Just

micromanaging 4 AI agents looks good on a timelapse

hello there the jacobian conjecture is false thanx to my close friend akhil for asking about it and my other close friend fable for working during the world cup final ((1+xy)^3 z + y^2 (1+xy) (4+3xy), y + 3 x (1+xy)^2 z + 3 x y^2 (4+3xy), 2 x - 3 x^2 y - x^3 z): \C^3\to \C^3, has jacobian determinant -2, and sends (0, 0, -1/4), (1, -3/2, 13/2), and (-1, 3/2, 13/2) to (-1/4, 0, 0)

The past years of agentic coding often felt like babysitting. But the latest wave of models feel like a breath of fresh air. I'm spending less time weeding out slop, feedback loops are almost instant, and I end the day actually satisfied with how much got done.

Being a founder is explaining to your family for the fifth year in a row that no, it's not profitable yet, and yes, you're still doing it

has anyone else realized there's nothin to watch on youtube anymore

Current bill situation: Codex $100 Claude code $200 AWS $17,622,128,552 Openrouter $45 Not able to understand how to reduce my bill. Any help would be appreciated

Under the hood, it generates a lightweight DSL, then a deterministic renderer turns that into SVG/PNG. Much cheaper than having the model emit the full output directly.

Been enjoying an agent skill I made called Fat Marker. Quick lo-fi UI sketches. Nice way to keep layout iterations cheap before rendering the final HTML. Inspired by Shape Up's fat marker sketches: #fat-marker-sketches" target="_blank" rel="nofollow noopener">basecamp.com/shapeup/1.3-ch…

woah 🤯

sit down and enjoy the singularity, I guess?

Actual solutions to environmental problems (like here) vs Virtue signaling solutions that do nothing

Embarrassment of riches. But looks like we might hit 9M soon. Should we reset the ChatGPT Work and Codex usage again or give it some space?

Lots of codex resets lately, yet here I am stuck on admin tasks. I am dying to build something.

The reason you do not verbally attack a dead person, no matter how vile and morally offensive he was, has to do with you & not with him. He is no threat to mankind any longer; have some skin in the game & go focus on living assholes.

Thank you to the 7M active users who are now using Codex and ChatGPT Work. We have added a banked reset to everyone's account to celebrate the milestone. You can apply the reset in the desktop app or on web and it will replenish the weekly usage for you. Have fun out there.

Someone DM'd me saying they're using /grill-me in technical interviews Watch the candidate work. See how they answer questions and push back. See whether the AI drives them, or they drive the AI. Honestly genius

Updates for Codex and ChatGPT Work users. No nerfing, only good stuff! - We have landed inference optimizations and are passing down savings to all the subscriptions for GPT-5.6 Sol. That should result in around 10% more usage on its own. - We noticed that by changing the context size limit in the product to 372k for GPT-5.6 Sol, up from 272k for GPT-5.5, it resulted in more usage being charged than intended. We have reverted to 272k and will work to roll back out to 372k in the days to come. You should notice that usage drains significantly less after this change. - To understand where the extra usage was coming from, we ran some experiments where reasoning efforts were changed (referred to as juice values under the hood) and have reverted this. - There is slightly more usage of multi-agent than intended in high and xhigh reasoning effort, we are fixing this going forward. Also fixing a small other thing we noticed with auto-review where we can be more efficient. And we continue to have the 5h limit temporarily not apply. Enjoy the rest of the weekend!

I made a huge mistake thinking that manual technical jobs such as electician, appliance technician etc. wd be spared by AI. Had pbs w/: HVAC, security system, outdoor lighting, pool programming etc. & was forced to solved them w/AI because technicians are slow at showing up.

How to tame GPT-5.6 Sol usage. A piece of advice: 1. Add this to your config.toml: [features.multi_agent_v2] enabled = true hide_spawn_agent_metadata = false tool_namespace = "agents" 2. OpenAI increased the default context window for the 5.6 series from 258k to 353k. Given how good compaction in Codex is, you don't actually need anything above the initial 258k. Ask your agent to set the context window back to 250k in your config. 3. GPT-5.6 is a different beast and requires some pipeline adjustments. Sol spends way more time gathering context before writing code, and it spawns agents on its own. It can choose a model for each subagent, but it doesn't always route tasks efficiently. Specify models and tasks in your AGENTS.md so the main thread doesn't burn your quota by using Sol Max for routine tasks and boilerplate. 4. You don't actually need reasoning higher than high 99% of the time. Use Max for planning and medium or high for execution.

🇪🇺 Chat Control has passed 😔 They can and will now legally scan any person's messages, emails and photos you send without a warrant The way they passed this law when the majority of the European Parliament was against it will shock you: They waited until most EP members were on holiday so only a few were present and then created an "urgent" vote for it to pass it through There's nothing democratic about any of it and big powerful forces are behind this that can manipulate the EU for whatever they want Democracy in Europe died a bit today 😔

Tldraw why T.T 😅