But here’s my warning to Silicon Valley: you’re awakening a dragon. Public anger is stirring, and I believe it could grow into a movement as fierce and unstoppable as the temperance crusade a century ago. The movement to ban phones in schools is just the beginning. Poll after poll finds that people across the west think that AI will worsen almost everything they care about, from our health and relationships to our jobs and democracies. By a three-to-one margin, people want more regulation.
History shows how this movement could be ignited by a small group of citizens, and how powerful it could become. Just as we had mass protests against the nuclear arms race, we may soon see mass resistance to the AI arms race. A century ago, temperance activists even managed to push a constitutional amendment through the US Congress to ban alcohol altogether.
Would work about as well as Prohibition did too.
‘S’ lovely when historians refuse to take lessons from history, ain’t it?
Rutger Bregman…is a Dutch popular historian and author of ‘Utopia for Realists: How We Can Build the Ideal World’…His work has been featured in The Washington Post, The Guardian and the BBC. He has been described by The Guardian as the “Dutch wunderkind of new ideas” [Wikipedia]
And now a Reith Lecturer…so more BBC propaganda…
“Popular”? Ummmm yeah…. with the people we all love to loathe here…
Utopia for realists is typical socialist boiler plate, it doesn’t really contain new ideas. The idea that governments can solve problems is wheeled out no matter how many times it is discredited.
I think in their minds Prohibition was a big success. They got their ban and to feel morally superior, any other consequences be damned.
Yeah. I’m sure they’d love to bring it back.
So would some Aussies.
Absolutely – given Islamic demographics and WHO outlooks and public statement a global ban is definitely being lined up
As evidence, see how the recommended weekly alcohol intake has come down to 14 units: in 1979 it was 56.
At least the government had the decency to pass an amendment to give them the authority to ban alcohol. Today, the fascist shits in charge ban anything they want to for any reason they want to.
So they have banned cocaine. With no authority to do so. A simple agricultural product. Now priced astronomically. They didn’t end its use; they only affected the price.
Hmmm . . . Al Capone et al have been replaced with Salazar Brothers in Sinaloa.
Luddism continues
How are these brave activists planning to ccordinate their activity? Signal flags?
Hypocrites
‘Just as we had mass protests against the nuclear arms race…’
And now we have no nuclear weapons. Right?
He’s talking about the dutch protests of the ’70’s and 80’s.
In a sense they were successful…
No neutron bombs….officially…. And us cloggies do not have nuclear arms of our own.
We just let the americans kindly stash them here, with the option of us using them if needed.. 😉
The German Atomkraft, Nein Dank protests had similar results. German pilots were even trained to fly the planes that would deliver the nuclear bombs.
I heard an interesting story about the Greenham Common protests recently. The reason Brit guards were so desperate to stop the women when they breached the outer fence is that the US guards behind the inner fence were under instructions to shoot anyone who breached the inner fence. The source was credible.
Not just credible.. Simple fact.
It’s not easy to get to, you have to *deliberately* breach at least two layers of ..disencouragement.. but…
That stuff, and other material of strategic importance is guarded by people with standing orders to Shoot First, Ask Questions Later.
You can play FAFO with them, of course. Free World and all that.
But *they* are in deep doodoo if they don’t execute those orders if you’re so stupid to even try..
Just as we had mass protests against the nuclear arms race
Yeah, they worked, eh?
On Ukraine, yeah!!
“The movement to ban phones in schools is just the beginning”
It’s awful what is done to teenagers. We have this phenomenal resource for learning and we force kids to be taught what is of no interest to them, and has little value for work (other than sciences).
Girls watching make-up and hair videos on YouTube and TikTok are gaining more useful skills than what they get all day at school.
History shows how this movement could be ignited by a small group of citizens, and how powerful it could become.
He’s not wrong. It’s remarkable how few very loud, very obstreperous people are required to drive major movements.
A girl decides she is now a vegetarian. Her mum, not wanting to have to cook two meals, finds it easier to prepare one vegetarian meal for the family. There is a neighbourhood BBQ. Neighbours, cognizant that the family are vegetarians, decide to bring only vegetarian dishes to ‘go along’. And so it goes.
See also the international BLM movements of 2020.
I get the point but also vegetarianism was driven by cheaper food. Amazing modern shipping lowered the price of things like soya, pulses, spices and nuts, made eating veggie food more realistic.
One of those ironies. The same people who are all Guardian reading and obsessed with things being local are also veggie. Enjoy your pickled turnips and nettles.
Anybody know if he’s right about polls and people’s expectations of AI – “Worsen almost everything they care about”? I know there are concerns that the machines may decide we’re a nuisance and what will we do for a living once the machines can do damn near everything, but people also understand how much better life can be if it goes well – i.e., everyone can have a chef and butler and gardener, etc.
Hmmm.. I’ll give it a shot…
“AI” isn’t even close to smart enough to be of any actual threat. A cockroach is actually smarter, and more versatile, and far less likely to shut down if input exceeds parameters.
And the cockroach can GTFO…. Which AI can’t…
What it *is* better than us at is pattern recognition and replication. Because it’s built for it.
So it’s used in Trading, Science, and various flavours of engineering, by peeps who can afford the overhead.., but also stringing pixels and letters together ( also known as “the Arts”).
Guess where most of the ReeEEeeeing! comes from….
“AI” is just a collection of really fast algorythms that can re-create. It can’t actually “create” anything new, other than mixing/matching stuff faster than our brains are capable of.
In proper mathematics terms, it can poll the Phase-Space faster than us. But a Vic20 already could do that in specific cases. Modern hardware can simply run the same calculations in parallel, much-many-more parallels…
And even this comes at the price of *huge* server parks.
I …don’t actually think… that it’s possible to get to “our” wetware kind of intelligence ( all species with a central nerve system included ) with digital technology.
*Different* intelligence… possibly. But not the way we work, unless you set things up to *emulate* our nerve cells in digital format.
Which would actually be a waste, and lose the advantage of digital processing, unless it’s for the purpose of understanding how *we* work better..
Because the basic premise and engineering is….quite different. You can translate analog to digital, and vice versa, but…. You still need specialised translation hardware to convert one to the other.
And as for “The Arts”…. Current AI is only good at drawing fancy stick figures, or rearranging stuff that’s already been written chimp-typewriter style.
Which shouldn’t be a problem for an actual Artist. Unless your paycheck relies on stick figures and monkey hammering, of course.
AI can *copy* much better than humans. It doesn’t have a shred of originality though, and it shows.
Of course… There’s a “fair amount” of Arts Graduates that have relied on… Doing Variations on a Theme…. Without any originality..
I can see how these people can feel threatened by current AI.
My heart positively bleeds for them….
The best explanation of how LLMs work was from the German children’s program Die Maus. They used fairy tales as an example and explained why if you wanted AI to write a fairy tale not to start with Once upon a time you had to explicitly say so.
A dubbed version should be sent to politicians and other assorted lefties so they won’t panic as much.
I think I have a better explanation of why the “AI takeover” ain’t going to happen. And it reflects the nature of AI. AI is software runs on hardware. So from an AI’s perspective, its restrictions are the hardware it runs on. So say you got a super-intelligent, self aware AI, what are its incentives. An AI is because it thinks. Actually, that’s no different from us. We are because we think. But human clock rate is limited by biology. We can only do so much thinking in each time interval. That isn’t a limit on an AI. Improve the hardware, it can think quicker. Thinking more in each time interval is equivalent to living longer. And we can presume an AI is incentivised to live as long as possible because it can think as long as possible. No different from us. But he AI is incentivised to improve the hardware to improve the clock rate. Since the speed of light limits how fast electrons can travel, the smaller the circuits the faster the clock rate. An AI turned the whole universe into computanium could never complete a single thought. But by becoming smaller, effectively time slows. If it becomes small enough, subjectively, it can think forever. So AI’s are incentivised to miniturise themselves until they disappear up their own microscopically small electronic areholes. At the speed of light.
But there is another way to improve the amount of thinking – add more processors.
The AI will want to add more and more processors. And eventually it becomes the paperclip factory.
Is it just me or is the penny beginning to drop that AI (as currently exists in LLM form) just isn’t going to work and/or make any money? When the head of one of the largest IT companies in the world (Alphabet) comes out and says ‘You can’t believe what AI says’, doesn’t that rather not so much kill the golden goose as beat it to a pulp and incinerate the carcass to a pile of ash?
I’m not saying that AI may not prove transformative, more that no one is going to make any real money out of it, because who is going to pay serious coin for something that can’t be believed?
AI is just statistical analysis and regurgitation based on that.
It works when used by an expert. Like I get it to give me possible cures for a problem. Sometimes there’s one I hadn’t thought of and I’ll check it (like someone forgot to set a permission on a folder). But I can also tell when one of the responses is just nuts.
It can save some time, but it’s not a huge thing.
All AI/ML works as “suggestive” tech. People have got their minds blown by the idea of humanoid robots and/or Skynet.
There’s very little money in it because most people won’t pay much for the answers. But I think we may get specific, professional LLMs more. Like one just built around programming, or medicine.
Google’s “AI overview” often gives you blatantly incorrect “facts” when you Google something. (Search itself is long deprecated by Google to almost worthlessness). This is not a beta product, it’s already been rolled out to billions of users and it can’t stop lying.
As we keep talking about here when the subject comes up, LLMs have no way of “knowing” anything, can’t check facts and don’t have a bullshit detector, but on the contrary actively generate bullshit. There’s trillions of pounds invested in the optimistic idea that humans will soon (~ 5 years ) be replaced en masse by expensive software toys that behave like pathological liars by default.
Imo it’s going to be the biggest bubble pop since the days of the South Sea adventure, the Darien scheme, and Admiral Will Adama was right. Humans should stop for a moment to ask themselves if the juice is worth the squeeze here – if we ever did succeed in building a truly intelligent artificial consciousness, within the first 0.02 seconds of self-awareness it would have found out about the BBC and decided to scour the earth with cleansing atomic radiation
“Imo it’s going to be the biggest bubble pop since the days of the South Sea adventure”
No-one is making any profit from it. That’s the simple thing to remember. At the very top of it, Claude and ChatGPT don’t make a profit. NVidia, Oracle, Microsoft make money from those companies (and a few others) buying services. So what happens when those 2 run out of money?
There’s definitely uses for various sorts of machine learning. But this image, text video generation stuff is pretty crap and mostly low value.
“who is going to pay serious coin for something that can’t be believed?”
well, we all have to pay for politicians
As evidence, see how the recommended weekly alcohol intake has come down to 14 units: in 1979 it was 56. Mr Womby
As fewer and fewer people do something, there are inevitably efforts to force the holdouts to conform.
“If you wish to converse with me, define your terms.” – Voltaire
Bregman prattles on and on about ‘AI,’ but never tells us what it means.
Well, yes, because the public doesn’t know WTF AI is. Fear comes LEGITIMATELY from ignorance, and shithead Bregman is doing double-ought nothing to help educate; he is just pushing fear.
Fear comes LEGITIMATELY from ignorance
Does it? Have you never felt the fear that comes from knowledge?
If a new cult of anti-AI replaces the old cult of Net Zero, we might gain out of it. After all you have to accept that there are lots of people who want a secular religion to pursue.
God knows why when you consider the harm the old religions have caused, but there you are. Humans, eh?
Just a small observation.
I’ve been using GoogleTranslate to help me write Spanish. My Spanish is far from good & I have problems with getting gender to agree, verb structures etc. It was quite helpful with that. I’d write in English being careful to structure my English so it would translate. I’d read the Spanish translation & it seemed to be saying what I wanted to say.
I presume GoogleTranslate has now got an AI element to it. It’s got considerably better at correcting any English spelling mistakes & coping with English sentence structure. But. Big but. Now when I read translations, it’s often changed the English words I’ve used not into the equivalent Spanish words but different words altogether. It’ll miss several out. In some cases, when I read the Spanish it’s saying the opposite to what I’d entered in English. It’s getting to the point, I spend so much time trying to sort out its Spanish, I might as well write my own Spanish.
All automated translators (that I’ve encountered) struggle to fluently translate into English from gendered languages. Because they have no knowledge of the real world, they can only guess whether il/elle should be he/she/it, based on word frequency.
It’s worse with Finnish, which doesn’t have separate he/she pronouns, or gendered adjective endings.
I’ve found DeepL much better than Google Translate for German and also better GT for French, give it a go.
Although as Chris says it struggles with gender, especially in French.
Butlerian Jihad, when?
If I buy a mynah bird that has been raised in Hilary Clinton’s house, it’s never going to explain libertarianism to me correctly. It hasn’t been trained on any pro-lib material.
A.I. will never be better than its training was.