Find published research by company, product, platform or technology.
Showing 1–20 of 222 dossiers
Rashomon's experimental local recorder can expose discrepancies between a coding agent's closing claims and its tool execution. It is an observability aid, not a sandbox or tamper-proof security product.
The important change is not simply that Claude can run several agents. Projects now owns decomposition, shared context, branch isolation and progress coordination across full Claude Code sessions, while the trade-offs become usage burn, cloud-only execution and ordinary merge conflicts when parallel work overlaps.
A restricted private-preview detector changes the compliance workflow, but Claude watermarking still cannot prove authorship or reliably detect short or lightly edited passages.
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.
Haiku 5.5 resets the economics of high-volume classification, extraction and agent sub-tasks, while Anthropic also cuts Sonnet 5.5 cache-read prices and introduces API credits for Max/Team subscribers.
The Anthropic procurement fight changed materially on September 25: a 2–1 federal appeals-court ruling backed the Pentagon’s supply-chain-risk designation. Builders serving defense customers should no longer rely on the August district-court ruling as evidence that the Claude procurement barrier is gone.
The interesting change is economic as much as benchmark-driven. Anthropic is compressing capability that previously justified its larger Fable tier into Opus pricing, while cutting Opus list prices and expanding immediate availability across the major clouds.
The material issue is not ordinary model distillation. Anthropic’s evidence suggests a customer-facing AI product may have used a rival model as an undisclosed backend while simultaneously harvesting those interactions for training, turning routing architecture into a privacy and trust boundary.
Together Link connects six existing coding-agent/desktop harnesses to open models with reversible profiles, per-session routing and cost receipts. The important shift is portability at the harness boundary, not Together's unverified savings claim.
A detailed engineering case study links a 68% AI-review cost reduction to prompt-cache structure and effort tuning, with security-review trade-offs still unresolved.
A hard shutdown date makes this more than a hosting refresh: unmigrated Codeberg Pages sites will go offline, and the replacement changes deployment and URL behaviour.
Two distinct access paths matter more than the benchmark headlines: Muse Wonder is available through Meta Model API and Muse Code, while Wonder Full is currently subscriber-only in Muse Code.
Pi’s first stable release is interesting less for another coding-agent version number than for what its deliberately minimal core now considers mature enough to include: MCP, code-driven tool orchestration and model routing.
The Agent Host’s environment boundary has moved from local Dev Containers to remote development hosts, making persistent coding-agent sessions more portable across real remote projects.
The scanner itself is not the new part. The September 16 change removes the CodeQL-default-setup gate that GitHub’s July rollout originally required, making AI-assisted vulnerability detection easier to add to repositories with different code-scanning configurations.
The important shift is that agent orchestration itself becomes a managed API surface: context compaction, tool discovery, programmatic tool calls and subagent coordination can now come from OpenAI’s maintained Codex harness rather than an application team rebuilding those layers.
Copilot code review now moves from advisory assessment toward a governed merge gate. The public preview remains off by default, and GitHub’s current docs let administrators separate AI approval itself from whether that approval counts toward required-review policy.
The interesting part is not another AI scanner. Cloudflare is connecting source-code evidence to what is actually deployed and being attacked at the edge, validating findings outside the model, then preparing both a code patch and, where appropriate, a narrowly scoped WAF mitigation for customer review.
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.
The August 28 transition is now active, and Railway’s current documentation removes an earlier ambiguity about new services in existing projects. Config as Code is legacy-only from here; production users should migrate and validate `.railway/railway.ts` before the December hard cutoff.