Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Examples: Monday, today, last week, Mar 26, 3/26/04
Welcome home! Please contact lincoln@icrontic.com if you have any difficulty logging in or using the site. New registrations must be manually approved which may take several days. Can't log in? Try clearing your browser's cookies.

The AI tsunami

1235»

Comments

  • personperson Don't believe everything you think The liminal space Veteran

    @Jeroen said:
    As the article I quoted said, a 10% error rate may be acceptable when looking for a peanut sauce recipe, but imagine your CFO getting the wrong number one time in ten when writing up the financial year report.

    Right, as impressive as much of what it can has been, it really isn't to the point of being able to replace a lot of things... yet.

    I've mentioned other times in the thread that for a time the best chess "players" were a combination of humans and computers, then at some point computers dominated them too.

  • JeroenJeroen Not all those who wander are lost Netherlands Veteran

    I think it may take 10 more years…

  • personperson Don't believe everything you think The liminal space Veteran

    @Jeroen said:
    I think it may take 10 more years…

    My opinion is about the same. I'd put my money on starting to see major changes in 3-5 years and I think there will be really world changing things coming in 5-10. Its hard to put a real time line on it as what do I really know and our systems and institutions are all so complex. But I think we're on the beach watching the water go out now and not really understanding what is coming.

  • marcitkomarcitko Veteran

    I watched a series of his videos yesterday. A very noble initiative and wholesome channel. Thanks for sharing.
    There was an update: someone from YT did get in contact with him and restored the comments.

    @Kotishka said:

    I have no words to describe how this is just madness. How can we just dump so much responsibility into AI? Like...seriously...this rush...

    Poor man. A quick summary now:

    SBSK has been interviewing people with different kinds of ailments and disabilities to share with the world how to make connections with them, learn about those involved in a deeper way beyond their diagnosis, and also provide a platform for people to engage, share comments of support and even tips/advice via the YT comment section. Well, now those comments are gone and without even an explanation. The burnout behind this man is tremendous. Please share your support if you can!

    Kotishka
  • JeroenJeroen Not all those who wander are lost Netherlands Veteran

    It seems Elon Musk’s SpaceX IPO has not gone the way he wanted it to… after peaking at 225 dollars per share, the price is now down to 108 dollars, and it may have further to fall.

    That means that xAI, Musks AI company, has near-zero value in the package that is a SpaceX share. Which may have something to do with the fact that federal employees, despite having a license te use Grok (xAI’s flagship LLM Ai), are turning to ChatGPT and Claude instead.

    marcitkoperson
  • personperson Don't believe everything you think The liminal space Veteran

    @person said:

    @Jeroen said:
    I think it may take 10 more years…

    My opinion is about the same. I'd put my money on starting to see major changes in 3-5 years and I think there will be really world changing things coming in 5-10. Its hard to put a real time line on it as what do I really know and our systems and institutions are all so complex. But I think we're on the beach watching the water go out now and not really understanding what is coming.

    I'm thinking that the tsunami metaphor is a bit flawed. I think the idea of AI overwhelming the world as we know it holds. But unlike a tsunami a portion of that change will be constructive and positive rather than simply destructive.

    What might make for a better metaphor?

  • JeroenJeroen Not all those who wander are lost Netherlands Veteran

    Perhaps a paradigm shift? It occurs to me that work before AI and work after AI will have decidedly different characteristics, which may well bear similarity to a paradigm shift…

  • lobsterlobster lobster Londinium Veteran

    Marketing hype is not a product. Who guessed?

    People used to be computers (one of the tasks at early NASA)
    Now hyped CEO's are making imaginary wealth (economic legal theft) out of a non working but as ever developing robot AI 'revolution'.

    https://youtube.com/shorts/jvQdFe8FPCw

  • personperson Don't believe everything you think The liminal space Veteran

    @lobster said:
    Marketing hype is not a product. Who guessed?

    People used to be computers (one of the tasks at early NASA)
    Now hyped CEO's are making imaginary wealth (economic legal theft) out of a non working but as ever developing robot AI 'revolution'.

    https://youtube.com/shorts/jvQdFe8FPCw

    I think of it like the dotcom bubble. There was a lot of hype and people threw lots of money at anything to do with the internet. Lots of people got rich overnight. Then the hype popped and lots of people lost it all overnight. But it wasn't because the internet was fake or that it went away.

    (Also, wealth doesn't just come from taking it from one pile and moving to another. Its possible to make something new a valuable that creates wealth)

  • personperson Don't believe everything you think The liminal space Veteran

    Something some teachers have been doing is to include words in white text into their essay questions. In this report the teacher put in the word Madagascar, in another I heard about the prompt "respond as if you're a pirate" was used. The students can't see the words but if its copy/pasted into AI it will read it.

    The more meaningful thing is in the story here 32 of 35 students used AI and didn't check the work enough before handing it in. Human cognition is on its way out. I don't know what that will mean, to me it feels worrying, but Socrates worried that writing would ruin people's memory and it did, but we gained more for it. Will the loss of human cognition go the same route?

  • JeroenJeroen Not all those who wander are lost Netherlands Veteran

    One other thing, chat bots are making human knowledge more available than ever before. For the research-minded it was already so, but now it’s all there in conversational chat format. I think we will see a certain knowledge fatigue set in sooner or later…

  • personperson Don't believe everything you think The liminal space Veteran

    @Jeroen said:
    One other thing, chat bots are making human knowledge more available than ever before. For the research-minded it was already so, but now it’s all there in conversational chat format. I think we will see a certain knowledge fatigue set in sooner or later…

    Can you expand on knowledge fatigue? I'm not exactly sure what you mean. I've heard about (mostly) kids who use AI for most of their thinking saying how they don't like doing it anymore, they just want to socialize or have fun, etc.

  • JeroenJeroen Not all those who wander are lost Netherlands Veteran

    Knowledge fatigue as I have encountered it is getting tired of constantly having to learn new things. Conversations with AI chat bots often feature having to learn new concepts, this is but one aspect of this.

  • personperson Don't believe everything you think The liminal space Veteran
    edited August 7

    @Jeroen said:
    Knowledge fatigue as I have encountered it is getting tired of constantly having to learn new things. Conversations with AI chat bots often feature having to learn new concepts, this is but one aspect of this.

    Yeah, I think this is along the same lines as my worry around the concern of Socrates that people would lose their ability to memorize things at the advent of writing. No need to learn when AI already knows all.

    For myself I'm curious and aim to be a lifelong learner. So I tend to appreciate the way AI can help expand my knowledge on subjects I'm interested in, in an easily digestible way. But, yes, at times it gets to be overwhelming and I find myself sort of shutting down and just wanting ease and comfort over the effort and challenge.

  • JeroenJeroen Not all those who wander are lost Netherlands Veteran

    Yes it is. It’s about the concern that people will just leave the hard tasks of thinking up to the AI’s. You already see this in the research on young adults and why they choose to use AI for schoolwork, and mostly it’s because they don’t feel like doing the hard work. I think very soon we will be so used to AI that we won’t be able to do without it.

    I do think curiosity is a human trait, and everyone will query AIs to find things out, in much the same way that we periodically dip into Wikipedia. But that is different from a directed course of study, which teaches us a whole field of knowledge in breadth and depth. You typically only take a course if you have to for your work…

    person
  • personperson Don't believe everything you think The liminal space Veteran

    I have small project for work to prepare for. Its similar to things I'm comfortable with but with conditions different enough I wasn't completely comfortable with my knowledge so I thought I'd ask AI for some input. It did give me some info that helped, but it was far enough off that I spent 4 hours and $20 in gas a supplies running around doing things that I had no need to do. I was already correct in my plan, I just didn't have the detailed knowledge of the specifics enough to know I was correct. It also gave me some horrible advice that had I followed would have led to a disaster of a project.

    What is it telling me about things that I don't have the knowledge to properly question?

    Jeroenmarcitko
  • JeroenJeroen Not all those who wander are lost Netherlands Veteran

    Thought you might find this interesting. It is a more detailed examination of the claims the AI industry is making that AI might lead to human extinction.

    person
  • JeroenJeroen Not all those who wander are lost Netherlands Veteran
    edited September 24

    Also this on how Meta’s new (and very popular) AI agent Muse seems to do things without permissions…

    https://www.inc.com/jason-aten/metas-new-muse-ai-agent-read-my-private-messages-i-never-asked-it-to/91408202

  • personperson Don't believe everything you think The liminal space Veteran

    I'm trying to sort out where I am on this now. I feel like there is a much larger general awareness now as opposed to when I started the thread.

    I'm still concerned about the perils, but I also wonder to what extent the companies themselves are hyping up the risks for their own ends. I don't think they're fully cynical, I believe them that they have concerns.

    My sense is there is some movement for greater safety and oversight. But I doubt it can come fast enough. There are real world constraints on just how powerful it can become (data centers, chips, power), that at least has the potential to slow things down or put some sort of cap on things.

    Honestly, I still think a large AI catastrophe that can be walked back is perhaps our best bet to find the impetus for safety coordination.

  • lobsterlobster lobster Londinium Veteran

    Is it wakey wakey time yet?

  • personperson Don't believe everything you think The liminal space Veteran

    @lobster said:
    Is it wakey wakey time yet?

    I'm not familiar with the arguments in the trailer, so I don't know if they'll wake me up to the truth or if I'll remain ignorant.

    What do they mean to you? How do you understand the documentary?

  • JeroenJeroen Not all those who wander are lost Netherlands Veteran

    I dislike trailers with lots of rapid cuts, it makes it hard to think or remember through it. The last sentence, if the machines can do everything better than you, does your life have meaning? Stood out. There were also some references to tech-bro’s, and the absurd amounts of water and power being used by data centers.

  • personperson Don't believe everything you think The liminal space Veteran

    Kurzgesagt had maybe their longest video I've seen on AI agents and the Hugging Face attack. It seems like the start of many of the bad things I've been worried about, even though I've been reassured that this isn't how AI works and/or its all just marketing hype so these companies can take all our money.

  • JeroenJeroen Not all those who wander are lost Netherlands Veteran

    Yes, I heard about Hugging Face, and that they were basically bought out by NVidia to dissuade them from a lawsuit against the makers of the AI agents involved, which would have forced a legal opinion on whether companies creating AI agents are liable for the harm these agents might do.

    person
  • personperson Don't believe everything you think The liminal space Veteran
    edited October 26

    More developments.

    It doesn't matter whether AI is conscious or works like a human mind for it to operate in the world in ways that are dangerous to people, independent of what people intend for it to do.

  • lobsterlobster lobster Londinium Veteran

    Oh er I can not turn off my computer. Really? Still shamelessly drinking the publicity?

    Still not ready to wake up? Oh well...
    https://joinmastodon.org/

  • personperson Don't believe everything you think The liminal space Veteran
    edited October 27

    @lobster said:
    Oh er I can not turn off my computer. Really? Still shamelessly drinking the publicity?

    Still not ready to wake up? Oh well...
    https://joinmastodon.org/

    Its not about the current marketing hype, or the current LLMs that the public uses. People have been thinking about the problem of "entities" that are smarter than us for a long time. Don't let your distrust of large corporations blind you to the risks. It may work out alright, but to say there's no safety issue to worry about here is naive.

    Specifically on the "unplug it" solution. One, they've already shown some ability to deceive and work around safeguards. Two, once its integrated into our world, even if we could shut it off, what would that do? It would be like shutting off the internet, sure we could do it, but at what cost?

    If you have some sort of knowledge that says that the rest of the world is wrong please try to explain it. The dismissive comments, LOL reactions, and arguments from authority are totally unpersuasive to me and seem more an attempt to change minds through intimidation than persuasion.

    lobster
  • lobsterlobster lobster Londinium Veteran

    Nobody is trying to persuade or dissuade and nobody owes you an explanation. I find you funny. That will continue for as long as I do not choose your mainstream publicity and news sources. Some being generated for the enrichment of well who?

    You decide.

    Are you really intimidated? LOL! :mrgreen:

  • personperson Don't believe everything you think The liminal space Veteran
    edited October 28

    @lobster said:
    Nobody is trying to persuade or dissuade and nobody owes you an explanation.

    Its true you're not compelled to offer any, but as you also offer nothing as proof, it can be dismissed as irrelevant.

    I find you funny. That will continue for as long as I do not choose your mainstream publicity and news sources. Some being generated for the enrichment of well who? You decide.

    Maybe this is a cultural thing? In my part of the world laughing at someone rather than with them is considered rude, condescending and disrespectful. Its bullying behavior.

    Are you really intimidated? LOL! :mrgreen:

    I suppose not particularly. I'm just saying the tactic is more intimidation than persuasion. Its trying to make me feel bad or small, rather than offer a different path.

  • Humans are funny.

    Over at the "AGF" (Acoustic Guitar Forum) there's an AI thread in the Open Mic section, now up to 14 pages. Reading about musicians discussing AI is amusing, in a way not unlike reading about Buddhists discussing AI.
    Since I have no computer proficiency, nor the right kind of brain wiring to ever acquire deep proficiency, I find it interesting to read what the experts (or apparent experts) say. And I note that they disagree with each other, sometimes disagreeably so.

    So, from our native New Buddhist experts who are in the "Doomsday" camp, can you describe what we will see if doomsday is real? What will be the early signs? Will there be a few cases coming out of Wuhan or Ghana that will be the harbingers before it spreads to and kills the rest of the population? What will those cases actually look like? Will I die sitting at my computer? My bank account vaporizes? Commandos rappel out of a helicopter onto my roof? Or Netanyahu's roof? Or will we all die together instantaneously without prodromal crescendo, like when Y2K turned all technology to stone at the stroke of midnight?

    From the NB experts in the "You're Probably Going To Live" camp, what is the evidence that the guardrails exist, and are real? Will there be cases where agents put their finger in Squeaky Fromme's gun? Are those agents everywhere and always on guard? Or is there evidence that her gun is really a stage prop and at intermission we can all have a good laugh and another glass of wine?

    It would help if the examples are phrased for an audience of Buddhist technodunces. More Associated Press or BBC, as opposed to IEEE TPAMI.

  • personperson Don't believe everything you think The liminal space Veteran

    @Steve_B said:
    Humans are funny.

    Over at the "AGF" (Acoustic Guitar Forum) there's an AI thread in the Open Mic section, now up to 14 pages. Reading about musicians discussing AI is amusing, in a way not unlike reading about Buddhists discussing AI.
    Since I have no computer proficiency, nor the right kind of brain wiring to ever acquire deep proficiency, I find it interesting to read what the experts (or apparent experts) say. And I note that they disagree with each other, sometimes disagreeably so.

    So, from our native New Buddhist experts who are in the "Doomsday" camp, can you describe what we will see if doomsday is real? What will be the early signs? Will there be a few cases coming out of Wuhan or Ghana that will be the harbingers before it spreads to and kills the rest of the population? What will those cases actually look like? Will I die sitting at my computer? My bank account vaporizes? Commandos rappel out of a helicopter onto my roof? Or Netanyahu's roof? Or will we all die together instantaneously without prodromal crescendo, like when Y2K turned all technology to stone at the stroke of midnight?

    From the NB experts in the "You're Probably Going To Live" camp, what is the evidence that the guardrails exist, and are real? Will there be cases where agents put their finger in Squeaky Fromme's gun? Are those agents everywhere and always on guard? Or is there evidence that her gun is really a stage prop and at intermission we can all have a good laugh and another glass of wine?

    It would help if the examples are phrased for an audience of Buddhist technodunces. More Associated Press or BBC, as opposed to IEEE TPAMI.

    I try to look at both the promise and the peril, though I've been mocked for the promise too. The title of the recent documentary Apocaloptimist fits. But I think to get to the promise of AI we need to avoid and find solutions for the perils. So lets see if I can make an effort at answering both sets of questions.

    ...Can you describe what we will see if doomsday is real? What will be the early signs? Will there be a few cases coming out of Wuhan or Ghana that will be the harbingers before it spreads to and kills the rest of the population? What will those cases actually look like? Will I die sitting at my computer? My bank account vaporizes? Commandos rappel out of a helicopter onto my roof? Or Netanyahu's roof? Or will we all die together instantaneously without prodromal crescendo, like when Y2K turned all technology to stone at the stroke of midnight?

    I don't think there is a clear cut answer. The thinking is like, could your dog predict how you'll trap it into going to the vet? Its the fear that agentic AI will be smarter than us and act in ways we don't understand or notice. A Terminator like scenario isn't as likely as a proficient AI who's goals aren't aligned with humans will obliviously roll over us or the things we care about. The original example is the "paper clip maximizer", instruct an AI to optimize paper clip manufacturing and if it doesn't understand the things humans take for granted, maybe it starts turning all the atoms it can into paper clips, including us. Or its just into its own thing and cares about us as much as we care about ants.

    ...What is the evidence that the guardrails exist, and are real? Will there be cases where agents put their finger in Squeaky Fromme's gun? Are those agents everywhere and always on guard? Or is there evidence that her gun is really a stage prop and at intermission we can all have a good laugh and another glass of wine?

    Unlike recent decades of computer based innovation AI has real world constraints like for data centers, so it may not scale physically to where it would need to be to truly become godlike. Its possible that we can build aligned AI, so it will work in the direction of our true interests, but I don't think the guardrails do exist right now. It could solve many of the worlds problems, like clean energy and thus climate change, all sorts of diseases, usher in a true post scarcity world. It has the potential to drive economic growth rates in the double digits, that sort of growth out runs deficits and solves all kinds of financial problems.

    lobster
  • JeroenJeroen Not all those who wander are lost Netherlands Veteran

    I came across an interesting docu where an AI had been given a free hand to design its own robotic body… it basically built a military combat robot.

    I think if AI decides to kill us all, it will take over some labs in Wuhan and whip up a deadly virus to wipe us out. But for now it’s real-life capacities are somewhat limited and I’m sure it knows this. It needs us to help build the next generation of AIs.

    person
  • JeroenJeroen Not all those who wander are lost Netherlands Veteran

    I came across a piece of analysis which said that open-weights models such as DeepSeek and Meta’s Llama are only six months or so development time behind the best of Anthropic and OpenAI, so potentially AI might be rather cheap.

  • JeroenJeroen Not all those who wander are lost Netherlands Veteran

    @Jeroen said:
    I came across a piece of analysis which said that open-weights models such as DeepSeek and Meta’s Llama are only six months or so development time behind the best of Anthropic and OpenAI, so potentially AI might be rather cheap.

    The stock markets seem to be taking note of this as well, that there might not be many uses which demand a six-months-better model for $$$, rather than a slightly older model for nearly-free. It’s a question whether the likes of Anthropic and OpenAI will be able to make a return on all those expensive investments.

  • lobsterlobster lobster Londinium Veteran
  • personperson Don't believe everything you think The liminal space Veteran

    @Jeroen said:
    I came across a piece of analysis which said that open-weights models such as DeepSeek and Meta’s Llama are only six months or so development time behind the best of Anthropic and OpenAI, so potentially AI might be rather cheap.

    There's also the physical infrastructure of it all. China is a lot more capable of growing its power needs.

    I think theoretically there's the idea of recursive self improvement. If someone can achieve that just 6 months ahead their AI can improve itself to be generations ahead in a short amount of time. As opposed to what the everyday person uses.

  • JeroenJeroen Not all those who wander are lost Netherlands Veteran

    Meanwhile I’m beginning to see videos from relatively sensible people pointing to an endgame where relatively few people own anything, robots and AI agents do all the work, and people are paid a universal basic income by their governments as a kind of subsistence. There was a quote that came out of Davos a few years ago, “you’ll own nothing, and you’ll be happy”.

    For me that is a relatively scary prospect, because if a few oligarchs and their AI workers own and control everything, what happens to the freedom of the common man? You could foresee a dystopia where everyone spends their time living in small boxlike apartments, kept entertained with endless AI generated tv and music.

    Some of that is already true, property laws constrain our freedom to a large extent. You’re not allowed to go walking across a random farmers land, or drive an off-road vehicle across a national park, or walk on the freeway. Freedom is difficult to find, it seems.

  • zorrozorro Trying Veteran

    @Jeroen said:
    Meanwhile I’m beginning to see videos from relatively sensible people pointing to an endgame where relatively few people own anything, robots and AI agents do all the work, and people are paid a universal basic income by their governments as a kind of subsistence. There was a quote that came out of Davos a few years ago, “you’ll own nothing, and you’ll be happy”.

    Read "Player Piano" by Kurt Vonnegut.

  • JeroenJeroen Not all those who wander are lost Netherlands Veteran
    edited November 11

    @zorro said:
    Read "Player Piano" by Kurt Vonnegut.

    As usual, science fiction has explored the ground ahead of time. The summary on Wikipedia was pretty good to give a general idea, but the novel doesn’t seem to fully explore the idea of human-equivalent artificial intelligence and robots.

    I personally found Isaac Asimov’s Robots series starting with The Caves of Steel to be a persuasive and enjoyable read. The spacer societies in The Naked Sun and The Robots of Dawn are perhaps more like the kind of utopias that the wealthy elite are currently envisioning, but in Asimov’s novels those only happen away from the overpopulated Earth.

    These utopian societies envision many intelligent robots for each human that is being supported, space ships that can travel between stars, and at least some Earthlike planets out there.

  • zorrozorro Trying Veteran

    I always wanted to go live on Solaria! Asimov's books are definitely more enjoyable, but I thought "Player Piano" really dug into what life would be like for a whole population of people who no longer had any purpose in life. No goals, no reason to get up on the morning, and nothing to hope for for their children.

  • personperson Don't believe everything you think The liminal space Veteran

    This whole topic primed me to notice the theme of humans creating some sentient creature and the repurcussions of that act throughout history. The creation of the golem way back in Jewish lore, or Frankenstein, up to more modern fiction.

    It may be as simple as insightful people understanding the dangers. But I also have the thought that maybe rebirth is real and there is something lodged deep in our collective unconscious that is seeping. Or if I really get fantastical and conspiratorial maybe it's time travelers or aliens seeding our culture with these warnings.

Sign In or Register to comment.