Tyler

evals / rl envs
@micro1_ai

[Tyler @tyhouch]

My weekend project is live! I need some feedback on the UI/UX & feature requests (also if anything breaks) DMs are open!

inspired by @GregKamradt 's tweet

[Greg Kamradt @GregKamradt]

has anyone made an arXiv tracker yet?

I want to monitor for papers that meet a search criteria

fuzzy llm search, not keyword

[Tyler @tyhouch]

@growing_daniel have u tried ‘brew install claude-code’

[Tyler @tyhouch]

yo @SlackHQ please add support for viewing .html files in chat. everyones tossing these around and the download->open in chrome friction is slowing me down bigtime

[Tyler @tyhouch]

@ivanburazin and those making infra DX easy will benefit greatly from this 😆

[Tyler @tyhouch]

@GregKamradt fancy pizza? wine?

cmon you gotta come do this in nyc

[Tyler @tyhouch]

tbh I’m very excited and optimistic for Siri AI

typical Apple fashion of waiting till it’s good enough

[Tyler @tyhouch]

@DeepankarBhade cap

[Tyler @tyhouch]

@signulll are we sure this is an 'untapped' market 🧐

[Tyler @tyhouch]

@maharshii have you tried ‘npx skills add fix-issue’

[Tyler @tyhouch]

@tbpn @dylan522p ultimate picks and shovels play is just talking about the picks and shovels

[Tyler @tyhouch]

@mattpocockuk didn’t @karpathy start building a company for this after leaving OpenAI? Where’d that go?

[Tyler @tyhouch]

@ben_burtenshaw @ClementDelangue @huggingface @reflection_ai @UnslothAI @modal @PrimeIntellect @NVIDIAAI @mercor_ai @fleet_ai can someone explain simply why I would need OpenEnv and Harbor? It seems like harbor solves this harness problem itself. Why’s another interface layer needed

[Tyler @tyhouch]

@gabriel1 damn congrats beast. looking forward to followng this

[Tyler @tyhouch]

@SeanZCai traditional business either become data companies or data companies will just consume their business. Especially true as the cost of high quality software creation gets closer to zero.

[Tyler @tyhouch]

This is fs gonna show up on model cards moving forward

[Rishi Desai @rishi_desai2]

Can coding agents stay coherent over a 1 billion token budget?

Can they build Slack from scratch?
Rewrite a JAX codebase in PyTorch?
Build a C compiler in Rust?

Enter SWE-Marathon: a benchmark for autonomous long-horizon software work.

[Tyler @tyhouch]

feeling like a verifiable task today

[Tyler @tyhouch]

@himanshustwts building infra sometimes makes me feel like just a means to an end 😆

agree it’s wildly unappreciated (im mega biased)

[Tyler @tyhouch]

@tszzl wispr flow usage is best metric

[Tyler @tyhouch]

@rishi_desai2 do you plan to release the full tasks on HarborHub so I can try to reproduce 😀 super interested in the implementation of the CUA verifier

[Tyler @tyhouch]

@rishi_desai2 This agentic verifier is so simple im pissed I didn’t think of it 😃

[Tyler @tyhouch]

@rishi_desai2 This is awesome. Congrats!