Name: Anthropic admits fault (Claude limits to be INCREASED)
Uploaded: 2026-03-31
Duration: 388 s
Description: Anthropic is admitting, in PR language, that Claude Code usage limits are burning way too fast, and limits will almost certainly be raised — but don't expect real transparency on request counts anytime soon.

Click to load the YouTube player

Key takeaways

Track your Claude Code burn in absolute requests, not percentages — log every prompt yourself until Anthropic ships a request counter like MiniMax.
Pause heavy Claude Code workloads (browser use, long agent loops) until Anthropic publishes revised Max/Pro limits, which the apology post strongly implies are coming within days.
Test the new `computer use` in Claude Code for app automation tasks — it now clicks native UI elements and renders pixel art, eroding the open-source Claude clones' main use case.
If you hit `way faster than expected` limits on Pro or Max, screenshot the percentage and post it in the community thread — the creator argues collective pressure is the only thing moving `senior managers` on compute allocation.
Treat Anthropic status posts (`actively investigating`, `still looking into it`) as `LinkedIn speak` and wait for a concrete number change in the dashboard, not a tweet.
If transparency on usage keeps stalling, start prototyping the same prompts on GPT-5/4.5 — the creator flags active user migration as Anthropic's real leverage point right now.
Don't burn a Max plan on one-shot tasks like trip planning right now — the creator's `50%` cap on `eight prompts` shows the new rate limiter is brutal on multi-step agent flows.

Video summary

Companion notes

Anthropic publicly admits Claude Code limits are being hit way faster than expected and will likely be raised.

## What users reported Pro and Max plan subscribers said their quotas were burning in hours, not days. One user described a job that previously ran to completion now stopping at 20%. The creator burned 50% of his weekly limit on eight prompts while asking Claude to plan a trip — and that included browser use, which is expensive. The common thread: nobody knows how many requests they're actually spending because Anthropic only shows a percentage.

## Anthropic's response Lydia from the Claude Code team posted on community channels: We are aware that people are hitting our usage limits in Claude code way faster than expected. We're actively investigating. She followed up with We're still looking into it. The creator reads this as LinkedIn-speak for Oops, we messed up, arguing that the phrase way faster than expected is an admission of guilt — Anthropic overcorrected, didn't introduce a bug, and is now backtracking from its earlier it's a feature, not a bug stance.

## The likely cause Compute is being diverted to a model the creator says is in training — Me False (Claude Mythos). Anthropic is balancing GPU allocation between the new training run and keeping current paying users functional. The fix is gated on how much compute senior managers approve to give back.

## What got released anyway 10 hours ago (relative to recording) Anthropic shipped computer use inside Claude Code: the model can now click on UI elements of native apps and even build pixel art. This directly undercuts the open-source Claude clones that previously owned the "control your desktop" niche.

## What the creator wants MiniMax-style usage tracking: show the actual number of requests, let users BYO API key, and stop treating loyalty as a captive audience. Until then, switching back to GPT is on the table.

## Bottom line Limits are going up because Anthropic can't afford the churn, but the opacity stays unless users keep pushing.

Watch on YouTube

Prefer the native player? Open it on YouTube: https://www.youtube.com/watch?v=WiAx9sPw69U