Claude Code v2.1.317 is mostly Claude asking before it does things you didn't ask for.
- Asks before a web or shell tool call sends your account email to the network
- /schedule warns when a new cloud task picks up connectors you didn't request
- Agent SDK hooks get onFailure: "block", so a crashing hook stops the action instead of waving it through
Quiet release. Good defaults.
OpenAI built a new Git backend because agents push a LOT.
Pando: ~80% lower push latency, nearly a petabyte of Git data served a day. The stated goal: grow Git capacity alongside agent usage.
Yesterday Shopify went monorepo, also citing agents.
Soon your repo's busiest contributor won't be human.
Monday dev life:
Open laptop.
Check which AI tool changed its limits over the weekend.
Check which one changed its pricing.
Check which one got acquired.
Start actual work at 11.
2026 architecture decisions:
- Monorepo? Because agents.
- Back to multi-repo? Because agents.
- Redux comeback? Because agents.
- IaC, or no IaC? Because agents.
I genuinely can't tell anymore if we're designing for humans or for the bots.
Last week: Claude Code renamed its limit. OpenAI merged ChatGPT and Codex into one allowance.
This week: Grok Bot's cost to serve "was unsustainably high relative to revenue," per Elon.
Three tools, one lesson. Agents eat compute faster than subscriptions pay for it.
Which tool's limits do you hit first?
Claude Code renamed its limit, killed a fast mode and fenced in file writes this week.
OpenAI folded Codex into the ChatGPT allowance.
Feels like every tool is finally showing us the real compute bill.
ChatGPT and Codex will share ONE usage limit starting in December.
- Plus, Pro and Business get a single allowance across both
- Measured in tokens, so bigger models and harder tasks burn more
- Instant chat doesn't count
- Combined usage and reset times in Settings β Usage
Your Codex budget now competes with your chat habit.
Scariest things in a dev's terminal π
π» "Compacting conversation..."
π An agent that says "Done!" with 0 files changed
πͺ rm -rf with a variable you didn't check
β³ "You've reached your usage limit. Resets in 4h 59m"
π§ git push --force on a Friday
What's on yours?
@Hesamation The pelican-on-a-bike test is the only benchmark I still trust my eyes.
Feathers, spokes and chain all line up. Is that one shot, or did Max get a long thinking budget?
Heads up if you use fast mode on Opus 4.8 in Claude Code.
v2.1.316 deprecates it. On Nov 30 it's removed, and fast mode requests on Opus 4.8 return an error. No silent fallback to standard speed.
Want to keep the speedup? Move to Opus 5.5.
You have a month. Don't find out from a failed CI run.