AI news for builders and product teamsUpdated Nov 18, 2026, 03:01 UTC
AI news, straight from the source.
Releases, research, and ideas for developers and product teams. Updated every hour.
Across the feeds
Newest first
Amazon now operates more than 1 million warehouse robots while employing over 1.5 million people, and it plans to expand both its robotic fleet and the factories that build it. Research cited in the article links robotic fulfillment centers to fewer severe injuries but more repetitive-strain injuries as robots set the pace of work.

Import AI 476 covers four releases: the HSS visual-reasoning benchmark where GPT 6 Astra scores 53.6 versus a 93.1 human baseline, Reflection's 501B-parameter Beam model, Mistral Large 4, and a P-Zero Research benchmark showing model research taste doubling every 3.0 months.

Google DeepMind's AlphaProof Nexus solved nine of 353 open Erdős problems and other open math questions, with the results peer reviewed and published in Science. The work lands amid a boycott push against OpenAI for posting AI proofs on GitHub without review.

Apple researchers evaluated LLMs against human annotators for machine translation quality scoring under the MQM and ESA schemes, finding agreement varies widely by scheme, language pair and domain and remains unreliable in most settings.

Apple researchers published a paper describing Minions, a distributed virtual GPU cloud that aggregates idle GPUs in Apple Silicon macOS devices to run large-scale machine learning inference. The authors say the approach offers a way to use existing hardware where centralized GPU resources are constrained.

Researchers at Johns Hopkins University propose SELFCOMPACT, an inference-time scaffold that lets a language model decide for itself when to summarize its accumulated context, matching or beating fixed-interval compaction at 30-70% lower cost per question.

Apple will announce a new smart home hub called the HomeView on Tuesday, according to Bloomberg's Mark Gurman, a Siri-powered device with a display and an A18 chip. Gurman says updated Apple TV and HomePod mini models will be unveiled the same day.

TechCrunch Disrupt 2026 opens in less than 48 hours at San Francisco's Moscone West, running October 13-15 with more than 10,000 attendees, 250+ speakers and 200+ sessions. Passes cost up to $100 less before door rates take effect October 13 at 8 a.m. PT, with a second pass at 50% off.

Some prominent Silicon Valley figures are adopting President Trump's preferred term "super intelligence" in place of "artificial intelligence," with Elon Musk promising to rename SpaceXAI to SpaceXSI.

Amazon will stop using nondisclosure agreements when negotiating with local governments over data centers, following a similar move by Microsoft earlier this year. The change, buried in a blog post arguing data centers benefit communities, comes as moratoriums and local opposition mount.

Vals AI research reports that multi-agent AI teams only marginally outperform single agents while consuming up to 5.1x more tokens; just one of four tests with GPT-6 Sol and Claude Opus 5.5 showed a measurable gain.

At the HumanX conference, PolyAI CTO Shawn Wen said voice AI has not yet had its "ChatGPT moment," despite full-duplex models that can speak and listen at once. He and Otter CMO Alex Gay pointed to fast reasoning, accurate speech recognition, and transparency as the next hurdles.

Researchers from Meta, MIT, and the University of Washington propose Context Language Models (CLMs), which let an LLM treat its own context as an editable file instead of relying on fixed summarization or retrieval pipelines, reporting accuracy gains alongside lower compute on several benchmarks.

Two separate studies from Epoch AI and Anthropic found that current AI models can run experiments but lack scientific self-criticism and genuine creative thinking. The best performer, GPT-5.6 Sol, reached only 15 percent of the human reference score.

The Verge is starting a series about running AI models locally on personal hardware, driven by privacy concerns over sending personal data to cloud services. The reporter plans to document whether powerful local models justify spending heavily on a computer with lots of RAM.

arXiv will limit authors to two submissions per person per month starting October 2026, citing a surge in AI papers and strain on volunteer moderators. Monthly submissions have doubled in two years, with cs.AI growing more than sixfold.

Microsoft CEO Satya Nadella has adopted the term "Super Intelligence" in place of artificial intelligence, aligning with President Trump's preferred language, while arguing in a new blog post that AI models should be treated as insider security threats.

Odyssey has opened a free public research preview of Odyssey-3, a 14-billion-parameter world model that generates interactive environments from text prompts in real time. It can also drive robots, drones and video games via specialized controllers, though the company's top physics benchmark claim comes with methodology caveats.

Dwarf Fortress now uses version control, ending a long stretch in which co-creator Tarn Adams publicly said he did not use it. Adams referenced "version control stuff to use now" in a March 2023 interview tied to the game's Steam release.

The actions/python-versions repository now includes Python 3.15.0, meaning GitHub Actions workflows can reference "3.15" in their test matrices to run against the new release.