You can now give an AI a GPU and let it improve its own model overnight. This repo turns LLM research into an automated search loop. @karpathy just open-sourced a system that lets AI iterate on its own training code. Karpathy’s AutoResearch turns LLM improvement into an
CODE
-
NVIDIA Scales Agent Inference with Engineering Stack Innovation
By
–
NVIDIA's AI Engineers: Agent Inference at Scale and "Speed of Light" https://
latent.space/p/nvidia-brev-
dynamo
… @nvidia is embracing Engineers at all levels of the stack: from the sublime DX of @brevdev
, to the open source datacenter framework of Dynamo. @KranenKyle and @NaderLikeLadder join -
AI Coding Tools Becoming Required for Software Developer Roles
By
–
Out of 539 poll respondents here who had recently interviewed for software developer roles, 32% reported that experience with AI coding tools didn't come up at all, 25% said it came up as optional and 43% said that it came up as required
-
Structured-RAG: Solving Aggregative Queries in RAG Systems
By
–
Standard RAG falls apart on aggregative queries. Things like "average ARR for companies with >1k employees" require reasoning across hundreds of docs, and vector retrieval just can't handle that reliably. We built Structured-RAG to fix this:
– induces a schema from your -
Amazon Imposes Senior Review for AI-Assisted Code After Incidents
By
–
Proceed with caution Lukasz Olejnik (@lukOlejnik) Amazon is holding a mandatory meeting about AI breaking its systems. The official framing is "part of normal business." The briefing note describes a trend of incidents with "high blast radius" caused by "Gen-AI assisted changes" for which "best practices and safeguards are not yet fully established." Translation to human language: we gave AI to engineers and things keep breaking? The response for now? Junior and mid-level engineers can no longer push AI-assisted code without a senior signing off. AWS spent 13 hours recovering after its own AI coding tool, asked to make some changes, decided instead to delete and recreate the environment (the software equivalent of fixing a leaky tap by knocking down the wall). Amazon called that an "extremely limited event" (the affected tool served customers in mainland China). — https://nitter.net/lukOlejnik/status/2031257644724342957#m [Translated from EN to English]
→ View original post on X — @madhumita29, 2026-03-10 12:53 UTC
-
Run Claude Code with local LLMs for free
By
–
Run Claude Code using local LLMs for FREE. No API costs. No data leaving your machine. Here's how it works: Claude Code lets you swap its backend via a single env variable. Point `ANTHROPIC_BASE_URL` to a local llama.cpp server, and it'll route all requests to whatever model
-
Mermaid Diagram Tool Adds Zoom and Pan Controls
By
–
I've not found anything I like more than mermaid yet – I added zoom/pan controls to it recently to help deal with a really large diagram https://
github.com/datasette/data
sette-table-diagram?tab=readme-ov-file#usage
… -
We are in the early Uber/Lyft stage of AI coding agents
By
–
We are now in the early Uber/Lyft stage of AI coding agents. Do you remember how you used to be able to commute to work with Uber for less than $10 a ride? Those were fun times.
-
Vibe-coding versus production-ready AI development
By
–
vibe-coding it vs making it production-ready
-
RLHF Quality Issues Human Feedback Volume Incentives
By
–
The problem with RLHF is that the humans providing the feedback are rewarded for volume, not quality.