Login
You're viewing the front-end.social public feed.

Replies

  • PetraPetraOleum@cloudisland.nz
    Oct 10, 2026, 8:43 AM

    @iris_meredith You can make some level of analogy between LLMs producing shoddy work easily and various advances in the automation of textile production not (at least initially) matching the quality of a master weaver or spinner - except that a large amount of substandard fabric was and is incredibly valuable, but the product of LLMs seems largely to highlight the things we never really needed in the first place

    💬 0🔄 0⭐ 2
  • 💬 1🔄 0⭐ 1
  • 💬 0🔄 0⭐ 1
  • Elton Cora (Honorary)cora@hachyderm.io
    Oct 10, 2026, 9:52 AM

    @iris_meredith The notion of LLMs being most effectively employed as weapons of weak rivals, itself, is really interesting framing for examples and events unfolding in realtime.

    - The recent onslaught of ripoff/decomp clones of commercial games and major applications, which don’t really produce anything with competitive advantage and simply reveal a pervasive opportunism and dearth of original thought, comes to mind.

    - The fact that little that seems truly inspired has emerged, even as the tools continue to produce produce faster, in higher quantities, with increasing complexity, and how reliably quickly the sheen wears off of successive generations of models, once the details resolve and human expectations calibrate.

    The most excited among those using them seem to be those who are impressed with how much better the LLM can perform at a given task than the user prompting it, rather than how well the results compare in the real world, in others’ eyes, over time.

    And it is notable that these tools are ascendant in **the industry-friendliest, least-adversarial environment imaginable,** in which even those who command armies and economies are in their thrall.

    With all of the bad that comes with this, I do see a glimmer of light, least of which is that presuming we make it past whatever Jonbar hinge we’re living through, we’ll have a more refined understanding of just what it means to be a human creating.

    💬 0🔄 0⭐ 1
  • Dr. Bruce Simpsonbms@bsd.cafe
    Oct 10, 2026, 11:03 PM

    @iris_meredith Are the people who are saying that this skewed experiment is "inevitable" paying any attention to the financial collapse of what's driving it also looking, to put it bluntly, rather inevitable?

    There is no conceivable way the pre-IPO valuations reflect the reality of the revenue streams any frontier lab or hyperscaler are going to be achieved, vs the financial burden of the take-or-pay compute contract commitments when they crystallize.

    These are facts.

    Why do these people keep ignoring or glossing over these things? That's a rhetorical question, sure, but clearly the people advancing these narratives seem to think people are uninformed or stupid.

    💬 0🔄 0⭐ 1
  • 💬 1🔄 0⭐ 1
  • 💬 1🔄 0⭐ 1
  • 💬 1🔄 0⭐ 1
  • 💬 0🔄 0⭐ 0
  • Manniconfuseacat@mastodon.social
    Oct 11, 2026, 7:57 AM

    @iris_meredith Your posts always make me think. And I consider that a good thing.
    They also always give me the feeling that I'm not the last person on earth who thinks that words have (and should have) meaning.

    💬 0🔄 0⭐ 1
  • 💬 0🔄 0⭐ 1