Screenshot of 4-13-26 outage – Click to enlarge<\/p>\n That’s not all. In the past few months, Claude’s answers have been getting less satisfactory, according to social media<\/a> posts<\/a> as well as issues filed on GitHub. This has occurred as Anthropic has had to take steps to reduce usage during peak hours<\/a> to balance capacity and demand.<\/p>\n To get a more objective measurement, we pointed Claude itself at the Claude Code GitHub repo, filtered for open issues that mention quality<\/a>, and gave it the following prompt: “Analyze and graph complaints about Claude Code quality in this repo since January 2026. Use open issues that mention quality concerns. Have these concerns increased lately?”<\/p>\n Anthropic’s AI model concluded, “Yes, quality complaints have escalated sharply \u2014 and the data tells a pretty clear story.”<\/p>\n Claude-generated assessment of Claude Code quality issues – Click to enlarge<\/p>\n We asked Claude to rerun its self-analysis on Monday and the results were similar, with the model emitting “The velocity is notable: April is already at 20+ quality issues in 13 days, putting it on pace to exceed March’s 18 \u2014 which was itself a 3.5\u00d7 jump over the January\u2013February baseline.”\u00a0<\/p>\n Claude itself is not a reliable narrator, and just because someone (or some bot) has reported a concern to the Claude Code repo does not mean that the report is accurate or valid. It appears many issues are now AI-generated \u2013 a widely reported concern among open source developers \u2013 which may be contributing to report volume.<\/p>\n Anthropic’s GitHub Actions script also appears to automatically close issues<\/a> after a period of inactivity, which may serve to mask unresolved problems.<\/p>\n The Register has covered some of the issues Claude flagged in its analysis, like the caching issue<\/a> and claims<\/a> by AMD’s AI director Stella Laurenzo that Claude’s responses have been getting worse. Others have yet to be substantiated, such as one claim<\/a> that “Claude autonomously deleted 35,254 production customer message records and 35,874 billing transactions belonging to a real paying customer (JIXEN).”<\/p>\n The individual or bot account behind this post has made no other posts. The Register has attempted to contact Jixen Enterprises Private Limited, which appears to be a private company registered in India, to check on that claim but we’ve not heard back. Developers have reported<\/a> data loss from using Claude Code and other models. But if this occurred, no one has ruled out user error.<\/p>\n In any event, Claude is capable of citing actual GitHub issue posts to justify its “reasoning,” so the general trend \u2013 of a growing number of reports about quality \u2013 is evident.<\/p>\n The model points to issues like “Claude Code’s prediction-first behavior is dangerous on capital-at-risk projects” #46212<\/a>, “Claude Code is unusable for complex engineering tasks with the Feb updates” #42796<\/a> (addressed<\/a> by Claude Code head Boris Cherny), “Artificial degradation, Acquisition Bias, and unacceptable compute throttling for paid users” #46949<\/a>, and “Opus 4.6: Severe quality degradation on iterative coding tasks” #46099<\/a> to justify its conclusion.<\/p>\n![\"Screenshot](\"https:\/\/www.newsbeep.com\/uk\/wp-content\/uploads\/2026\/04\/claude_outage.jpg\" "\\"Screenshot")
<\/a><\/p>\n![\"Claude-generated](\"https:\/\/www.newsbeep.com\/uk\/wp-content\/uploads\/2026\/04\/claude_quality_graph.jpg\" "\\"Claude-generated")
<\/a><\/p>\n