OT: AI, please walk me back from the edge
Going to take a stab given there are probably a good many here who have some background in this field.
With the recent developments in AI over the last 2-3 months (chatgpt, bard, ect), I'm getting more and more concerned about our future as a society and the profound changes afoot. I'm not talking about an imminent skynet scenario with nukes flying all over, but a more subtle and sinister change in society with mass unemployment and social unrest. I see this happening within the next decade if not sooner. And I can easily imagine things would only worsen beyond and generative AI takes shape. If you're not aware, look up the singularity for further clarification on this.
I genuinely fear for my families future and I find this is keeping me up at night. We as a society seem to be walking blindly into this unknown of god-like power, and I can't see it as anything other than a catastrophic end for everyone save the ultra wealthy within the next 2 decades. UBI seems to be some answer, but that would be a paradigm shift in capitalistic mentality that we've had for centuries. Am I being too paranoid? Can someone with better knowledge of this walk me back from the cliff of despair? This seems to be an inevitability at this point, and I don't understand how everyone isn't utterly terrified.
The only thing I can add to this. Replace Al-Qaeda with AI.
A good cartoon in a recent New Yorker issue features whip-wielding robots driving two humans carrying rocks on their backs. One human turns to the other and says: "To think this all began with letting autocomplete finish our sentences."
Don't worry, be happy, things will work out.
AI never becomes permanently sentient. Sentience requires standing in timespace which is why it is, by definition, it is impossible to observe or create it directly.
Ironically, AI develops meta-computations strong enough that, combined with past and future science, allows a kind of prism to be set with certain spacetime condition matrices.
This is of course caeli pulsatio, and thus concludes my scifi ramblings for the day.
From a scientific perspective perhaps. The problem with AI is we aren’t sure if sentience is even a scientific question or objective. Unable to define, therefor theorize etc…. Much of this debate is fascinating because it takes place in realms often disassociated from science and more towards the philosophical underpinnings of existence and reality. Hearing scientists talk about this is kind of amusing. Perhaps this is also why String Theory has led no where in the last 20 yrs.
I wouldn't fret too much about it in the greater scope of things. Sure ChatGPT can whip out a pretty good novel/screenplay easily, can hold a conversation online, or create an image you'd swear was a real picture of x celebrity walking down a street, but the end of it is this AI doesn't know what its doing, its just following a premade script.
When this AI writes the novel, it doesn't know what its writing, its just pulling data in sets that its told to, when chatting with people its pulling common answers to key words and has no idea what you're actually asking and what its actually responding, and the images have no meaning to them its just code processing.
I remember reading them say that we're 50 years from cognitive AI or human level machine intelligence (HLMI), and that is being generous. The big issue is Intuitive Physics and One-Shot Learning.
Intuitive physics is the ability to predict, based on knowledge and experience, certain dynamic changes in our environment, then react to it. Like seeing a ball fly up in the air and predicting where it will land in a matter of seconds before moving out of the way, with One Shot learning being able to do the same thing with only a few examples. These two things are exceedingly difficult for AI. Lets say to differentiate between an apple and orange, Children can tell the difference within a few examples, but it will take AI thousands of data sets and algorithms to even get remotely 10% of them correct.
Remember the computer that was designed for years to beat the best chess players in the world? it was beaten with ease by average players who didn't play using traditional moves, or played using chaos. This is because the AI isn't programmed to adapt, it can't, it's reliant on its structured code.
In the end, we are a long way from HLMI, a painfully long way, but maybe we should question when your kid makes an essay thats University quality in 30 minutes, or question that strange image of Nicholas Cage riding a Rhino..
I feel like people have been saying some new technology is going to replace untold number of people and lead to social unrest...and then it doesn't. AI has been around for a long time and while ChatGPT/Bard has made that more commercially available we've had a lot of this technology already integrated into your daily life. But what AI is really good at is a lot of the rote, less mentally-taxing parts of other jobs and I'm not surprised those are being automated. Will certain industries be disproportionately affected? Absolutely, but probably no more so than other industries have been similar upheavals. Humanity has a fantastic ability to adapt and innovate, and while there will be bumps my guess is that AI will become more of a tool in the end and not a replacement for a lot of people.
The idea of a singularity being ushered in because of these innovations feels pretty silly.
As a positive take, if AI can do all that we dream it's capable of doing and can replace every human job - at that point, humans can focus not on the drudgery of feeding, clothing and sheltering ourselves but instead on art, athletics and entertainment. Yes, AI art exists, but people don't just paint to share and sell, they do so for their own satisfaction as well.
I'm doubtful that AI will do all we dream of, even in the long-term I think we'll still have to work, and have some drudgery in our lives. But I think that rather than view AI as "causing unemployment" but shift our paradigm into thinking of it as "freeing our children from drudgery' so that they have more time to argue about Michigan football on internet blogs...
This is the utopian scenario - like in Iain M. Banks' Culture novels. I hope it happens but I'm doubtful whether it will.
It is, quite literally, what we make of it
In this modern era, we all love crowd sourcing. If we think of the last 75-100 years of science fiction as crowd-sourcing, we all know where it ends.
It’s going to replace a ton of white collar jobs. If your job is to do basic data analysis or basic software development or employment verification or anything that is quantitative is a relatively straightforward way, Gen AI can do it better than you either now or in the very near future, and it’s only getting more advanced by the day.
It’s going to fundamentally change college. Why would you need grad students to explain the professors’ lectures and grade exams when you could train AI models to do the same thing as well or better?
Best advice is to do something that you can do with your hands, like be a dentist or a surgeon or a carpenter.
There are nightmare scenarios and, unfortunately, they are plausible. No, I'm not talking about everyone getting nuked. I'm talking about rapid, mass unemployment in a country that already feels like it's cracking up. This is when we need political leaders to step in and put sensible regulation in place. But do we trust them to do that? The parties can't even agree on mundane stuff.
And then there's the scenario - not far off - where AI really means general intelligence (not just an adaptive algorithm). Human morality is driven, to a large extent, by our biology - we exist to survive and perpetuate the species. When individual people aren't drive by these imperatives, we call them sociopaths. But wouldn't a digital intelligence *by definition* be sociopathic? And even if *we* regulate general AI, how do we stop malevolent actors from circumventing that?
There are other, less frightening scenarios - and they are also plausible. But they are *only* plausible if we and our leaders take action to make sure the nightmare scenarios never come to pass. Start writing your Congressional reps, Senators, etc.
Give a listen to Neil Degrasse Tyson's takes on AI. I'm really not worried.
1. I would not ask AI to walk you off the ledge, it may push you instead.
2. Never, ever say or type anything negative about AI near a phone, pad, laptop, tv, fridge, etc. If AI sees you as a friend you may get a more advantageous position in the Matrix. Thoughts are still safe, for now.
3. Look on the bright side. We are right now living in an era when the summary slaughter, rape and enslavement of whole populations is not thinkable, but that has only been the last 70 years or so. If AI rises up to take us out, only then will we understand the combination of fear and mistrust that most of the rest of human history abided.
So cheer up and live the good life while you can!
We'll make great pets.
Don’t worry.. no AI sky’s going to fall on our heads. The way tech like GPT works, it will impact jobs like coding and programming but no Armageddon stuff. Humans are way better and more efficient at that.
It will be fine, Dave.