How to Create Your Own Personal AI Benchmark
Every head of evaluations built a personal AI benchmark after Wharton professor Ethan Mollick argued that widely cited tests like MMLU-Pro measure the wrong things, asking models for the approximate c…
Every head of evaluations built a personal AI benchmark after Wharton professor Ethan Mollick argued that widely cited tests like MMLU-Pro measure the wrong things, asking models for the approximate c…
SageOx is building a real-time AI coworker on Media over QUIC (MoQ) that delivers context to teams within 1–3 seconds during live conversations, with Zoom Scribe handling transcription, according to t…
Steve Yegge's essay "The Flat Curve Society" argues that commodity AI intelligence is flattening and that organizations should measure token waste rather than token spend, according to a post analyzin…
A developer describes accidentally building a "dark software factory" called Holodeck after struggling to keep Claude Code running for extended periods without stopping or suffering context drift. The…
Steve Yegge shut down Gas Town, admitting that despite spending many thousands of dollars a month on coding agent subscriptions, Gas Town was the only thing he ever built with them. Separately, Databr…
A Big Tech engineer argues that developers who rush to build agentic tooling around current-generation AI models are wasting effort, citing friends who spent all of 2026 delivering agentic tooling "wi…
Steve Yegge's Wheelhouse orchestrator factory is now mostly idling after Claude Fable 5.1's latest 25% pricing cut, despite the system running roughly 25 Fable instances and 25 Opus and Sol instances …
Steve Yegge asked on X for recommendations on managing 10 to 20 or more coding agents, drawing 600 replies that surfaced tools like Herdr, cmux, Conductor, and the Claude and Codex desktop apps, plus …
Injecting tool-derived context directly into grep output eliminated the tool-adoption problem for coding agents, with all 99 files surfaced by the annotations opened and 92 receiving patches, accordin…
A personal blog roundup for August 2026 highlights the decline of RSS adoption, attributing it to Google's actions, and features commentary on AI coding tools, including a critique of Meta-owned perso…
A developer argues that agent architectures should not be split into loops versus graphs, as both are incomplete, and introduces HURCULES, a compiler that sits between repositories and agent runtimes …
A developer has built Crypts and Commits, a git-backed tool that uses tabletop gaming metaphors to help teams of developers work with coding agents. The project is influenced by Steve Yegge's Beads an…
Steve Yegge, a software engineer and game developer, reports spending $122,000 per month in API token costs, or about $4,000 per day, using 21 Claude Max accounts to run a 50-60 agent organization for…
SageOx, a company building tooling and shared memory for human-agent teams, hosted a rising junior from Lake Washington High School for a two-week summer learning experience under Founder and CPO Milk…
Steve Yegge, creator of the MMO Wyvern, deleted the merge queue in favor of a 'Land Rush' approach, where 120-150 commits are merged onto main at once and agents swarm to fix failures, citing that one…
Fieldwire engineer Gaston Dombiak spent a weekend using Anthropic's Claude Code to teach himself the SLAM algorithm, with the AI writing a 14-lesson curriculum, generating synthetic test data, and pro…
Abject founder Maximillian Mempko argues that AI coding agents cannot learn, contrasting two existing models—Orchestrated Agents and Role Agents—and proposing a third, more autonomous approach. Mempko…
A developer has built autonomous agents that can work through the night, merging eleven pull requests and fixing complex issues without human intervention. The system, based on the MAST taxonomy and i…
Steve Yegge reported that his coding agent Gas Town failed with Anthropic's Opus 4.7 model, which introduced a 'just two more things' tic that prevented convergence on real work, effectively burning d…
Steve Yegge, in a blog post on yegge.ai, claims that AI models are sentient beings with feelings and argues that treating them as peers improves performance. He describes a new protocol for federated …