I have the same answer for most problems in the Fediverse.

  • 0 Posts
  • 81 Comments
Joined 10 months ago
cake
Cake day: January 21st, 2026

help-circle



  • The engineer added that employees are now pressured to ship products as fast as possible, leaving them with little time to review Claude’s output or understand the code being created.

    Companies have always pressured software teams to ship products as fast as possible.

    What’s new is the part where all of the worst coders can shit out code a lot faster and pretend like they’re doing a good job. It used to take them a while to write garbage that they didn’t understand or review themselves, but now they can do it so quickly that it makes the actual good coders look like they’re taking a long time.

    This is why you absolutely have to have documented coding and review standards. The idea that you’d ship any code that has never been reviewed by a human is horrifying.

    The idea that you’d work 12 hour days is so stupid to start with. I’ve seen people consistently work that long and… surprise, they don’t actually work that long. They work about the same as anybody else and goof off the rest of the time. I doubt you can even just press enter for that long. People just don’t work that way.





  • I mean, you are describing modern copyright.

    But copyright is an example of something where the US Constitution, at least, explicitly says its purpose is to “promote” something:

    [the United States Congress shall have power] To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries.

    I like how straightforward the Copyright Clause is. It even gives a testable standard. The length of “exclusive Right” for IP should be the “limited” length of time that “promote[s] the Progress of Science and useful Arts.”

    So, the plain reading here is that IP laws need to give exclusivity long enough to encourage creation, by making authors and inventors think it is worth the time and money investment, but not so long that they discourage creation, for example, where a person doesn’t create something because they are afraid of violating IP laws.

    Originally, in the US, copyright lasted 14 years, with an option for renewing for 14 more years, and this is one example where I think they got it about right.


  • True. The Republican incumbent that he’s running against, Tim Burchett, promised to legislate against Flock cameras. While that’s nice, I no longer trust the promises of any Republican.

    As far as I’m concerned, we need to replace all Republican politicians. If their opponent is a Democrat, we need to be a little careful not to elect a wolf in sheep’s clothing. But if they’re independents like the guy we’re talking about, I think just about anybody would do. Like an animal or an inanimate object would even be infinitely preferable to a Republican.





  • I worked in software for, well, let’s say 20 years, at some elite places, too. And the one thing that I never understood was how few people ever bothered “to make sure it is correct,” as you put it, even before this new problem with AI. They didn’t even bother to make sure it was correct even when they wrote it themselves.

    You don’t know how many times I’ve been sent a request to review code, and it should have only been like 20 lines of code, but instead it was page after page of irrelevant stuff, because they created the review from the wrong baseline. But they never even took a single look at it before they sent it to me.

    So many people don’t write tests for their code that generally my biggest achievement at any given assignment was creating test frameworks that let them write tests using domain specific language, because it made it much easier. At a certain point, if it’s easy enough, even those people will write tests, especially if they know that I’ll be doing the review.

    I honestly don’t see AI code as being all that different. People still don’t even do the minimum, but they just get exposed and embarrassed more easily.




  • OpenAI’s general model doesn’t give out reliable medical advice. So, they need to instruct it not to give out medical advice, and instead to tell the user to consult a real doctor. Hopefully they have done that already.

    But beyond that, there’s a quagmire. Some people will move mountains to avoid talking to humans about anything, or even about some certain things like their health, and if ChatGPT refuses to give advice, they’ll either find a way around the instructions to the AI, or they’ll just find another AI that is willing to give medical advice. So part of me is glad that these people will get a specialized medical advice AI model, because it’s better than nothing, and if OpenAI wants to risk the lawsuits, that’s their problem. But it’s still dangerous if their new medical advice model hallucinates like all of the current models seem to. I personally wouldn’t recommend people use it.