Data and AI struggles

2,149 problems · 379 companies

LLM apps

Trouble with Anthropic usage

2 posts from 2 people2d activeSteady, +1 this weekunverified

Posts per day

Posts per day2 posts, Oct 25 to Nov 22

The posts behind it

GitHub issuesmlflow/mlflow · arthurgoussetNov 19

[FR] DSPy autolog: record token usage on LM spans, so trace cost is computed

### Summary With `mlflow.dspy.autolog()`, each LM span (`LM.__call__`, type `LLM`) has `mlflow.llm.model` and `mlflow.llm.provider`, but no `mlflow.chat.tokenUsage`. Token usage is only set on module spans, summed from `Prediction.get_lm_usage()`, and those spans have no model...

2 replies
App Store reviewsClaude App Store reviewsOct 25

Illegal Usage of Data

Anthropic is illegally handling your data.

Companies and products named

2 posts
1 post
1 post
About this problem

Evidence

2 posts from 2 people in 2 places, about 1 a week over 26 days. Mostly on Claude App Store reviews, mlflow/mlflow. Tools named alongside: Claude, MLflow.

Frustration Frustration 0 of 3· Seen on App Store reviews, GitHub issues

How it was grouped

Posts about Anthropic with no known issue, sharing the word "usage". First post Oct 25, 2026, latest Nov 19, 2026. Unverified: fewer than 3 posts, or one person or place only. Method

Other Anthropic problems

Anthropic users keep asking for help

Top post: "`HTMLHeaderTextSplitter` raises `ValueError` on init for the non-heading tags its splitter already supports", one of 37 posts from 35 people on GitHub issues, App Store reviews, Community forums.

Anthropic upgrades break working setups

30 posts from 24 people on GitHub issues.

Anthropic rate limits and usage caps block work

13 posts from 10 people on GitHub issues, Reddit, Community forums.

Trouble with Anthropic core

Top post: "_agenerate_with_cache merges streamed chunks on the event loop, and the merge is quadratic", one of 10 posts from 8 people on GitHub issues.

Anthropic fails to return valid structured output

6 posts from 6 people on GitHub issues.

Anthropic token costs are hard to predict

6 posts from 6 people on Hacker News, GitHub issues, Reddit.

Trouble with Anthropic cache

Top post: "LLM cache stores a generation that fails structured-output parsing and then replays it", one of 6 posts from 6 people on GitHub issues.

Trouble with Anthropic messages

Top post: "/v1/messages: client disconnect before the first chunk leaves the upstream LLM request running", one of 6 posts from 6 people on GitHub issues.

Anthropic runs out of context on real work

5 posts from 5 people on GitHub issues.

RAG on Anthropic returns irrelevant or missing context

5 posts from 5 people on GitHub issues, Reddit.

Trouble with Anthropic instead

Top post: "Tool returning a pydantic model yields a Python repr instead of JSON in ToolMessage.content", one of 5 posts from 5 people on GitHub issues.

Trouble with Anthropic call

Top post: "invalid_tool_call blocks still reach groq and mistralai during v1 conversion", one of 5 posts from 4 people on GitHub issues.

History

Momentum went from fading to steady.

Nov 22, 2026

First seen with 1 post; the top post now: "DSPy autolog: record token usage on LM spans, so trace cost is computed".

Nov 6, 2026