AI News
Built a local-first context engine for AI coding agents, symbol graph + semantic search, no cloud
Sharing a project I’ve been building: Argyph, an MCP server that gives AI coding agents (Claude, or anything…
AI News
Today’s Irony. We as small creators cannot use AI but big companies can ban us using same AI
Sharing my experiences from recent horrific fights among AI slop fighters and big companies. Lately I have been…
AI News
Rewriting model inference with CUDA kernels: the bottleneck was not just GEMM [P]
I’ve been working on a CUDA-first inference runtime for small-batch / realtime ML workloads. The core idea is…
AI News
llama.cpp MTP support landed – Qwen3.6 27B at 2.44× on a Strix Halo, 2.17× on a RTX 3090 rig
PR #22673 (commit 4f13cb7) landed MTP speculative decoding in mainline llama.cpp on May 16. I tested it on…
AI News
Warelay -> OpenClaw
In preparation for a lightning talk I’m giving at PyCon US this afternoon I decided to figure out…
AI News
Musk v. Altman proved that AI is led by the wrong people
The tech trial of the year, Musk v. Altman, was ultimately a fight for control. Elon Musk argued…
AI News
The FBI Wants to Buy Nationwide Access to License Plate Readers
The FBI wants to buy access to automated license plate readers (ALPRs) nationwide, which would likely allow the…
AI News
Need Advice, Should I retake two courses over the summer or prioritize research [D]
I am about to graduate from a masters program in Data Science. I had a 3.75 gpa before…
AI News
Sub-JEPA: a simple fix to LeCun group’s LeWorldModel that consistently improves performance [P]
World models learn compact latent representations for planning without pixel reconstruction. LeWorldModel (LeWM), from LeCun’s group at NYU,…
AI News
Configuration Qwen3.6-35b-a3b (12Gb VRAM)
Has anyone here tested different KV cache quantizations and compared their performance? I’m currently using the model in…