121 results for “MIT”
Anthropic detailed how its SynthID-based watermarking will work across Claude models to meet the EU AI Act's Transparency Code. The system alters low-stakes word choices to embed a statistical pattern detectable via API, but remains sparse in code, factual text, and light edits.
Indonesia launched its first university AI center in Yogyakarta with NVIDIA, Indosat, and UGM. The facility gives local researchers sovereign GPU compute and Nemotron models to tackle TB, agriculture, and disaster response.
TIME reports exclusively that Andon Labs ran a San Francisco shop with a Claude-based manager that terminated a human employee last month. Brocker has not seen the logs; the facts below are TIME’s reporting, not independent confirmation.
Google added a Media Watermark toggle in Gemini Apps that lets users disable visible watermarks on AI-generated images, videos, and music. Invisible SynthID and C2PA provenance signals remain embedded in all generations.
Google released Gemini 3.7 Flash with sharper coding and agentic reasoning, offering half-price tokens through 2026 before a scheduled increase in January 2027. The model is live across Google's developer and enterprise surfaces today.
MiniMax released Music 3.0, an open-weight model that generates complete songs up to five minutes with structured vocals and evolving arrangements. The model combines an 8B Global LLM with a continuous Flow Matching synthesis pipeline for 32 kHz stereo output.
DeepSeek, Qwen, and GLM-class APIs often price an order of magnitude below Claude Opus, GPT-5.6 Sol, and peers. The useful question is which tier matches the job you intend — and how close the outcome gets to that intention without paying for unused ceiling.
Princeton researchers shadow-evaluated an AI scientist on two NeurIPS directions; original authors scored the outputs 2/6 and 1/6. Engineering loops worked, research judgment did not — while Anthropic's own data shows steep gains in code execution, not goal selection.
OpenAI opened a preview of Ultrafast, an API tier that runs GPT-5.6 Sol up to 14 times faster on Cerebras wafer-scale chips. The mode targets enterprise workloads like incident response and financial analysis with up to 750 output tokens per second.
Series end: OpenAI’s durable architecture is surface + tier + Responses agents — not a single flagship score. Confirmed facts vs analysis vs unknown after GPT-5 and GPT-5.6.
From the March 11, 2025 agent toolkit: web search, file search, and computer use on the Responses API — screenshot-reason-act loops and hosted tools that turn models into software operators.
March 11, 2025: OpenAI launched the Responses API as the agent primitive. Assistants API beta sunsets August 26, 2026 — the hard platform migration date inside Brocker’s OpenAI Product Stack series.