Top rising problems today
What people in data, tech and AI struggle with, from public posts: 332 problems so far from 1,735 posts, +498 posts this week across 146 companies and products. Scanned daily.
Problem posts per day
Rising today
All rising| Problem | This week | ||||
|---|---|---|---|---|---|
| Service outages and slowdowns disrupt workOutages and change | 13d activesince Aug 26 | +1Seen on Reddit, fru.dev sites, GitHub issues, Hacker News, App Store reviews | +1010 posts in the last 7 days, 5 the 7 days before (rising) | ||
| Upgrades and new releases break working setupsOutages and change | 14d activesince Aug 27 | Seen on GitHub issues, Reddit, Community forums | +77 posts in the last 7 days, 2 the 7 days before (rising) | ||
| Context windows run out on real workloadsLLM apps | 8d activesince Aug 28 | Seen on GitHub issues, Hacker News, Reddit, Community forums | +66 posts in the last 7 days, 0 the 7 days before (rising) | ||
| Everyday tools feel slow and laggyDev tooling | 11d activesince Sep 2 | +1Seen on Reddit, GitHub issues, App Store reviews, Mastodon, Hacker News | +77 posts in the last 7 days, 2 the 7 days before (rising) | ||
| Login, SSO and OAuth flows breakDev tooling | 7d activesince Aug 30 | Seen on GitHub issues, Reddit, App Store reviews | +44 posts in the last 7 days, 1 the 7 days before (rising) | ||
| Vulnerability scanners flood teams with noiseSecurity | 16d activesince Aug 27 | +1Seen on fru.dev sites, Reddit, GitHub issues, App Store reviews, Community forums | +77 posts in the last 7 days, 5 the 7 days before (steady) | ||
| Fine-tuning is costly and results are unpredictableMLOps | 5d activesince Sep 14 | Seen on GitHub issues, Hacker News, Reddit | +44 posts in the last 7 days, 1 the 7 days before (rising) | ||
| Deprecations force unplanned migrationsOutages and change | 7d activesince Sep 1 | Seen on GitHub issues, fru.dev sites, Hacker News, Reddit | +44 posts in the last 7 days, 1 the 7 days before (rising) | ||
| Bad data slips into production unnoticedData engineering | 7d activesince Sep 1 | Seen on GitHub issues, DEV, Reddit, Stack Overflow | +33 posts in the last 7 days, 0 the 7 days before (rising) | ||
| DuckDB upgrades break working setupsOutages and change | 5d activesince Sep 17 | Seen on GitHub issues | +55 posts in the last 7 days, 2 the 7 days before (rising) | ||
| Dagster backfills are slow and riskyData engineering | 3d activesince Sep 21 | Seen on GitHub issues | +33 posts in the last 7 days, 0 the 7 days before (new) | ||
| Error messages are cryptic and unhelpfulDev tooling | 16d activesince Aug 28 | Seen on GitHub issues, Stack Overflow | +44 posts in the last 7 days, 4 the 7 days before (steady) | ||
| Small files, skew and partitioning hurt performanceData engineering | 9d activesince Aug 28 | Seen on GitHub issues, Reddit | +33 posts in the last 7 days, 0 the 7 days before (rising) | ||
| Tools are clunky and unintuitive to useDev tooling | 5d activesince Sep 1 | Seen on GitHub issues, Reddit, App Store reviews, Hacker News | +33 posts in the last 7 days, 1 the 7 days before (rising) | ||
| Claude fails to return valid structured outputLLM apps | 4d activesince Sep 7 | Seen on GitHub issues | +33 posts in the last 7 days, 0 the 7 days before (rising) |
Persistent problems
All persistent| Problem | This week | ||||
|---|---|---|---|---|---|
| No reliable way to evaluate LLM app qualityLLM apps | 15d activesince Aug 27 | Seen on GitHub issues, Hacker News, Reddit | +33 posts in the last 7 days, 3 the 7 days before (steady) | ||
| Breaches expose customer and company dataSecurity | 15d activesince Aug 27 | Seen on fru.dev sites, Reddit | +22 posts in the last 7 days, 5 the 7 days before (fading) | ||
| Permissions and IAM are confusing and error-proneDevOps and infra | 13d activesince Aug 26 | Seen on GitHub issues, Reddit, Hacker News, Community forums | +33 posts in the last 7 days, 1 the 7 days before (rising) | ||
| API rate limits and usage caps block real workLLM apps | 12d activesince Aug 28 | Seen on GitHub issues, Hacker News, Reddit, Community forums | +33 posts in the last 7 days, 3 the 7 days before (steady) | ||
| AI-generated code is buggy and hard to reviewAI coding | 11d activesince Aug 29 | Seen on Hacker News, Reddit, App Store reviews | +44 posts in the last 7 days, 4 the 7 days before (steady) | ||
| Claude Code upgrades break working setupsOutages and change | 10d activesince Aug 27 | Seen on GitHub issues | +22 posts in the last 7 days, 2 the 7 days before (steady) | ||
| LLMs give confidently wrong answersLLM apps | 9d activesince Sep 2 | Seen on Hacker News, Reddit, Bluesky | +44 posts in the last 7 days, 4 the 7 days before (steady) | ||
| Dependency conflicts break installs and buildsDev tooling | 9d activesince Aug 30 | Seen on GitHub issues, Stack Overflow | +33 posts in the last 7 days, 4 the 7 days before (steady) | ||
| GitHub upgrades break working setupsOutages and change | 8d activesince Aug 27 | Seen on GitHub issues, Hacker News | +22 posts in the last 7 days, 5 the 7 days before (fading) | ||
| Vendor lock-in makes switching tools costlyCloud and tool cost | 8d activesince Aug 27 | Seen on GitHub issues, Hacker News, Reddit | +33 posts in the last 7 days, 1 the 7 days before (rising) |
New today
All new| Problem | This week | ||||
|---|---|---|---|---|---|
| Trouble with Cursor agentAI coding | 2d activesince Sep 24 | Seen on Community forums | +33 posts in the last 7 days, 0 the 7 days before (new) | ||
| YouTube users keep asking for helpOther | 1d activesince Sep 25 | Seen on Reddit | +11 post in the last 7 days, 0 the 7 days before (new) | ||
| Lenovo users keep asking for helpOther | 1d activesince Sep 24 | Seen on Reddit | +11 post in the last 7 days, 0 the 7 days before (new) | ||
| Verizon users keep asking for helpOther | 1d activesince Sep 24 | Seen on Reddit | +11 post in the last 7 days, 0 the 7 days before (new) | ||
| SQL Server users keep filing bugsDatabases | 1d activesince Sep 24 | Seen on GitHub issues | +11 post in the last 7 days, 0 the 7 days before (new) |
Questions
What is a problem here?
A pain people state in public: a question nobody has answered well, a complaint, a workaround, a bug with many reactions. Posts that say the same pain about the same company or product are grouped into one problem with a short plain statement.
Where do the posts come from?
Public sources only: Hacker News (Algolia API), public Reddit feeds, Stack Overflow, the most-reacted GitHub issues of major data and AI repos, Discourse community forums (dbt, OpenAI, Hugging Face, PyTorch and more), Bluesky and Mastodon public posts, YouTube video titles, 1 and 2 star App Store reviews, and outages, deprecations, breaches and advisories from other fru.dev sites. Slack is read only where a community publishes public archives. No logins, nothing behind a sign-in.
How are posts grouped?
By rules, so every grouping can be explained: a post counts when it states a pain ("struggling with", "doesn't work", "workaround", "is there a way"), it is matched to a known issue (rate limits, schema drift, warehouse cost, flaky CI and so on) and to the companies and products it names. Posts about a product with no known issue are grouped by the words they share. The Method page lists every rule.
What do rising, persistent and new mean?
Rising: more posts in the last 7 days than in the 7 before. Persistent: raised on 5 or more days across at least two weeks, and still active this week. New: first posted in the last week. Problems backed by fewer than 3 posts from 2 people or places are marked unverified.
How can I use this?
As a map of unsolved pain: common, persisting problems are the signal to build something. Every problem links to the posts behind it, the companies involved and its trend over time. The data is also open as JSON, llms.txt and an OpenAPI spec for agents.