A minimum set of math for reading AI, LLM, and image-generation articles — the aim isn't to derive anything, just to recognize weighted sums, S-curves, probabilities, and the 'nudge toward the answer' step of training.
Hands-on Qwen3.6-35B-A3B (23GB 4bit GGUF) on M1 Max 64GB via Ollama 0.20.6. Generation speed stays at 27 tok/s — same as Qwen3.5-35B-A3B — but the same prompt produces 13× more thinking tokens. Multi-turn behavior, persona handling, and a three-tier NSFW probe included.
Hands-on run of trellis-mac (the CUDA-free port of TRELLIS.2) on M1 Max 64GB. Setup via uv with PyTorch 2.11.0 MPS, applied mps_compat.py patches, and recorded actual generation time vs the M4 Pro 24GB 3.5-minute reference, plus where the bottlenecks land on Apple Silicon.
Alibaba's Qwen3.6-Max-Preview and Moonshot AI's Kimi K2.6 were released within a 24-hour window on April 20–21, 2026. A side-by-side look at specs, benchmarks, distribution, and agent-side features for the two flagships.
Cloudflare's Artifacts is a Git-compatible version control storage designed for AI agents to operate tens of millions of repositories. The Git server runs as a Zig-based WebAssembly binary on Durable Objects, accessible through Workers Bindings, REST API, and standard Git clients.
A port that replaces TRELLIS.2's CUDA-only libraries (flash_attn, nvdiffrast, sparse 3D convolution) with pure-PyTorch equivalents and runs Microsoft's 4B image-to-3D model on an M4 Pro in about 3.5 minutes without any NVIDIA GPU.
Vercel's official incident disclosure published on April 19, 2026. A walk-through of how a compromise of Context.ai's Google Workspace OAuth app led to Vercel employee account takeover and access to environment variables in some customer projects, plus the checks users should run right now.
Anthropic Labs released Claude Design on April 17, 2026 — a new design product that reads your codebase and existing design files to generate prototypes, slides, and landing pages. Here is a feature comparison with Figma and Canva, and a scenario-by-scenario guide to which tool to open when.
A three-link chain of mmap → MTLBuffer(bytesNoCopy) → Wasmtime MemoryCreator that makes a Wasm linear memory share the same physical bytes as a Metal GPU buffer. Llama 3.2 1B runs at 9ms/token on M1.
A vulnerability in iTerm2 3.6.9 and earlier where simply displaying a malicious file with cat triggers local code execution. Caused by conductor impersonation in SSH Integration, fixed in 3.7.0.
Measured: Opus 4.7 burns 1.2-1.45x tokens vs 4.6 in community benchmarks (Bill Chambers' Tokenomics Leaderboard, Claude Code Camp), beating Anthropic's official 1.0-1.35x range. CLAUDE.md alone hits 1.4-1.45x.
Two simultaneous announcements from Cloudflare Agents Week 2026: Agent Memory manages agent recall via Durable Objects, Vectorize, and Workers AI, while isitagentready.com scores how well sites are prepared for agents.
