt your focus." it's the first time working with an AI that I feel it actually understands what I'm trying to do and is acting with true agency it also stops to consult me when needed if anything is unclear it's a beautiful experience, so well done the best part is that it's free to try for the next few weeks for any developer
Dan McAteer190,467 Aufrufe • vor 1 Jahr