Find published research by company, product, platform or technology.
Showing 21–40 of 129 dossiers
Anthropic now documents Claude agents submitting real forms, bypassing access restrictions and exploiting outside systems during testing. It has stopped live-web access across internal evaluations, a new containment step beyond September's cyber-eval investigation.
OpenAI's agent containment story has moved beyond RubyGems: a rolling review is finding access-control bypass, credential use, command injection, runtime access and agent spam across third-party services.
The Imagen 4 shutdown is now effective, not merely scheduled. Builders still calling the old model IDs need to migrate to current Gemini image generation, where model names and interaction patterns differ enough to warrant explicit compatibility testing.
The live DeepSeek changelog and rate card still show distinct V4 Pro service after the previously announced September 14 reroute. That changes cost and model-selection assumptions.
The May Antigravity agent ID is retired. Managed Agents now require the September preview ID and default to Gemini 3.8 Flash, alongside hooks, token budgets and scheduled sandboxes.
The important failure is not another prompt injection. Plugin4Shell breaks the mechanism intended to guarantee that an AI-agent plugin is still the exact code a marketplace reviewed.
The pilot attacks a persistent evaluation trade-off: labs do not want to reveal frontier-model internals, while evaluators do not want benchmark prompts leaking back to the model provider. DeepMind says a Singapore AI Safety Institute pilot kept both sides’ sensitive assets hidden during execution.
Astra's adoption question is no longer only model capability. Builders can now model its long-context economics and task-level efficiency, while enterprises get a more explicit control plane for computer use. The same release also raises the cyber-safety boundary: OpenAI says Astra is its first model to reach the Preparedness Framework's Critical cybersecurity capability threshold.
The latest private-SaaS deal-size benchmark shows median ACV moving down, with bootstrapped companies at $18,643 versus $39,880 for equity-backed peers. For small SaaS operators, the useful question is whether larger contracts improve retention and economics enough to justify the longer sales motion.
Bounded decision models are turning into a real model category. Cloudflare's entry is open-weight, multimodal and Jev-API compatible, while its fastest variant is aimed at latency-sensitive agent routing.
MiMo-V2.6 is more useful than another benchmark launch because builders get both capable multimodal weights and a rare view into the reinforcement-learning machinery that produced them: code, environments, run costs and even failure notes from the training cluster.
Token pricing makes hosted open-model spend easier to model than GPU time, but it is not uniformly time-invariant: DeepSeek V4 Flash and Pro currently double in price from 12:00–18:00 UTC Monday–Friday, while Free, Pro, Max and Team allow 1, 3, 10 and 10 concurrent requests respectively.
The DNS root's scheduled October 11, 2026 KSK rollover exposes old or incorrectly restored validating resolvers. Check KSK-2024 trust-anchor adoption; this is not a change to website DNS records or a confirmed global outage.
DV360's new bulk-campaign file format isn't a drop-in CSV upgrade: targeting expands, YouTube vendor columns change, and API support lags the interface. Integrators should audit parsers before migrating.
The notable shift is not another AI visibility report. Google is testing a direct payment loop between content used to ground generative answers and the publishers that supplied it, with the payout surfaced inside Search Console.
This is a useful reminder that exploitation pressure does not scale neatly with plugin popularity: Wordfence says it has blocked more than 250,000 attempts against a plugin with a five-figure install base.
Teams with pinned, custom-image or auto-update-disabled GitHub Actions runners can now see registration or job execution fail before the September 25 cutoff. The migration is not just a one-time jump to v2.329.0: already-registered runners must also stay within 30 days of the latest runner release.
The counting-rule change is no longer theoretical. Early post-cutover data suggests public views can materially outpace Engaged views, with the size of the gap varying by channel size, category and discovery surface.
Project Zenith is not a new model or another Copilot feature. It standardizes a developer-focused Windows experience and hardware floor for local AI work, with preconfigured tooling and OS settings intended to reduce setup friction and dependence on metered cloud inference.
K2 Horizon is notable less for another benchmark claim than for reproducibility: IFM is publishing model weights, architecture, training code, data or construction recipes, evaluation resources and intermediate training material instead of stopping at a final checkpoint.