Comment on Would an LLM AI model trained wholly on consenting open source projects with a license requiring all derivative works be open source licensed still be problematic?

SnailMagnitude@mander.xyz ⁨13⁩ ⁨hours⁩ ago

• SnoopSqueak@lemmy.today ⁨13⁩ ⁨hours⁩ ago

  “I put millions of lines of code in a blender and kept re-rolling the output until I got something that kind of does what I want it to” will always be less efficient and more error-prone than a competent dev or team building with specific intent.

  People who can’t tell the difference approve bloated code with myriad vulnerabilities. Often, AI output doesn’t actually do what it pretends to do.

  If you’ve worked in any IT project, you will remember the most common issues: unclear requirements, scope creep, silos or misaligned incentives.

  For AI projects, you can add to the list: “We’re not even sure this thing works the same way every time” and you’ve got a perfect storm for failure.

  towardsdatascience.com/why-ai-projects-fail/

  An annoying trend I’ve noticed in articles like this is that they still insist the AI is not to blame. Phrases like “devs love the technology,” fuck no I don’t. Or “95% of AI projects fail, but yours will totally succeed if you do it right”. It feels exactly like how the lottery exploits people’s hopes.

  The steps required to make AI function are too costly to make AI worthwhile in almost every case. It was a stupid scam that too many companies fell for.

  source

  • SnailMagnitude@mander.xyz ⁨13⁩ ⁨hours⁩ ago

    Breaking News: High failure rates reported in early stage Oldowan startups make millennials reconsider tool use.

    This just seems like anti-computer evangelism from the 1950’s and 1960’s, and anti internet stuff in the 90’s, being repurposed.

    I’m working with anti-Manichean attacks repurposing old anti-Marcionite attacks at the moment, same stuff.

    Bronze tooling is more than fine to get stuff done.

    source

    • SnoopSqueak@lemmy.today ⁨7⁩ ⁨hours⁩ ago

      “AI” is more of a scam than a tool.

      We do not yet have artificial intelligence. At all. They only call it that to scam idiots.

      source

      • SnailMagnitude@mander.xyz ⁨7⁩ ⁨hours⁩ ago

        It’s been an academic disciple since the 1950’s and a work in progress since at least Ada Lovelace.

        Calling AI a scam for idiots is fair enough: but please include writing, machines and computers there too.

        World class moron Linus Torvals on the matter 14 July 2026:

        ---

        “AI is a tool, just like other tools we use. And it’s clearly a useful one.

        It may not have been that “clearly” even just a year ago, but it’s no longer in question today.”***

        source

        • -> View More Comments
  • abbadon420@sh.itjust.works ⁨12⁩ ⁨hours⁩ ago

    There is a class of devs that do love the technology. It’s the same class of devs that also loved blockchain and crypto. The class of devs that fail another startup every other year.

    source