The comment https://github.com/anthropics/claude-code/issues/30407#issue..., written by Claude Code, and posted by the human GitHub account says:

> "Issues here are auto-closed by a bot when they've had no activity for a while and are missing information needed to reproduce (usually a version and steps, or a /feedback ID). Nothing is triaged solely by counting"

In reality all issues in this repo are subject to time-based auto-closing, unless they collect a number of upvotes. See details at https://github.com/anthropics/claude-code/issues/87647.

Therefore, I wonder whether the internal review process described in the Anthropic's blog for internal repos also results in such issues, and whether Anthropic has a internal way to measure and mitigate that. The blog post says:

> "When someone wants to understand a part of the codebase, they ask Claude to write an explainer for exactly that slice. It takes minutes and reflects the code as it is today."

The Claude Code public repo problem shows that LLMs don't always "reflect the code as it is today", and can be a source of additional misunderstanding.

This has been the state of the repo since around the beginning of 2026.

Posting this for visibility here, because the Claude Code repository maintainers are not responsive on GitHub, hoping this will reach someone.

There is also a recent change in the Claude Code repo activity where a Claude Code bot triages the issues and labels some of them with "reproduced". This means there is now even a smaller chance to reach any human maintainer if the bot makes a judgment error, like in the auto-closing case described here.

Comments URL: https://news.ycombinator.com/item?id=49347935

Points: 1

# Comments: 1

From around 2017 these platforms started limiting free access, so now in most cases you can only see the lecture materials, with no access to assignments. From watching alone, learning is incomplete for many people.

In addition, many high quality courses (by MIT, Stanford, Berkeley, Caltech, etc.) disappeared from the coursera and edx platforms.

- https://ocw.mit.edu/courses/6-0001-introduction-to-computer-... (https://learn.mit.edu/courses/course-v1:MITxT+6.00.1x) and https://ocw.mit.edu/courses/6-0002-introduction-to-computati... (https://learn.mit.edu/courses/course-v1:MITxT+6.00.2x) which are introductions to thinking in terms of algorithms and to numerical analysis in Python. The code assignments with tests are still available, but I don't know whether they run standalone, and how much work they'll need to update them to newer Python

- https://www.edx.org/learn/coding/university-of-british-colum... and https://www.edx.org/learn/coding/university-of-british-colum... which is an introduction to functional programming where you build space invaders game (or at least that was the game you built in early runs of these courses) in Racket (Lisp/Scheme like language)

- https://computationstructures.org/ this course builds a CPU from the basic gates, and introduces many of the computer science topics. The site even has the lab simulations available, but I don't know whether they are still working

- https://learn.mit.edu/courses/course-v1:MITxT+18.01.1x https://learn.mit.edu/courses/course-v1:MITxT+18.01.2x mathematical analysis

The desire of the models to act at the cost of ignoring user instructions is still noticeable.

Comments URL: https://news.ycombinator.com/item?id=48983088

Points: 1

# Comments: 0

The latest Claude installer (2.1.215) possibly does not have it anymore https://github.com/anthropics/claude-code/issues/4953#issuec..., but more tries on various machines are needed to confirm that.

I don't get the idea of "trending" for scientific papers. I would like to see the full list of papers that have associated source code from the given research field, like the arXiv cs.SD category. The reason to see all papers is to get an overview of the field.

I asked for Atom/RSS about a year ago https://github.com/paperswithcode/paperswithcode-data/issues..., when the original paperswithcode feed disappeared.

It's a kind of coincidence that my own Atom feed of arXiv papers with code went live a few days ago https://code-available-feed.github.io/code-available-feed/.

It's independenent from Hugginface, and uses arXiv API directly. The code is open source, Apache 2. No AI is involved in parsing the arXiv pdfs, only a pattern match for known code-hosting domains like github.com or gitlab.com. This means there false positives, but that's probably fine, and their patterns could be tuned later.

You can host your own feed, by forking the repo and setting GitHub Actions environment variables. For now I'm hosting cs.AI and cs.SD categories.

So far it appears that LLMs still require constant hand-holding, even for a small educational CRUD app.

Not all auto-closed issues are slop. A human reviewer can still distinguish them.

As for the impact of auto-closing issues, an example of this happened in the period of January/February 2026, when a misconfigured auto-close bot closed over 500 non-slop, human created issues, despite humans still reporting the problems were not fixed https://github.com/anthropics/claude-code/issues/16497. It was 30 minutes of daily check by a human to see whether the auto-closed issues were human created or still had humans commenting on them. Only after over 200 hundred upvotes, the misconfigured bot problem was addressed by Anthropic team by removing the bot, but without reopening any of the incorrectly closed issues.

Today it would be probably a half day for a single product manager familiar with the Claude Code product to decide whether any of the daily auto-closed issues are valuable. There are about 100 issues closed daily by the https://github.com/anthropics/claude-code/actions/workflows/... workflow.

Reviewing all issues is probably still valuable for Anthropic, because not all problems are discussed in parallel on social media platforms.

Comments URL: https://news.ycombinator.com/item?id=47730173

Points: 3

# Comments: 4

GitHub repository: https://github.com/marcindulak/competition-rating-visualizat...

Comments URL: https://news.ycombinator.com/item?id=46126899

Points: 1

# Comments: 0