Love this post by @antirez on developing Redis Array support. Its a great showcase of thoughtful AI usage and how AI can empower even the best developers while still producing high quality work.

I was curious if and how much Astra could improve antirez/ds4 inference on M4 Max & DGX Spark - here are the first test results after ~90 min >The strongest result so far is about 6-10% faster decoding on the M4 Max >On Spark, the strongest completed comparison is 3.30% faster decoding

Qwen 3.8 Flash Next PR 991 for DwarfStar ready for your review super @antirez Rebased to current main and squashed to a single commit. MoE is not using standard path and is custom for Qwen. Feel free to change anything and move models under your HF 🚀

🚨 OPEN SOURCE AI IS LITERALLY UNSTOPPABLE 🚨 The legendary founder of Redis (Antirez) just dropped ds4 - a custom native inference engine built specifically for DeepSeek v4 Flash This is earth shattering! Here is why: DeepSeek v4 Flash is a quasi-frontier model with a massive 1M context window You can now run it LOCALLY on a 128GB Mac using specialized 2-bit quantization The architecture is reimagined—he moved the KV cache from RAM directly to the SSD disk! 🤯 We already know DeepSeek v4 Flash is insanely good for agentic loops - Now you don't even need the cloud to run it Closed-source labs are burning tens of billions on massive GPU clusters while single brilliant developers are running frontier-level AI on laptops! They told us open-source would be worthless against trillion-dollar monopolies Instead, pure hacker culture + incredible open-weight models are completely rewriting the rules Open Source will ALWAYS win 💕

Local AI tip: if you've got a big-memory Mac, you can run frontier open models at home now. antirez, the creator of Redis, built ds4, a lean C engine that runs a few top open models on your own machine. the setup: * models → DeepSeek V4 Flash, GLM 5.x and Qwen 3.8 Flash Next * hardware → high-memory Macs, NVIDIA and AMD GPUs * speed → about 39 tokens per second on a 128GB M5 Max * tools → a CLI, an OpenAI and Anthropic compatible API, and a coding agent why it's different: → built for a few models only, instead of everything at once → 2-bit quantization so big models fit in memory → 23,000+ GitHub stars, MIT licensed One dev. One engine. Frontier models on your desk. the repo:

Small experiment: I'm using my Pi coding sessions as calibration data for a new DeepSeek V4 Flash Q2 quant, via antirez's ds4 imatrix machinery. Absolutely no expectations, but I'm curious to see what happens. Not sure how to eval though.

the 80s/90s F/OSS movement saw a cluster of tech intellectuals spring up all dedicated to software freedom* and hacker culture - RMS, @esrtweet , Richard Gabriel, Larry Wall, jwz, etc. Feels like a resurgence now on people dedicated to a similar spirit around AI - thinking of @realGeorgeHotz , @antirez , @AndrewYNg , @beffjezos , @dhh and others. * as in speech, not as in beer.

Thank you to the staff at the facility and to @AntiRecidivism, @ScottyBudnick and @samwlewis1 for having me. Happy Thanksgiving to all of the men and women that are away from their families this year.

A small clinical trial for people with Type 1 diabetes combined a transplant with an experimental antirejection drug, allowing some patients to stop taking insulin. More research is needed, but experts are cautiously optimistic about the results.