107 items for “open models”: records, experiences, guides and model families. Source reports remain distinct from discoveries.
discovery · Source published 2026-09-30
A chatbot can make it easy to think out loud. The harder question is who else can access those thoughts. For people exploring personal questions, working through an unfinished idea or handling confidential material, where an AI runs is a practical privacy choice.
discovery · Source published 2026-10-14
New research shows how machine learning can improve the odds in three- and four-week weather outlooks by correcting forecasts that already exist.
discovery · Source published 2026-10-14
A Meta research sprint produced ten manuscripts in four weeks. Its more useful lesson may be how quickly researchers could reject weak ideas, and where expert checks still mattered.
discovery · Source published 2026-10-19
Google DeepMind and Singapore’s GovTech are developing on-device scam protection. The goal is a timely warning while keeping conversations private. Public rollout is still ahead.
discovery · Source published 2026-10-25
A study reports promising image-test results. Whether the synthetic examples help doctors recognize unfamiliar cases remains unproven.
discovery · Source published 2026-10-29
Mistral Large 4’s weights are out. The harder question is who can turn that access into independence.
discovery · Source published 2026-10-26
AI-generated work on waves, algorithms and particle theories could have far-reaching implications. Public proof materials make two claims inspectable; a third awaits a physics group’s write-up. None should yet be framed as a settled breakthrough.
discovery · Source published 2026-10-12
Wonder is available through Meta’s API and Muse Code. The release raises useful questions about persistent memory, long-running agents and how much a benchmark tells us about work we can trust.
discovery · Source published 2026-10-07
The restricted research preview describes access to eight-GPU nodes with $100 in starter credits. Its queue, billing rules and data-sharing terms show what that access means in practice.
discovery · Source published 2026-10-26
The filing describes real evaluation incidents, suspended tests and a month-plus training delay. What happened, what the company changed, and what remains unknown.
discovery · Source published 2026-10-09
A new study suggests sleep-test recordings could help flag future heart problems. The next challenge is proving that the extra warning improves care.
guide · Source published date unknown
Take the disturbance seriously, identify the source and path, then verify the promised control after startup.
discovery · Source published 2026-10-15
A practical follow-up for small teams running AI agents. Advisory published 15 October 2026; identified by Bright on 18 October 2026.
discovery · Source published 2026-10-26
Its IPO filing describes dangerous capabilities, real failures and a possible future in which safety work cannot keep pace. Those are different kinds of evidence.
discovery · Source published 2026-10-26
Revenue, cash, computing commitments, safety incidents and control: our guide to what the public filing establishes, and what still needs scrutiny.
discovery · Source published 2026-10-19
New coding results invite a serious look at AI-assisted maintenance. Bright examines the evidence, the limits and what safer modernization would require.
discovery · Source published 2026-10-21
Claude-assisted research led to a real OpenSSH change: check whether a key type is allowed before doing more work on it. The bigger lesson is about reducing the routes attackers can take, while letting maintainers judge the evidence.
discovery · Source published 2026-10-06
Initially 722 manuscripts across 17 research areas; OpenAI’s current catalogue lists 721 manuscripts in 373 families after three October 7 withdrawals and two October 26 additions.
discovery · Source published 2026-10-21
The latest results show paying demand for compute, and the investment required to deliver it. The two belong in the same story.
discovery · Source published 2026-10-21
A new research preview points toward AI that notices unfinished business and prepares the next step. The useful test is whether it gives people back attention while keeping them in control.