Anthropic admitted that engineering missteps caused a monthlong decline in Claude Code performance, affecting paying users who reported the tool as less capable and more expensive to use. The company'…
Practices built to compensate for today's AI model have a half-life determined by the model's improvement rate, according to Stack's analysis of agent engineering trends. The last year saw labels shif…
Boris Cherny, head of Anthropic's Claude Code, said prompt engineering is largely not important and that users should focus on giving Claude hard tasks and enabling verification. In an interview with …
Anthropic engineer Thariq Shihipar revealed at the AI Engineer World's Fair that the team removed over 80% of Claude Code's system prompt for Opus 5 and Fable 5—shrinking it from roughly 2,686 words t…
Anthropic deleted over 80% of Claude Code's system prompt for Opus 5, and the model became more intelligent, according to Claude Code creator Boris Cherny at YC Startup School 2026. Cherny shared a pl…
SemiAnalysis reported on 5 February 2026 that Claude Code authored 4% of GitHub public commits, projecting it will exceed 20% of daily commits by end of 2026. Ryan Dahl, creator of Node.js, stated 'th…
Anthropic's Claude Code, launched as a CLI in February 2025, became a developer phenomenon after a native VS Code extension was released on September 29, 2025, as part of Claude Code 2.0. By February …
Boris Cherny, the creator of Anthropic's Claude Code, said at a Y Combinator event on Saturday that users should stop giving overly detailed, step-by-step instructions to AI agents. Cherny advised des…
Y Combinator's Startup School drew over 6,000 attendees to San Francisco's Chase Center for a two-day conference featuring AI executives including OpenAI CEO Sam Altman, Nvidia CEO Jensen Huang, Meta'…
Loop engineering's two core pieces — run-until-done and maker/checker — are simple to wire up but easy to get quietly wrong, Anthropic's Boris Cherny said. Building a real-feedback loop from scratch i…
A 780 KB Rust proxy called cursor-bridge lets users run Anthropic's Claude Code against Cursor's backend, billed to their Cursor subscription instead of a separate Claude plan. The tool, which hit Hac…
Jeffy Loop, an autonomous code improvement loop for Claude Code, audits a codebase across quality dimensions, prioritizes tasks with runnable acceptance checks, and fixes one verified task per iterati…
An agent run is stochastic gradient descent over solution space, and steering is how the true objective enters the loop, according to a July 26, 2026 post that grounds the industry ladder of prompt, c…
Anthropic's Claude Opus 5 is proving significantly more resilient to prompt injection than its predecessors, according to internal red teaming and prompt injection evaluations buried in the model's sy…
Boris Cherny, an Anthropic employee, stated that Claude Opus 5 is the company's least prompt-injectable model yet, based on prompt injection evaluations and red teaming results detailed in the system …
A new paper finds that long chain-of-thought reasoning in AI models often fails to converge on an answer, with many traces lost long before they stop talking. The analysis, from researchers studying t…
Loop engineering is a deterministic control loop around an AI coding agent — schedule, durable state, and verification — that lets developers move up from prompting to designing the machine. Boris Che…
Loop engineering, a term coined by Addy Osmani in June 2026, is replacing prompt engineering as the dominant paradigm for coding agents, shifting focus from prompting models to designing autonomous lo…
Only 2 of 9 documented production use cases qualify as truly self-improving loops, according to a new analysis by Loadbearing Tech. The analysis highlights Replit's product development agent and an Op…
Only 2 out of 9 self-improving AI agents running in production actually have a verifier grounded in real-world data, according to an analysis by an unnamed expert. The verifier is the key bottleneck, …