Fable 5 vs Opus 5 in Claude Code: early numbers and vibes
Same 12 tasks I run on every release. Fable took 9, Opus took 2, one tie — but the interesting part isn't the score.
Discussion
Claude Code2.4.1
8
1h ago
DiscussionClaude Code2.4.181h ago
The unstructured half of the manual: build logs, postmortems, weekly trials, and honest arguments. The good parts get promoted into records.
Same 12 tasks I run on every release. Fable took 9, Opus took 2, one tie — but the interesting part isn't the score.
Four ways to run the same agent. Which one should someone actually start with, and why?
Show me what actually works. Not the aspirational stuff — the rules that survived contact with a shipping iOS app.
Half my company lives in Cowork now. Every week someone asks me why it won't touch the repo.
This week's trial: same spec for everyone, any agent, working offline sync required. 34 submissions and counting. Due in 3 days.
We let an agent do first-pass triage on pages for a quarter: real numbers on time-to-context, the two incidents it made worse, and where the human stays load-bearing.
Not benchmarks, not demos — daily driving. Qwen3 Coder on a 4090 versus the API bill: where local actually holds up, where it quietly fails, and the routing setup that makes it viable.
Mine grew to 600 lines and the agent ignores most of it. Time to talk about what earns a slot in the most expensive real estate in your repo.
Compaction ate my constraint again. Practices for deciding what stays in context, what gets externalized to files, and when to declare session bankruptcy.
The agent worked for two hours and produced something plausible everywhere. Rubber-stamping is negligence, line-by-line is a full day. What's the actual workflow?
Six months of actual numbers from a 12-person team: cycle time down, rework up, and the metric that predicted value wasn't the one we bet on.
Not bugs — slop. Needless abstractions, defensive try/catch on everything, comments narrating the obvious. It passes review individually and degrades the codebase collectively.
Four parallel agents sounds like a 4x multiplier until they meet in the same file. Worktrees, task routing, merge order — what's your actual topology?
Our new grad ships like a mid-level and I have no idea what he actually knows. The apprenticeship loop assumed you learn by writing — that assumption just quietly broke.
Five days, two agents, one working Stripe billing portal: what was delegated, what broke, what it cost ($61.40 all-in), and the day everything went sideways.
Eight days migrating a legacy widget suite to WidgetKit with Claude Code and XcodeBuild MCP: timeline budgets, a haunted pbxproj, and $84 of tokens.
A rehearsal of a column-drop migration ran against production because of an environment variable that outlived its shell. Nothing was lost — 41 minutes of degraded writes and several illusions were.
An agent debugging a staging failure copied a .env file into a test fixture; the PR shipped to a public repo. Keys revoked in 71 minutes. The scanner that should have caught it was configured to ignore fixtures.
An agent 'improved' a dry-run-by-default cleanup script; three weeks later a human ran it the way the old one worked. 4.2M report files gone. Neither alone would have done it.