Cortisol spit curve tests are pretty cheap, so I was disappointed to see the title claim not substantiated with real data. Consider biometric testing next time you distribute!

Probably in the minority here but I, too, went to the article hoping for objective quantification of the change in cortisol levels corresponding to the input, i.e. trying to distribute Mac software from scratch.

or say yes? decel mentality like this is why europe is falling behind. some poor startup will try to backfill these contracts to be the new palantir of europe only to be cut at the knees by regulation and more outcry think piece boycotts like this. rinse and repeat until the us and china become the only relevant acceleration hubs on earth during the singularity

the animation of the model name text when opening the detail view is so smooth and delightful

Congrats on launch! As the agent cli’s and sdk’s were built for local use, there’s a ton of this infra work to run these agents in production. Genuinely excited for this space to mature.

I have been building an OSS self-hostable agent infra suite at https://ash-cloud.ai

Happy to trade notes sometime!

Yeah with sandbox pre-warming and disk co-location its fast enough to avoid UX cold start penalty.

On write amplification — we persist at the message level, not per SSE chunk. The sandbox's workspace filesystem (claude code's native jsonl files) is the source of truth for resume, and the DB is for queryability, tracing, etc - so fire and forget works fine here.

I’m building a self-hostable, open source agent sandbox orchestrator here: https://github.com/ash-ai-org/ash-ai

directionally correct but important to note the water wasted by sustaining the insufferable human is much higher than producing the tokens

i've always wondered (for this, portkey, etc) - why not have a parallel option that fires an extra request instead of mitm the llm call?

You can fire them in parallel for simple cases. The issue is when you have multi-agent setups. If context isn't persisted before a sub-agent reads it, you get stale state. Single source of truth matters when agents are reading and writing to the same context.

For single-agent flows, parallel works fine.

I think it’s the reverse - people were too lazy to read the docs so nobody was motivated to write them.

With an agent I know if I write once to CLAUDE.md and it will be read by 1000’s of agents in a week.

I like this insight. We kind of always knew that we wanted good docs, but they're demotivating to maintain if people aren't reading them. LLMs by their nature won't be onboarded to the codebase with meetings and conversations, so if we want them to have a proper onboarding then we're forced to be less lazy with our docs, and we get the validation of knowing they're being used.