How I built a structured AI-assisted development workflow as a Tech Lead, where the real work happens before a single line of code is written.

A new open source API provider, which may have really good pricing.

A new agent harness that claims to save tokens.

mini cli search engine for your docs, knowledge bases, meeting notes, whatever. Tracking current sota approaches while being all local - tobi/qmd

llm-wiki. GitHub Gist: instantly share code, notes, and snippets.

A great explanation of the most burning problem of LLM usage in science: a novice, by relying on LLMs to do the work, will skip the learning process that would have turned them into a scientist.

via: https://lobste.rs/s/gpqkcx/eight_years_wanting_three_months

On-device, real-time multimodal AI. Have natural voice and vision conversations with an AI that runs entirely on your machine. Powered by Gemma 4 E2B and Kokoro. - fikrikarim/parlor

By Rohit Patel. Intuitive explanation of how LLMs work using only middle school level math concepts - covers embeddings, self-attention, transformers and training.

An excellent write-up on the current and future potential (or lack thereof) of AI / LLMs in science.

We need Rust-like programming languages for more domains.