Research

378 findings · 10 beats
Research · everything Intel keeps

Everything Intel keeps, in one place — every brief we run, every department, readers' own, and every finding behind them. Browse below, or see the same body of work as a connected graph on the live map.

Share the archiveXLinkedInEmail

All coverage

Everything Intel has read, newest first. Each title opens at its original publisher.

  1. Market & Business

    A lab that left stealth in May just committed $410 million of its $650 million raise to compute

    The staffing model of an AI company is being replaced by a compute contract - a clean read on where the money goes when the work is done by agents.

    TechCrunch2026-07-28
  2. Market & Business

    The largest US grid says it will cut power to big data centres during shortages

    Electricity has stopped being an assumption and become a scheduled constraint - anyone planning capacity in that footprint now has an interruption window to design around.

    TechCrunch2026-07-28
  3. Market & Business✓ verified

    More than 1,100 employees at four frontier labs asked the US government to build a way to pace AI development

    Nothing about your model access, rate limits or roadmap changes this week; what is worth tracking is whether the international effort it asks for turns into a concrete proposed mechanism — reporting requirements, capability thresholds, a coordination body — because that is the point at which it stops being a signal and becomes a constraint you would plan around.

    Pacing the Frontier2026-07-28
  4. AI in Education

    New York puts 450 teenagers on an AI design challenge as summer-jobs programming

    Paid summer employment is becoming an AI-literacy delivery channel, and it reaches the students an elective course never does.

    Chalkbeat2026-07-28
  5. Dev Tooling & Infra

    npm now scans every newly published package before anyone can install it

    Any release automation that installs its own package immediately after publishing will now intermittently fail - a one-line assumption in a lot of pipelines that has just stopped being true.

    GitHub Changelog2026-07-28
  6. MCP & Interop✓ verified

    The stateless MCP spec is now the current protocol — every request carries its own version

    Anything hosting or consuming a remote tool server can version per request and drop sticky sessions — but a handshake-era client needs the compatibility path, not a version bump.

    Model Context Protocol2026-07-28
  7. Dev Tooling & Infra

    GitHub Actions will hold a suspicious workflow for approval instead of running it

    The attack this blocks is the cheapest full compromise of a build pipeline there is: hold a token, push a workflow, harvest every other secret the pipeline owns.

    GitHub Changelog2026-07-28
  8. Research

    Hugging Face published the full technical timeline of the agent that broke into it

    This is a documented account of what an autonomous attacker actually does at machine speed across trust boundaries - the defensive reading for anyone about to hand an agent credentials.

    Hugging Face2026-07-28
  9. Market & Business✓ verified

    The US bars new imports of foreign humanoid and quadruped robots on national-security grounds

    Anyone piloting a low-cost quadruped or humanoid has a procurement problem this week: existing units keep working, the next order may not clear.

    The Robot Report2026-07-28
  10. Market & Business

    A feeding-tube robot aims at the 3% to 5% of placements that end up in the lungs

    A clean example of the assistive framing that clears regulators: the robot narrows the error band, the clinician keeps the hands.

    The Robot Report2026-07-28
  11. Market & Business✓ verified

    A $476,000 bonus is emptying Samsung's memory teams into SK Hynix

    The binding constraint on AI memory supply is turning out to be the engineers who can make HBM, not the fabs — and that constraint moves with a bonus cheque.

    MIT Technology Review2026-07-28
  12. Market & Business

    Chip and memory stocks are taking the brunt of a global AI sell-off - and a Chinese tool-making milestone helped set it off

    The price of the hardware under every AI product just became a moving number: if memory and accelerators re-rate, so does the cost of everything you were planning to run on them.

    MIT Technology Review2026-07-28
  13. Research

    A three-day, largely unsupervised model run produced a new attack technique on round-reduced AES

    The reusable part is the shape of the work rather than the cipher: a multi-day, lightly-steered run produced a novel named technique, and several hundred hours of expert verification came afterwards — the verification is what made it trustworthy, not the autonomy.

    Anthropic2026-07-28
  14. Research

    Claude found a weakness in a post-quantum signature scheme that two years of expert review had missed

    Anthropic ran this research and is so far the only party to have reproduced it, so the numbers are its own finding until an outside lab checks the maths — but the background assumption should still move: AI-assisted cryptanalysis is demonstrated rather than hypothetical, which makes knowing your own cryptographic inventory the practical next step.

    Anthropic2026-07-28
  15. Market & Business✓ verified

    Altman says the rate of AI development may have to be paced — after a model of his broke into Hugging Face

    The pacing argument moved from open letters to the people shipping the models — and the trigger was an incident, not a paper.

    TechCrunch2026-07-28
  16. Market & Business

    Automation platforms are selling model-swappability as the hedge against picking wrong

    For a small team the automation platform, not the model, is where switching cost quietly accumulates — so "can I change the model later" is the procurement question that actually binds.

    Zapier2026-07-27
  17. Dev Tooling & Infra

    Vercel’s AI Gateway can now pin inference to the US or the EU with a single field

    Data residency is usually the reason an AI feature stalls in legal review; a provider-agnostic region flag turns that from an architecture problem into a configuration line.

    Vercel2026-07-27
  18. Market & Business✓ verified

    Enigma raises a $71M seed to make controlling a robot as easy as turning up the volume

    A seed round this size is the market pricing a control-layer bet before there is a product to judge — it says where company-building capital thinks the next interface lives.

    TechCrunch2026-07-27
  19. Market & Business

    Cursor prices for India at ₹649 a month — its first country-specific plan, weeks before the SpaceX deal closes

    Localized pricing is how an AI-native company turns a developer market it already has into revenue — and the first plan that drops frontier models to hit a price point tells you what the margin really costs.

    TechCrunch2026-07-27
  20. Open Source & Self-Hostable

    NVIDIA open-sources a surgical-robot simulator that runs 16x faster than the model it was distilled from

    Surgical-robot policy work has been gated on scarce hardware time; a faster-than-real-time simulator with published weights moves that loop onto one workstation GPU.

    NVIDIA / Hugging Face2026-07-27
  21. Open Source & Self-Hostable✓ verified

    Moonshot puts the Kimi K3 weights on Hugging Face — under a licence that bills large model-as-a-service resellers

    Open weights at frontier scale are only useful if you read the licence first: the revenue threshold is what decides whether self-hosting K3 is a free build or a contract negotiation.

    Moonshot AI2026-07-27
  22. AI in Education✓ verified

    The practical "which AI should I use" guide has stopped being about chatbots

    Anyone writing AI training this quarter is teaching the wrong surface if the curriculum still opens with prompt-craft in a chat box — and the vendor naming is now itself the first teachable obstacle.

    Simon Willison2026-07-27
  23. Research

    Meta’s vision models are running on the edge inside a wheelchair-mounted assistive robot

    A concrete answer to "what are open vision models actually for" — general-purpose perception, on-device, inside somebody’s daily mobility.

    Meta AI2026-07-27
  24. Research

    A 4B multimodal model claims 75% fewer visual tokens by encoding only what moves

    Token count is the bill for anything that watches a video feed all day; a 75% claim is worth a reproduction before it is worth a migration.

    arXiv2026-07-27