AI — Page 37

The limited-release offering combines voice and chat agents, evaluations, guardrails and hands-on deployment by OpenAI engineers.

The Teams and Enterprise feature sends each request to a different model and gives admins control over cost, quality and access.

Sam Fok's Stanford spinout is pushing sparse silicon into headphones, smart glasses and appliances, while its 5x growth claim leaves the metric undefined.

The former Accel investor assembled a Palantir-heavy team and reached a reported $140 million valuation within seven months.

On July 22nd 2026 the X founder revealed a sandbox where AI creates persistent societies, previewing a multi‑citizen test before a public release.

Brothers Aditya and Rishab Jain are using YC to sell employers a benefits stack spanning plan administration, care navigation and payments.

Alex Yeh's GPU cloud tied the commitments to capacity coming online after live ARR grew 2.4x in the first half.

The workspace adds member controls as Sara Hooker and Sudip Roy push Adaption's model-building tools toward organizational use.

The July 21st-28th contest asks developers to bring their own IDE or coding agent while Base44 supplies the infrastructure.

The open‑source model Nanbeige4.2‑3B adds a Looped Transformer layer that the team says improves agent performance while keeping the parameter count at three billion.

The public beta generates physics-ready 3D scenes for Isaac Sim, MuJoCo and Gazebo, with an API built for automated testing workflows.

The video‑AI startup’s new agent lets users ask Anthropic’s Claude to pull clips, stats and themes from thousands of videos at once.

The new tool creates people, ledgers, board minutes and contracts for a full‑year of operations, using a simulated DTC brand as a pilot.

The Mintlify engineering manager built a shared editor where humans and coding agents can write, comment and suggest changes in the same document.

Jeff Seibert's ledger software will enter CLA's $2 billion-plus services operation through its client accounting practice.

Frozen v2 could reach servers in 2028, setting a new benchmark for specialized inference-chip startups if Google keeps Gemini's architecture stable.

SYNAPS-I fine-tunes SAM 3 and DINOv3 on beamline data, then runs them across 300 Nvidia A100 GPUs inside government infrastructure.

The warning turns Anthropic's distillation allegations into a trade weapon, though Bessent named no target or sanctions authority.

Its contrastive finetuning method changed lying rates from 9% to 87% in one example, though OpenAI omitted sample sizes and model details.

The open-weight model activates 8B parameters per token, while Poolside's benchmark claims still await independent leaderboard verification.

The lawsuit, filed by the University of Tennessee Research Foundation, alleges that Anthropic's AI systems infringe two patents related to neuroscience-inspired computing.

Co-founder MacKenzie Price's film posits AI frees teachers from standardized instruction to focus on human-centric coaching and relationship building.

Gemini 3.6 Flash cuts output pricing, Flash-Lite raises rates for more throughput, and the delayed Gemini 3.5 Pro remains in partner testing.

Block's self-hostable workspace uses signed Nostr events, while its current relay architecture remains centralized within each deployment.