Is Claude AI acting dumb right now?

ClaudeDown.com tracks real-time public complaints about Claude AI on Twitter/X to detect quality regressions. Visit the site to see the current status.

How does ClaudeDown detect Claude AI issues?

We monitor public Twitter conversations about Claude AI using keyword-matched complaint detection, compare against historical baselines, and use statistical spike detection on a 6-hour rolling window.

Is Claude being dumb right now?

Yes.

Something appears broken.

5.2% complaint rate (3hr avg)

Claude Code v2.1.212

AI-generated summary

Complaint Timeline

3hr avg · 7 days · UTC

What share of all Claude mentions on X are complaints? Higher means more people are having problems. All times UTC.

Complaint rate

Elevated (>3%)

Critical (>5%)

In progress

Want to know when things go wrong?

No spam. Only spike alerts.

Daily Pulse

Last 30 days · UTC

How does each time slot compare to its usual level? Hover or tap any cell for details. All times UTC.

Normal

Elevated

Pending

06-18

06-21

06-24

06-27

06-30

07-03

07-06

07-09

07-12

07-15

12am

3am

6am

9am

12pm

3pm

6pm

9pm

Cell color shows complaint share relative to the median for that time of day. 1.0x = typical. Above 2.0x = elevated. Tooltip shows both the complaint rate % and the relative level.

Recent Tweets

A sample of recent tweets about Claude AI, sorted by engagement. All dates UTC.Sample of recent Claude tweets (UTC).

TheStalwart@TheStalwart

07/10

I used to cynically think “AI writing sucks, but it’s better than 90% of people at it.” But I’m less sure. I think it’s gotten worse. At least in the Anthropic family, 4.8 and Fable 5 outputs are so larded with Claude-isms I increasingly get straight up unintelligible outputs.

376 likes7 reposts45 replies

View on X →

_clarktang@_clarktang

07/13

I think there is general confusion around how AI works, AI tokenomics, and ultimately *what is actually priced in* for the AI trade - and that some of the existing arguments are at odds with one another Firstly to clear this up - what Brad and Gavin are saying are completely in ... show more

85 likes3 reposts2 replies

View on X →

shortmagsmle@shortmagsmle

07/11

I’m not a big fan of “modern life sucks” discourse because I’m a romantic and an optimist, but when I see an image like this and think about how a guy of yesteryear could have this view all day with no idea what “Claude code” was…maybe things were better https://t.co/MwCrbXQQKt

62 likes3 reposts5 replies

View on X →

CryptoTweets@CryptoTweets

07/10

⚡️DAILY RECAP⚡️ • 🇺🇸 Apple sues OpenAI for allegedly stealing trade secrets. • 🇺🇸🇮🇷 US imposes new sanctions on Iran. • 🇺🇸🇨🇳 OpenAI and Google are selling AI access to Pentagon-blacklisted Chinese firms, FT reports. • 🇨🇳 China considers restricting foreign access... show more

49 likes5 reposts9 replies

View on X →

nikesharora@nikesharora

07/10

Summary: I spent time trying to figure out this orchestration layer problem, can we design a multi model architecture in the long term. The more I dug in the more I understand that trying to build an abstracted layer is hard. As agentic activities increase and agent chaining and ... show more

115 likes11 reposts22 replies

View on X →

jerallaire@jerallaire

07/16

The Convergence and the Decomposition of the Firm — from The Agentic Economy Every major platform shift of the internet era arrived the same way. Not as a single invention, but as several technologies that had matured separately and suddenly came together. The web was one such ... show more

72 likes14 reposts16 replies

View on X →

ren_stocks@ren_stocks

07/13

Rosenblatt is bullish photonics and you’re selling? $LITE $AAOI $COHR $FN $CIEN The mid-quarter weakness in optical isn’t a broken thesis, it’s a better entry. The whole bear case rests on China ramping indium phosphide capacity and crashing laser prices, except Rosenblatt’s o... show more

57 likes4 reposts3 replies

View on X →

cyrilXBT@cyrilXBT

07/14

MOST PEOPLE'S AGENT INSTRUCTIONS GET WORSE THE MORE THEY ADD TO THEM. Every time something goes wrong, they add another rule to the same file. Weeks later, it is a dense legal document the model can no longer hold in working memory so it pattern-matches on whatever feels most re... show more

48 likes11 reposts5 replies

View on X →

About ClaudeDown

Read the launch post

Built by Prabal, Saransh, Sarthak

Have you ever wondered if Claude is actually getting worse, or if it's just you? ClaudeDown tracks public complaints about Claude AI across Twitter/X in real time, so you can see whether other people are experiencing the same issues.

Every hour, we collect complaint tweets, compare the volume against historical baselines for the same time of day, and flag when something looks off. The complaint-to-mention ratio helps separate genuine regressions from normal background noise.

This is a community project. The data is estimated, not exhaustive. But when hundreds of people independently start complaining about the same thing at the same time, that signal is hard to ignore.

Complaint volumes and baselines are estimates based on sampled Twitter data. Spikes are detected algorithmically and may include false positives.