This is the fediverse, so this is probably going to be extremely unpopular, but yea here goes.

Not saying you have to love it, or trust it blindly. Skepticism is healthy. But there’s a difference between “I use it carefully and check everything” and “I refuse to touch it.”

Historically, refusing new tools has never stopped them. Coopers didn’t stop the steel drum by refsuing to make them. They just ended up out of work while the industry moved on.

The realistic threat isn’t AI replacing you. It’s a colleague who uses AI well doing your job faster. And the skill of using it properly (knowing when it’s wrong, veriyfing output, knowing what not to use it for) takes time to build. Starting late means competing with people who’ve had years of practice.

You can have legit concenrs about ethics, energy, or jobs and still learn the tool. Opting out doesn’t fix those problems. It just removes you from the people who get a say in how it’s used.

PS. Ethically trained models exist and you can run models locally.

you are viewing a single comment's thread
view the rest of the comments
[–] 26 points 11 hours ago (4 children)

It’s a colleague who uses AI well doing your job faster

My professional experience so far has been that "faster" doesn't actually work unless you abandon the whole concept of responsibility entirely and just blindly trust the AI. Writing code has never been the long tail of software engineering. It has always been code review, software management and debugging, and so far AI has actually made that significantly slower in my experience.

Just this week my colleague and I pointed out a very laughably obvious flaw in a "true believer's" code, and he just... argued about it, not because the AI can't be wrong, but because of the implication of him submitting code he clearly did not understand. So far, that's the norm, not the exception.

Obviously it's not going anywhere, but the pendulum is so far past the equilibrium it's going to settle into eventually, that the loudest evangelists might as well just be wrong.

  • source
  • hideshow 4 child comments
  • [+] [S] -7 points 9 hours ago (3 children)

    This all or nothing extreems aren't reality.

    You can have AI help you add unit-tests you wouldn't have thought of.

    Once you are "Done" writing tests, and can't think of more scenarios, AI more often than not can. And can help you add better test coverage.

    I've heard many people say this, what is the downside here?

  • source
  • parent
  • hideshow 3 child comments
  • [–] 11 points 9 hours ago (2 children)

    A unit test you don't understand is worse than useless.

  • source
  • parent
  • hideshow 2 child comments
  • [–] [S] -1 points 9 hours ago* (1 child)

    Then read the extra tests it came up with and make sure you understand it.

    It catches things like a null check you might have forgotten, trivially easy to understand.

    You don't even have to have it write any tests if you want, you can ask it "what Unit-Test scenarios did I miss?" And write them yourself.

  • source
  • parent
  • hideshow 1 child comment
  • [–] 3 points 8 hours ago*

    That last scenario is the correct usage IMO, and largely how I've been using it. However, in my experience the actual result in actual organizations is that every software engineering taboo just falls to some combination of ego, laziness, and whip cracking, one by one, until something gets properly fucked, and only then are strict rules established.

    My major beef is that it is too tempting to abuse it, and the abuse inevitably becomes too voluminous to properly inspect, and something inevitably breaks. The same thing happens with humans, but humans moving at human speed make it exponentially easier for humans to comprehend. For bullshit slopware, code nobody actually has time to understand is honestly fine, but for serious systems it is absolutely not ok.

  • source
  • parent