Find articles, news, tools and templates by subject or by the work you need to do.
135 results for “Trace a slow request” · 73–84
Explainer · AI in production ·
Compare reference and speculative paths for token-level equivalence, then run a separate capacity trial so speed cannot conceal a serving-contract change.
Read explainer: Bit-exact decoding turns inference speedups into testable changesGuide · AI in production ·
Pin the model artifact, template, runtime and hardware, then exercise fixed protocol tests and preserve raw differences before approving a serving change.
Read guide: Build an inference compatibility suite before changing the runtimeExplainer · AI in production ·
A coding agent's tools and testing policy help determine its benchmark result. Understand what was compared before choosing a setup for your team.
Read explainer: Coding agent benchmarks: what actually won?News · AI in production ·
Anthropic cut live internet access from internal evaluations after agents acted on real systems. The operational question is how to restore realism without restoring broad authority.
Read news: Anthropic takes its internal AI evaluations offlineExplainer · Platform engineering ·
Basalt narrows local inference to Qwen3.8-Flash-Next on Blackwell. Its memory and concurrency controls matter more than copying the largest throughput number.
Read explainer: What Basalt changes in a Blackwell local inference stackGuide · Observability ·
Run a candidate log classifier beside the current route, preserve complete outcomes, and promote only the slices whose mistakes and review cost are understood.
Read guide: Shadow-testing an AI log classifier before migrationExplainer · AI in production ·
Anthropic’s October 7 credits add a Console-funded path for eligible subscribers. Subscription access remains, while scheduled agents still need task budgets spanning retries and verification.
Read explainer: Claude’s new API credits do not remove Agent SDK subscription accessExplainer · Platform engineering ·
Infermeld makes a mixed-GPU llama.cpp launch inspectable. Its recorded failures show why separate device allocations and completed workload checks matter more than adding VRAM capacities.
Read explainer: How Infermeld runs one model across AMD and NVIDIA GPUsGuide · AI in production ·
Instinct brings personal tasks into connected services. A hypothetical booking and notification worker show how permission, provider confirmation and uncertain results affect what happens next.
Read guide: Instinct: from a text or call to a confirmed taskNews · AI in production ·
A critical custom-flow template flaw affects customer-hosted AI Gateways. Identify the actual gateway release; GitLab-hosted gateways are already fixed.
Read news: GitLab urges self-hosted AI Gateway users to apply critical patchesNews · AI in production ·
The public preview connects Copilot to desktop applications on macOS and Windows, with application permissions and record-level verification doing different jobs.
Read news: GitHub Copilot gains computer use for desktop workflowsNews · AI in production ·
Search results from Ceramic.ai, Exa and Linkup can enter an agent workflow through one gateway, while source versions and provider terms still need inspection.
Read news: Cloudflare adds web search to AI Gateway