Benchmarking >80 LLMs shows: The best model is not necessarily the best for your programming language 😱
— Markus Zimmermann (@zimmskal) September 2, 2024
- Best overall: Anthropic’s Sonnet 3.5
- Best for Go: Meta’s Llama 3.1 405B
- Best for Java: OpenAI’s GPT-4 Turbo
- Best for Ruby: OpenAI’s GPT-4o
Good models for one… pic.twitter.com/EYUphEI5rH
Great to see more concrete results published on how different models are “the best” at writing different programming languages.
Iterating on Cogno, improving the “remaining guesses” and sharing functionality.
Shipped a very-beta version of a recipe ingredient extraction app, which can then generate a grocery list from many recipes. Calling it snapcart.
Having used fasthtml and sqlite of a couple projects now, I was interested in trying to run an app as a Cloudflare Worker to make use of D1, which is a serverless, sqlite compatible database.
It looks like the library needs to be added to the worker runtime first, according to this issue.