Latest
Aug
26
Apple's M5 Ultra Mac Studio bets on bandwidth for local AI
Apple kept M5 Ultra's 512GB ceiling but raised bandwidth to 1.2TB/s. The 256GB model starts at $9,499; prompt-prefill gains remain untested.
3 min read
Aug
25
Every major browser engine is converging on JPEG XL. Keep your fallbacks anyway
Safari already decodes JPEG XL and Firefox targets default support in version 157, but Firefox's SDR-only HDR handling, no Chrome milestone, and near-parity with lossless WebP keep fallbacks necessary.
3 min read
Aug
24
MCP’s new roadmap names the right gaps, but the fixes are still proposals
Stateless scaling and several authorization foundations shipped in July. Workload identity, progressive discovery and server-pushed events have no delivery dates, and the groups responsible for identity and discovery are still forming.
4 min read
Aug
21
The Rust arrayref attack turned a routine Cargo build into code execution
A poisoned dependency could execute code during compilation, so Rust teams should inspect every lockfile and Cargo cache touched on August 20.
4 min read
Aug
19
Apple's EU per-install fee is dead. A 5 percent sales commission replaces it.
Apple's October terms replace the per-install fee with a 5 percent revenue commission, retain monthly reporting, and lock payment choices for 12 months.
4 min read
Aug
18
DuckDB v2.0 preview targets a stable release for its Quack server mode
The preview also outlines a stable extension API and new storage format, while Quack remains beta and migration details await the fall release.
3 min read
Aug
17
Stripe’s reported OpenRouter deal is a reason to audit gateway costs and lock-in
Neither company has confirmed the reported $7B deal, and the price could still change. But the two already share billing infrastructure, which gives OpenRouter users reason to audit their routing costs, data policies, and exit path.
3 min read
Aug
13
Grok 4.6 arrives with long-running coding agents in focus
xAI positions Grok 4.6 as a frontier competitor with $2/$6 token pricing. Repository-level tests will show whether that translates into cheaper completed tasks.
2 min read
Aug
12
Zuckerberg says AI should be for everyone. Muse Glimmer is Meta’s first test
Muse Glimmer is a real 30B Apache 2.0 release, but its performance on common hardware, independent benchmark results, and Meta’s promised follow-ups remain open.
3 min read
Aug
11
Claude’s output marks are here, but developers still lack the details to test them
Claude output watermarking now applies worldwide on supported Anthropic models, but code-specific testing and older-model rollout details remain unavailable.
3 min read