I find both to be true. I use Claude for most of the implementation, and Codex always catches mistakes. Always. But both of them benefit from being asked if they’re sure they did everything.

To be clear, for those reading these comments and thinking “oh no Azure”, this is an addition to the list of cloud companies that provide “cloud infrastructure worldwide” for “all products”. Alongside GCP and AWS. This is not a GitHub style announcement that they’ve moved all operations to Azure.

It's also down for me here in Brazil. Getting overloaded errors for about one hour now. It's been happening a lot this week. Is this normal for Anthropic?

Working fine for me right now, from Brazil. Claude via Github Copilot at least.

I'm using Claude Code on the terminal. Not sure if it matters.

  API Error: 529
  {
    "type": "error",
    "error": {
      "type": "overloaded_error",
      "message": "Overloaded. https://docs.claude.com/en/api/errors"
    },
    "request_id": "..."
  }

The promotional double usage period is just about to end too. Sucks.

It's down for me too. A colleague says it's up though - it's possible they're shedding different groups of users (he has the Max subscription, I don't).

They absolutely do. In this case litellm 1.82.8 had been out for at least a week (can’t recall the exact date offhand). The compromised version was a replacement.

It actually wasn't. That was one of the reasons why I looked into what was changed. Even 1.82.6 is only at an RC release on github since just before the incident.

So the fact that 1.82.7 and then 1.82.8 were released within an hour of each other was highly suspicious.

Ah, my mistake! Thanks for the correction.

But I believe you can replace versions on both, nonetheless. It’s a multi step process, unpublish then publish again. But the net effect is the same.

PyPI enforces immutable releases.

https://pypi.org/help/#file-name-reuse

> PyPI does not allow for a filename to be reused, even once a project has been deleted and recreated...

> This ensures that a given distribution for a given release for a given project will always resolve to the same file, and cannot be surreptitiously changed one day by the projects maintainer or a malicious party (it can only be removed).

If you lock your dependencies, it should fail if the hash doesn't match.

1.82.7 and 1.82.8 were only up for about 3 hours before they were quarantined on PyPI.

This is fantastic, thank you. Your reporting has been great. But also, damn, the playlist.

I get the email notifications from Anthropic’s status monitor, and I think they might be my most frequent emailer these days.

I think you’re confusing capital c Claude Code, the desktop Electron app, and lowercase c `claude`, the command line tool with an interactive TUI. They’re both TypeScript under the hood, but the latter is React + Ink rendered into the terminal.

The redraw glitches you’re referring to are actually signs of what I consider to be a pretty major feature, a reason to use `claude` instead of `codex` or `opencode`: `claude` doesn’t use the alternate screen, whereas the other two do. Meaning that it uses the standard screen buffer, meaning that your chat history is in the terminal (or multiplexer) scrollback. I much prefer that, and I totally get why they’ve put so much effort into getting it to work well.

In that context handling SIGWINCH has some issues and trickiness. Well worth the tradeoff, imo.

Codex is using its app server protocol to build a nice client/server separation that I enjoy on top of the predictable Rust performance.

You can run a codex instance on machine A and connect the TUI to it from machine B. The same open source core and protocol is shared between the Codex app, VS Code and Xcode.

OpenCode works this way too

not sure if same reason but window resize feels better in claude than codex.

on my m1, claude is noticeably slower when starting, but it feels ok after that.

I had a nasty slow claude code startup time at one point something like 8s, a clean install sorts it all out. Back up your mcp config and skills and you're good.

That sounds like poor engineering on Anthropic's part. Good software doesn't slow down over time like that.

That's the same reason I don't like Opencode, but Codex doesn't use the alternate screen. I remember it did when it was very very new, but now it doesn't.

Ah nice, good to know. I hadn’t used codex in a while. I actually really like opencode and its ui, just wish it didn’t clear the screen on exit. It could at least redraw whatever was last in the chat, that would be better than nothing.

well it still does something weird which breaks scrollback in zellij, it's a known issue, but who knows when it's getting fixed

I’ve looked at that a bit. Roff and mandoc etc have specialized tagging that’s not easily representable in markdown. You’d wind up with a lot of boilerplate or special non-standard markup, which would undermine the point.

The LLMs are super good at doing that translation, though. They can write those formats no problem.

I can’t tell if you’re joking or not, but it’s funny either way.

On one hand this is a neat idea. I've thought about how nice it would be to have a visual layout tool for text-based designs. The current offerings are slim. Of course, you could easily argue that if you need a visual tool for it, you've gone too far; even the most sophisticated TUIs are still extremely simple.

On the other hand, for this work as they describe, it needs to be a complete UI framework across a bunch of languages and built on top of a bunch of existing frameworks. That seems... ambitious. Building one UI framework for one language is plenty hard enough.

The opening sentence “Those who have never endured the relentless ringing of tinnitus can only dream of the torment” does not mean what they think it means. Unless this is a very niche kink.

Thats what I thought when I first read it. I don't this is something that people look forward to in dreams. Or possible to imagine in dreams.

There are limits to dreams.