Comment on An investigation of the forces behind the age-verification bills

• misk@piefed.social ⁨1⁩ ⁨day⁩ ago

  Yes, because those sources weren’t read by a human. They could state one thing while the LLM hallucinates another.

  source

  • BrikoX@lemmy.zip ⁨1⁩ ⁨day⁩ ago

    Every factual claim in this repository cites a primary source (IRS filing, Senate disclosure, state database, legislative record, or published reporting) that can be independently verified. The tool does not change whether Meta’s LD-2 filing lists H.R. 3149, whether DCA has an EIN, or whether Stefanski admitted tech funding under oath. The records exist or they don’t.

    Unless you can point to incorrect primary source you are just wrong.

    source

    • misk@piefed.social ⁨1⁩ ⁨day⁩ ago

      LLM output isn’t „correct unless proven wrong”, nothing ever is. It’s that there was no human effort involved and the technology won’t be ever reliable enough to trust it with research work.

      source

      • BrikoX@lemmy.zip ⁨1⁩ ⁨day⁩ ago

        LLMs aren’t a reliable technology for research and and all it has done so far is that it has lowered the bar on presenting plausible sounding text.

        Don’t disagree. But again, have you checked those human written primary sources and have a particular dispute or are you just blindly saying everything is wrong without even checking the records for yourself?

        Also as the diclaimer says LLMs weren’t used for research only for grunt work. You seem to only care about complaining without even having any particular issues to complain about.

        ProPublica, OpenSecrets and US governments own fillings are human records which I checked and I can see they support the conclusion that was written.

        source