For the First Time, Artificial Intelligence is Being Used at a Nuclear Power Plant: California’s Diablo Canyon.

Comments

• AnonomousWolf@lemm.ee ⁨5⁩ ⁨days⁩ ago

  We really should distinguish between a Algorithm, a LLM and AI

  They are very different things, if you let a LLM run your nuclear reactor you’re in for a bad time.

  source
• who@feddit.org ⁨6⁩ ⁨days⁩ ago

  Diablo Canyon, California’s sole remaining nuclear power plant, has been left for dead on more than a few occasions over the last decade or so, and is currently slated to begin a lengthy decommissioning process in 2029.

  So this AI is apparently not operating a nuclear plant, which would be concerning.

  For now, the artificial intelligence tool named Neutron Enterprise is just meant to help workers at the plant navigate extensive technical reports and regulations — millions of pages of intricate documents from the Nuclear Regulatory Commission that go back decades — while they operate and maintain the facility.

  Ah, that makes more sense. I hope it doesn’t end up leading humans away from correct understanding of safety regulations.

  source
• IllNess@infosec.pub ⁨6⁩ ⁨days⁩ ago

  For now, the artificial intelligence tool named Neutron Enterprise is just meant to help workers at the plant navigate extensive technical reports and regulations — millions of pages of intricate documents from the Nuclear Regulatory Commission that go back decades — while they operate and maintain the facility.

  As long as the data sets is directly from the technical documents without out interference, it sounds fine to me.

  We probably spend about 15,000 hours a year searching through our multiple databases and records and procedures,” Zawalick said. “And that’s going to shrink that time way down.”

  I am very curious on how much this will actually help.

  I am also curious if reorganizing the database and training the staff would’ve been a better or worse time saver.

  source

  • technocrit@lemmy.dbzer0.com ⁨6⁩ ⁨days⁩ ago

    I am also curious if reorganizing the database and training the staff would’ve been a better or worse time saver.

    That’s literally what this is though. There is no actual “AI”. They’ve just got a new search algorithm.

    source

    • IllNess@infosec.pub ⁨6⁩ ⁨days⁩ ago

      Sounds like a waste of a quarter of a million dollars then.

      source
• technocrit@lemmy.dbzer0.com ⁨6⁩ ⁨days⁩ ago

  Spoiler: There is no “AI”.

  source
• Vanth@reddthat.com ⁨6⁩ ⁨days⁩ ago

  My company recently released a beta “AI” tool to do similar.

  What I’ve found but was 0% surprised by, our documentation has conflicts. It still takes an experienced, knowledgeable person to notice that there’s a conflict, then dive into source documentation to resolve them.

  What going to happen is they think they can cheap out on hiring/retaining employees with those abilities.

  I would bet a dollar the documentation in nuclear has similar conflicts and similar if not greater experienced human knowledge to resolve them.

  source

  • BussyCat@lemmy.world ⁨5⁩ ⁨days⁩ ago

    The closest conflicts I have seen is varying codes required and there is usually clear language to use the strictest code. For example one code might require piping with 1/8” thickness if it holds water at X pressure and another will requires 3/16” thickness if that water is going to have a chloride content above X amount so you if it meets both you use the 3/16”.

    source
• Goldholz@lemmy.blahaj.zone ⁨5⁩ ⁨days⁩ ago

  Kaboom?

  source
• RizzRustbolt@lemmy.world ⁨5⁩ ⁨days⁩ ago

  Has no one seen The Forbin Project?

  source
• J52@lemmy.nz ⁨6⁩ ⁨days⁩ ago

  Didn’t read, first thought on headline: ‘ah, just great, what could go wrong?’ (Sarcasm - for those in doubt)

  source
• bjoern_tantau@swg-empire.de ⁨6⁩ ⁨days⁩ ago

  Image

  source

  • sem@lemmy.blahaj.zone ⁨6⁩ ⁨days⁩ ago

    This is the same as just passing -y

    source