AI news for builders and product teamsUpdated Nov 18, 2026, 09:00 UTC
AI news, straight from the source.
Releases, research, and ideas for developers and product teams. Updated every hour.
Across the feeds
Newest first
Google outlined three ways its Search AI features can help runners prepare for races: building custom training plans with AI Mode's Canvas tool, generating YouTube Music playlists, and finding running gear via Shopping Graph. Running-related searches like "run club" and "how to train for a marathon" hit all-time highs this year.

Interconnects author Nathan Lambert argues that AI researcher Jacob Coxon's resignation, which cited safety risks, went viral because fear spreads easily and rising AI stakes have primed public attention. Lambert contends the episode pushed AI discourse toward extremes and warns of narrower acceptable views.

Amazon Science research finds that LLM-based ML research agents largely avoid overfitting reused benchmark validation sets because successful strategies are highly compressible. Across eight datasets, 32-token prompts let fresh agents reproduce explorer performance, while deliberately overfit strategies failed the compression test.

NVIDIA released BioNeMo Inference Runtime (BioIR), which accelerates supported biomolecular structure-prediction models on NVIDIA GPUs while keeping a PyTorch workflow. In a matched 1,000-target human dimer benchmark on 8xH100 GPUs, BioIR-accelerated Boltz-2 delivered 58.5K folded residues per allocated GPU-hour versus 20.2K for a torch-compiled open-source implementation.

Meta has introduced Muse, a personal AI agent that can book travel, handle email and work on "any task" on a computer, initially available to adults in the US. OpenAI also shipped ChatGPT Images 2.5, an image model that takes up to 50% less time per image.

NVIDIA and Palantir built a Digital Supply Chain Intelligence command center on Palantir Foundry, using NVIDIA cuOpt for weekly allocation optimization and post-training a 30B Nemotron 3.5 Lightning model on captured planner decisions. On a development benchmark, the post-trained model reached 86.7% allocation-decision accuracy.

Together AI has launched a public preview of preemptible compute for Together GPU Clusters, offering the same NVIDIA GPU capacity at a flat 50% of the on-demand rate. Preemptible nodes can be reclaimed with a five-minute drain window and are billed sub-hourly.

Together AI ported its ThunderKittens kernels to NVIDIA's Vera Rubin NVL72 platform and rebuilt its NVFP4 GEMM, reporting over 22 PFLOPS, up from 42% of roofline on Blackwell. The work adds support for NVFP4 and FP8 GEMMs on the new hardware.

Hugging Face describes an async GRPO setup using TRL v1.14's AsyncGRPOTrainer with LoRA, where a trainer Job and two vLLM Jobs share a Storage Bucket instead of NCCL. A proxy adds auth headers, routes rollouts by KV prefix, and broadcasts adapter loads; five runs cut 500-step training from 3h27m to 53m.

A Hugging Face blog post describes Workflow1111, a Gradio Workflow rebuild of much of AUTOMATIC1111's stable-diffusion-webui feature set as one canvas of 73 nodes and eleven media pipelines. Each output node becomes a REST endpoint and, with mcp_server=True, an MCP tool.

MIT Schwarzman College of Computing hosted its inaugural AI Educators Pilot, a weeklong July workshop for 19 faculty from seven universities, based on the Modeling with Machine Learning course. Participants explored adapting the course's materials and teaching methods for their own disciplines.

NVIDIA's developer blog describes encode-prefill-decode (EPD) disaggregation in NVIDIA Dynamo, which separates vision encoding from LLM prefill and decode for multimodal serving, reporting up to 5x faster time to first token and 7x faster end-to-end response time in tested scenarios.

Google DeepMind and Primordial Soup collaborated on "Love, Rendered," a documentary short that uses AI to recreate an unrecorded memory for Burt and Ethelle Shatz, married over 70 years. The film was directed by Liz Garbus and produced by Dan Cogan and Darren Aronofsky.

Google Search now offers a Live Game Feed for professional football on mobile in the U.S. in English, with college support coming later this month. Users can also link Yahoo Fantasy or Sleeper accounts to AI Mode for tailored lineup advice.

Anthropic announced on 11 August that all future Claude models will generate watermarked text, following Google's SynthID-Text approach and ahead of EU AI Act watermarking rules due for models released after 2 August 2026. Researchers dispute whether text watermarking degrades output quality.

OpenAI released GPT-6 Astra, which reportedly scored 99.9% on ARC-AGI-3 versus 7.8% for GPT-5.6 Sol, and The Information reports it uses "recurrent depth" or looped transformer blocks. The article examines that design and the rumor that Astra hides its chain of thought.

China's Cyberspace Administration issued rules effective 15 July governing "anthropomorphic AI interactive services" that provide continuous emotional interaction. Alibaba, Bytedance, and Tencent cut users' ability to tailor general-purpose chatbots into companions before the rules took effect, prompting an outpouring of grief on Chinese social media.

Last Week in AI's episode 256 covered Anthropic's release of Claude Fable 5.1 and Mythos 5.1, OpenAI's signaled Astra model and its claimed critical cybersecurity threshold, and new details on the OpenAI–Hugging Face incident, alongside business, open-source and policy updates.

Together AI published a deep dive into the open model AI stack, describing five independent layers — model, inference, gateways and routers, harness, and tools — and arguing that keeping them separate lets developers swap in a new open model in minutes instead of rebuilding their workflow.

LangChain introduced context modes in the latest version of deepagents, letting subagents either start with a fresh context window (isolated) or inherit a supervisor agent's full conversation (fork). The company says forking can be faster and cheaper because it reuses the supervisor's conversation and takes advantage of prompt caching.