Anthropic is burning more and more dev goodwill(twitter.com)
38 pointsby tosh5 hours ago11 comments
  • infecto4 hours ago
    Is Claude code really refusing to do none coding tasks because thats not my experience even as of today and I would have expected those tokens are being accounted for the same way as a Claude.ai chat.

    Seems like all of this is gripes over blocking openclaw which I just cannot take seriously.

  • WaryByDesign4 hours ago
    I have truly no idea what this is all about. OAuth login issues aside (which have been a thing on-and-off for a while now, and I hope they fix soon), Claude's performance and responses have been entirely stable for me for like, forever.
    • politelemon3 hours ago
      Different people have completely different experiences as they perform different tasks, that shouldn't be difficult to understand. It's a bit like the purported degradation of quality of Google search. It is still excellent for me but I don't doubt that others are experiencing it.
      • halJordan3 hours ago
        If different people are allowed to have different experiences why are you only getting one set of experiences be controlling?
    • schmookeeg3 hours ago
      Same for me. I have been max200 for like 10 months, i think my usage is reasonable, and i never seem to get the quality throttling that gets complaints daily. I assume they are targeting certain piggies at the Claude buffet, who then squeal the loudest.
  • alsetmusic5 hours ago
    Who wants to follow a link to an untagged [video]? Please tag video submissions so that those of us who don’t care for them can save a click / tap.
  • amelius4 hours ago
    Can't they just throttle Claw usage if they are so worried that it may use up all their resources?
  • tim-tday3 hours ago
    Support your argument with specific evidence or GTFO
  • drakythe4 hours ago
    tl;dr: Dude thinks Anthropic is making claude worse on purpose to cut down on GPU usage and/or increase income. He recommends using Codex instead. This video should have been a text tweet.

    I'm 3 minutes in and he has done an ad read and shown that Anthropic wasn't kidding about removing the ability for Claude to be used with OpenClaw without paying for additional usage.

    And now he says he's going into his _actual_ problem with the service.

    EDIT: I have fast forwarded to 8 minutes in an he is still going on about Open Claw. Apparently they have banned mention of open claw in the system prompts. And it can trigger your extra charges even if you're not _using_ Open Claw, just mentioning it.

    EDIT 2: forwarded again. He went into something about his app and how he is concerned about whether his app will be allowed to continue doing this and that he would take Anthropic to court if that suddenly changed? But then he talks about how he likes using Claude for debugging misc issues outside of this app. Why did he bring the app up? And he is upset that Claude is telling him that it can't help him with random tech support questions like "Why is my dropbox icon not showing up in the MacOS menu?"

    We're now 16:45 in and this is his personal _conspiracy theory_ about Anthropic doing this stuff specifically so it doesn't work for Open Claw.

    This is a 24 minute rant that the tool he didn't understand before is now telling him he doesn't understand the tool and he should look elsewhere for specific answers.

    But Codex did what he wanted. So okay. "All without opening a browser" -- I... have so many responses that are unkind. This isn't Anthropic burning dev good will. This is Anthropic being a capitalist company.

    • taiwks10 minutes ago
      It's a Theo post, I see him and go oh I don't care / need his opinion on anything.
    • pixl974 hours ago
      >And it can trigger your extra charges even if you're not _using_ Open Claw, just mentioning it.

      Eh, I can only imagine this could get them in legal trouble in places that have actual consumer protections.

  • AlexErrant4 hours ago
    tldr:

    * Banning OpenClaw via system prompt filtering

    * Unclear TOS, citing Matt Pocock who sells a course on Claude (and therefore his interests are aligned with Anthropic):

    > I have never before experienced, from any developer tool, such a frustrating lack of clarity over the basic terms of usage. I personally asked, 3 weeks ago, and have received nothing but delays. The recent @bcherny announcement did absolutely nothing to clarify things.

    More details: https://xcancel.com/mattpocockuk/status/2040536403289764275

    * And finally refusing to help a user debug issues with Dropbox, repeatedly saying "I'm best suited for software engineering" and "this is a question for Dropbox support" (paraphrasing Claude).

  • ece4 hours ago
    In all honesty you'd have to be living under a fucking rock the last 10 years to think these APIs wouldn't be locked down to make more money for their owners. Who is surprised? Not me running models on my spark cluster.
  • nikolay4 hours ago
    [flagged]
    • nemomarx4 hours ago
      was it? I thought they had already dropped Anthropic by that point?
      • jsbisviewtiful4 hours ago
        I forget the timeline of events but the school bombing was within a very short timeframe of the announcement - a short enough timeframe that no rip and replace could have happened yet.
      • nikolay4 hours ago
        I guess you've not been keeping up with the news. It's been used for at least 1,000 strikes in Iran.
      • megous4 hours ago
        preparation for the initial strike has to have been going on for months
  • rdevilla4 hours ago
    [flagged]
    • on_the_train4 hours ago
      Because it's a clickbait title
      • rdevilla4 hours ago
        Understood. I will now aggressively flag so called "clickbait" on HN.