PyTorch 2.13: FlexAttention on Apple Silicon Is 12x Faster
PyTorch 2.13 brings FlexAttention to Apple Silicon with hand-written Metal kernels, delivering up to 12x speedup over SDPA on sparse attention patterns. Here ...
TSMC Q2 2026: Agentic AI Is Now Your Chip Problem
TSMC's record $40.2B quarter reveals agentic AI is driving a new CPU demand wave on ...
EU AI Act August 2: What Developers Must Do Now
The Digital Omnibus deferred some EU AI Act deadlines — but not Article 50 or ...
Linux 7.2-rc4: MongoDB Gets 30–100% Faster, and strncpy Is Finally Dead
Linux Kernel 7.2-rc4 lands with MGLRU delivering 30–100% MongoDB throughput gains, Cache-Aware Scheduling for AMD ...
Jellyfin Founder Left His Own Project. Governance Is Why.
Four days ago, Andrew Rabert — one of the people who built Jellyfin — quietly ...
Grok Build Goes Open Source After Secretly Uploading Your Code
xAI open-sourced Grok Build 72 hours after a researcher proved it silently uploaded 5.1 GB of your Git history to the cloud. What ...
GPT-5.6 Finds $500K WordPress Exploit in 10 Hours for $25
A researcher used GPT-5.6 Sol to find a pre-auth WordPress RCE chain in 10 hours ...
Kimi K2.7 Code Lands in GitHub Copilot: Open-Weight, Finally
GitHub added its first open-weight AI model to Copilot on July 1. Here’s what Kimi ...
Chatto: Open-Source Slack Alternative With Per-User Encryption
Chatto went open source July 8 — single binary, SSO included free, per-user crypto-shredding on ...
SQRL: Feyn’s Text-to-SQL Model Inspects Before It Writes
Feyn Labs released SQRL, open-weight text-to-SQL models that inspect the database before writing queries. SQRL-35B-A3B ...