AI — Page 16

Matt Dailey's method front-loads constraints and review before generated UI reaches Ref's production code.

Kushal Mittal and nine Intel contributors packaged setup, sizing, benchmarking, profiling and CUDA migration into one Apache-2.0 repository.

The free September 10th-13th event pairs match screenings with a serve simulator, AI-generated ping-pong recaps and a photobooth beside IBM's Manhattan office.

ReViSQL-K2.6 reached 88.55% with greedy decoding and 91.37% with 16-sample self-consistency on the corrected Arcwise-Plat-SQL benchmark.

Judge Rita F. Lin ordered the administration to lift its ban on Anthropic technology after finding the exclusion inadequately justified, Bloomberg reported.

The Y Combinator Summer 2026 startup is betting that production traces can push routine work onto cheaper models customers own.

The company raised its full-year outlook as custom silicon and optical interconnects outpace traditional segments of the business.

Google DeepMind says its pilot keeps external evaluation prompts and model weights private, though its first-of-its-kind claim remains independently unverified.

First disclosed on March 3rd, the production research system uses up to 48 reference frames to preserve a person's identity across new scenes.

The Tsinghua-born AI lab held back the coding model's checkpoints for two weeks after its cyber capabilities grew faster than expected.

The public preview gives Azure customers access to Grok 4.6 with a 500,000-token context window, extending xAI's multi-cloud distribution push.

In a 50-task run of Newsroom Reliability v0.2, Claude Opus 5 (Fast) led the field with a 0.73 score at $0.0460 per task. GLM 5.3 Flash, Gemini 3.7 Flash, and Qwen3.8 Flash followed at 0.69, each at substantially lower cost.

A test interface shows users speaking directly with agents that already browse websites, operate software and run jobs in the cloud.

In a 12-task Editorial Craft evaluation, Z.ai’s GLM 5.3 Flash ranked first with a 0.94 score and an estimated cost of $0.0002 per task. Step-3.7-flash followed at 0.92, while Qwen3.8 Flash, Claude Opus 5 (Fast), Gemini 3.7 Flash, and DeepSeek-V4-Flash-0731 rounded out the leaderboard.

The third-party provider offers metered access without identifying the checkpoint, revision, modification method or endpoint evaluation behind its service.

The AI chip giant crushed Street expectations on top line and provided a massive $108 billion outlook for the next quarter.

Nick Walton's open beta separates world state from AI narration, with free public multiplayer and paid private servers across web and mobile.

Imran Siddique's cryptographic receipt format has backing from AMD, Intel and Microsoft, though its governance and core verification work remain unfinished.

Manil Lakabi's YC-backed tool plans, designs and packages React Native apps for testing and App Store review.

Portable Computer keeps models and files on a DGX Spark, while cloud escalation still requires approval and credits.

The public-preview model handles 85+ languages, custom jargon and up to three speakers across live and recorded audio.

Stanford, Oxford and METR analyzed three separate samples of about 250,000 conversations through Anthropic's privacy filter.

Google says its latest speech-to-text model handles structured strings, filler removal, formatting and custom vocabulary, while pricing and the access path remain undisclosed.

Blackroot's new guide translates a January architecture that trades some MoE experts for hashed N-gram embeddings.