: Anthropic says most code is not human “reviewed” in the typical sense for them anymore. I do like …
: “**Engineers are no less valuable, but their work is changing. **All of the engineers who worked on …
: That face when AI agents take your coding velocity so high that git can’t keep up. Followed shortly …
: A nice less technical take on using local AI. As someone who has gone deep down the rabbit hole of …
: Temperature 0 does not mean reproducible or deterministic for an LLM. Even if setting fixed random …
: Is sandboxing enough to stop AI agents from sending, receiving, or executing malicious instructions? …
: Sometimes gen AI is just scaffolding until you figure out the programmatic or deterministic was to …
: Revisiting my original introduction to AI safety: Rationality by Eliezer Yudkowsky 📚 (not a general …
: Revisiting one of my favorite books on AI. Lots of good intuition and memorable imagery: You Look …
: Writing something yourself forces you to take the time to make it clear and organize your thoughts. …
: Noticing a pattern with Gemini where when I ask it to help me find some primary source document, it …
: Some empirical work on what kinds of tools in agentic harnesses help LLMs do the best work. They …
: “AIs do not have rights, feelings, or consciousness. And we must not train them to act as though …
: Sometimes AI coding agent sessions inject their own custom instructions during context compaction. …
: I think this is a great core idea: a generative AI that can respond only with the choice options the …
: A welcome perspective on thinking about management vs individual contributor. I need to go rethink …
: Not sure how long it’s been in production, but I noticed with OS 27 that Apple has two built-in …
: MIT committee’s report on AI in education has refreshingly concrete recommendations like …
: A good reminder, especially as the rate of technological advancement only increases, migrations are …
: Causing me to consider system-prompting all my models to not speak in first person or express desire …
: Big if true, “Run Long-Horizon Inference Faster on CPUs Than on GPUs” [eLLM: Run Long-Horizon …
: Cool idea. Basically knowing how variables impact each other (not just the outcome) with Shap …
: “Coders at these labs have sunk untold resources into creating the ideal coding interface. What …
: I don’t think I agree, but this is an interesting take. Basically math is about the process of the …
: Interesting reflections on how a digital and AI world changes our sense of humanity This Jungian …
: An idea I’ve yet to see well executed, but I’m interested in hugginface’s “ML Intern,” which …
: “If I use ChatGPT to help me partially solve a Millennium Prize problem, what are the chances that …
: https://github.com/maziyarpanahi/openmed Very cool use of Apple MLX in the example app. Cool small …
: Two fascinating recent articles. One on integration of generative AI into clinical pathways (from …
: Yellow-Bellied Whistling Marmot Yellow-bellied marmot seen on the way to Pikes Peak
: Impossible Creatures Currently reading: Impossible Creatures by Katherine Rundell 📚 Interesting world, compelling plot. …