• rbn@sopuli.xyz
    link
    fedilink
    arrow-up
    40
    arrow-down
    2
    ·
    19 hours ago

    I mean that’s a fascinating piece of technology, but…

    1. it sounds like a desaster from an environmental perspective to shoot down every insect that’s approaching a house

    Operating a high-powered laser in a domestic environment introduces significant safety hazards, particularly regarding potential permanent eye damage from direct exposure or reflections off shiny household surfaces. To achieve the necessary operational effectiveness, the manufacturer selected a Class 4 laser source, which represents the highest hazard category and can cause severe injuries even during extremely brief exposures. To mitigate these risks, the device incorporates multiple protective mechanisms, including artificial intelligence cameras and radar capable of detecting humans and pets even in dark conditions

    I mean - what could go wrong?

  • Xerxos@lemmy.ml
    link
    fedilink
    arrow-up
    19
    ·
    19 hours ago

    Okay, so it can detect humans and pets. What about easily flammable materials or mirroring surfaces? This seems like it could be just one error away from causing a burning house or permanently blinding someone.

    • cynar@lemmy.world
      link
      fedilink
      English
      arrow-up
      5
      ·
      8 hours ago

      In this case, it’s likely not a bad word choice. The plans for this type of device use a mix of optical and acoustic tracking and identification. They are almost certainly based on modern machine learning and training. Basically, part of the modern “AI” family of methods.

    • davel@lemmy.ml
      link
      fedilink
      English
      arrow-up
      5
      ·
      19 hours ago

      Yeah… though many everyday products technically have had AI for a while now. Take this rice cooker, which came out in the 1990s.

      • TootSweet@lemmy.world
        link
        fedilink
        English
        arrow-up
        7
        ·
        edit-2
        17 hours ago

        many everyday products technically have had AI for a while now.

        Given that there really is no objective or universally-agreed-upon definition of “AI”, I’m not sure what this sentence could even mean. How can something “technically” qualify under a term that has no “technical” definition? Something can be “technically” increasing or “technically” underclocked or “technically” alcoholic. But something can’t be “technically in love”. I’d probably even ascribe more meaning to a claim that someone is “technically an indego child” than I would the claim that such-and-such technology is “technically AI”.

        An example I often give: I took a class toward my Computer Science degree in like 2005 or 2006 called “Introduction to Artificial Intelligence”. In that class, they taught me about the A* algorithm. When you do an escort mission in Skyrim and an NPC follows you, navigating around/through terrain features, trees, rocks, creatures, other NPCs, bodies of water, etc, that almost definitely uses the A* algorithm or some pretty slight variation on it.

        The A* algorithm doesn’t involve collecting/scraping training data or training anything. If you wonder why it’s behaving a certain way, a PHOSITA can easily trace through the logic and figure out why it produced the very specific behavior it did. (Unlike LLMs or Stable Diffusion or whatever.) A GPU wouldn’t make it more efficient/performant than a CPU would, and more efficiency is hardly needed because it’s super fast. (I chose Skyrim as an example, but to illustrate how performant it is, I’m sure it was used in NES games.) If I had to pick the one pair of most dissimilar algorithms I could, the A* algorithm and the one(s?) that drive(s?) LLMs would be a good pair to pick.

        And yet, they taught me about the A* algorithm in a class called “Introduction to Artificial Intelligence”.

        You called that rice cooker (your link to it was messed up, but this appears to be where you intended to link) “AI”. Why? Because it mentions “Fuzzy Logic”? Whether that qualifies as “AI” or not really depends who you ask. But if your threshold for what qualifies as “AI” is that low, then basically every technology is AI.

        Have you ever looked at the inside of a really old wall thermostat? They used coils of two-layered metal and a bulb with mercury in it. As the temperature changed, the metals would expand/contract at different rates, unfurling or… “furling”… the coil, moving the bulb. When the bulb moved one way, the mercury would roll/run over to that side and short a couple of contacts, acting as a switch closing the circuit which would then kick on the heat or AC. Does that qualify as “AI”? I don’t think it’s any less arguable than your rice cooker example.

        Which is what pisses me off about this article title. That “AI” at the front is clearly just to provoke attention. It doesn’t really mean anything. It’s just feeding into this insufferable hype cycle. And average people will read that and think it has something to do with LLMs. It (probably… didn’t read the article) doesn’t. But Sam Altman 100% wants you to think that his “advancements” (in quotes because it’s no advancement at all) in “AI” (whatever that might mean) have contributed to everything from humanoid robots to this bug zapper. He wants you to think that “AI” is some monolithic thing because if you believe that, the hype around LLMs sounds more realistic than if you don’t. And that’s how I know this article was written with the snivelingest bootlicker that ever licked a boot.

        Thank you for coming to my TED talk.

        • davel@lemmy.ml
          link
          fedilink
          English
          arrow-up
          3
          ·
          edit-2
          17 hours ago

          The meaning of AI changes depending on context, and it changes over time. When I first started studying AI, fuzzy logic was on the leading edge of it.

          Anything that was ever considered AI by the AI community technically is AI, because it historically came from that community.

          • queermunist she/her@lemmy.ml
            link
            fedilink
            arrow-up
            3
            ·
            15 hours ago

            It’s basically a buzz phrase on par with “high tech” or “cool car” - it has cultural meaning, but that’s kind of it.

    • XiJinpingStanAccount@lemmy.ml
      link
      fedilink
      arrow-up
      11
      arrow-down
      3
      ·
      18 hours ago

      Despite the constant fear mongering about Chinese products being used for nefarious means there’s no evidence of any of them even having the capability to do so. But Intel does have the ability to brick any Intel machine remotely via the Management Engine for example. I think if these had some sort of remote takeover and start zapping Americans in their back yard function the US government probably wouldn’t allow them to be imported anyway lol.

      • 0_o7@lemmy.dbzer0.com
        link
        fedilink
        English
        arrow-up
        6
        arrow-down
        1
        ·
        14 hours ago

        Um Chinese smartphone rootkit/spyware are absolutely a thing, and network devices or do you think they aren’t nefarious? And yes, other smartphone manufacturers do the same.

        Don’t act like no ones checking.