4 ms·

Being bored, scared, and on Twitter is a bad combo. I'm mostly down on the current AI, as it's ML leveraging the vast "intelligent" conversations we put out on

by mimd 4y ago

Being bored, scared, and on Twitter is a bad combo. I'm mostly down on the current AI, as it's ML leveraging the vast "intelligent" conversations we put out on the internet. It's wheel of fortune and it failed my tests on reasoning miserably, even with the answers in hand. (I'm not going to tell you which tests, because then OpenAI will scrap it and tune their model to claim victory in GPT5. I saw the feathers bricks cheating!). That doesn't mean that these technologies could not do damage, the greatest risk bring hubris in it's capabilities. I wouldn't rely on it's reasoning for launching an ICBM for example. But I suspect that it's contributions/detractions to the workforce will be more muted than thought, as the current use cases seem marginal in cost savings. Ie, HR letters from management time savings using templates versus a AI generated one. But I don't see how we get from a wheel of fortune completer to, in the usual example, an AI whose goal is to eliminate trash by eliminating humans. It's missing components and training necessary to learn those "functions", and their's little incentive to do so. I'd rather have the benefits of understanding protein folding for lifesaving drugs, decreased costs on physics simulations for fusion reactors, and so on, than stagnation due to someone's bad scifi.

smrtinsert 4y ago
When exactness, liability and oversight matters, it's just not ready yet, period. Even when its ready for that, it will take years for legislation to make it possible to use in key areas such as finance or medicine or law. While some of the public is scared, others are thrilled as they expect an AGI takeover somehow results in a egalitarian UBI cushioned existence. It's a mess.
- mimd 4y ago
  It is. The AGI takeover crew is a bit nauseating. But I sorta understand them considering our daily panics. Sorta a dream of being whisked away by a prince/princess. But the scared ones are much more organized, and have been at it for a good while now. And they're more of a threat, as they could shut this down for no gain, while the AGI takeover ain't happening with this wave of AI. Still, I worry about a period in which doctors/lawyers forget their liabilities and decide to consult the great chatgpt oracle. Makes for a wonderful daily fail story. ;)
  - low_tech_love 4y ago
    They won’t shut anything down. Maybe some authoritarian leader in some insignificant country does that, and all the weirdos will hail him as the future of humanity (like the El Salvador crypto fiasco), but the mind power and money behind current AI is unstoppable. Whether that is a good thing, I don’t know, but we’ll see.
    - mimd 4y ago
      "Shut it down" is too strong a quip, but it's more to illustrate that their goals will do something while AGI take over crew is dreaming. I'm more worried about losses of incentives that fund development or the use of it as gatekeeping that will limit entry. Ie, Europe's plans or getty's lawsuit. Yes, there is a massive amounts of money, incentive, and mindpower to ward them off. But they're also just mostly letting the doomers run free without pushback. I suppose it could work, if it's to ignore the doomers as they'll make themselves look foolish. But theirs a lot of incentives to attack ai research, from "it will take your jobs", to doing it out of pure spite to hurt tech. Eg. payments to newspaper's is an example of fairly recent line that had money behind it and partially succeeded. It could get stupid too, such as Australia saying linking needed payment. Most of the main barrier is fear from US/China about letting the other get the upper hand.

by mimd 4y ago

smrtinsert 4y ago

When exactness, liability and oversight matters, it's just not ready yet, period. Even when its ready for that, it will take years for legislation to make it possible to use in key areas such as finance or medicine or law. While some of the public is scared, others are thrilled as they expect an AGI takeover somehow results in a egalitarian UBI cushioned existence. It's a mess.

mimd 4y ago
It is. The AGI takeover crew is a bit nauseating. But I sorta understand them considering our daily panics. Sorta a dream of being whisked away by a prince/princess. But the scared ones are much more organized, and have been at it for a good while now. And they're more of a threat, as they could shut this down for no gain, while the AGI takeover ain't happening with this wave of AI. Still, I worry about a period in which doctors/lawyers forget their liabilities and decide to consult the great chatgpt oracle. Makes for a wonderful daily fail story. ;)
- low_tech_love 4y ago
  They won’t shut anything down. Maybe some authoritarian leader in some insignificant country does that, and all the weirdos will hail him as the future of humanity (like the El Salvador crypto fiasco), but the mind power and money behind current AI is unstoppable. Whether that is a good thing, I don’t know, but we’ll see.
  - mimd 4y ago
    "Shut it down" is too strong a quip, but it's more to illustrate that their goals will do something while AGI take over crew is dreaming. I'm more worried about losses of incentives that fund development or the use of it as gatekeeping that will limit entry. Ie, Europe's plans or getty's lawsuit. Yes, there is a massive amounts of money, incentive, and mindpower to ward them off. But they're also just mostly letting the doomers run free without pushback. I suppose it could work, if it's to ignore the doomers as they'll make themselves look foolish. But theirs a lot of incentives to attack ai research, from "it will take your jobs", to doing it out of pure spite to hurt tech. Eg. payments to newspaper's is an example of fairly recent line that had money behind it and partially succeeded. It could get stupid too, such as Australia saying linking needed payment. Most of the main barrier is fear from US/China about letting the other get the upper hand.