I am particularly worried that our politicians are buying so heavily into the idea of these new clothes, and setting up our economy to rely on their production.
It was Starmer doing deals with the tech companies when Trump came over. I'm clueless as to what it's about. We give them all our data and they consume our energy and cooling systems - yea, but what do we get? I feel out of my depth but increasingly relying on Carol C to work it out before anyone else.
American here; It’s your strength over years that is giving me my strength now, to have faith that huge effort will succeed. The good and strong have assuredly enable a cosmic strength to do the human job. I’m not exaggerating. I needed the example of your brutal journey, to know what efforts would be needed for our whole country. Now I believe in the possibility of great things and have a realistic view of what long term and what’s needed to accomplish…you had people across this pond thinking of and cheering for you in the past. And grateful.
Absolutely Tim. Carol C has been vocal about this, and I have echoed her in trying to raise awareness of Starmer’s deals with Palantir and others. Letting them harvest our data to control and manage us, and to cosy up to Trump, is inexcusable.
I agree and each time I hear “jobs”. I think “what jobs?” Construction I guess but all the compute comes from elsewhere and how how many people are needed to babysit warehouses full of banks and banks of computers?
What a great read, Carole--BRILLIANT! Are the suity-fucks getting it all plotted now, to do it again as the Quantum Computers roll in for them to pitch. AI ain't seen nothun yet! Are you kidding me: data sorting at Quantum Speed! Holy shit!
The suity-fucks are trying there damnedest to steer us away from the nitty-gritty like this:
THE RED HATS ARE COMING TO MOW YOU DOWN WITH AUTOMATIC WEAPONS
THE RED HATS ARE COMING TO TAKE YOU TO ALIGATOR CAMP
Just like the Red Coats. They came day and night until Sam Adams and a group of Patriots took serious action against the brutalization of people from the Colonies.
The red hats swept into Charlotte today, 11/16/2025,
to beat up on WE THE PEOPLE. We need 10% of the population—39 million people—in the streets for serious non-violent resistance just like Thoreau called for.
Don’t get violent cause the red hats are just looking for an excuse to murder you in your own country! They will mow you down with automatic weapons the first chance they get.
Charlotte today. Your city is next. Their goal is to cover the voting poles to scare WE THE PEOPLE AWAY!
One way or another, my prediction is that this ends in revolution, they violate the Constitution daily, either all of it applies, or none of it does. This is to say, we no longer live in a democracy, that is gone, they didn’t bother rigging the last election, they WILL rig the midterms, or simply try to cancel them. Either way, it is my bet they will not be going peacefully, it’s do or die for them now, they KNOW they’ll be going to prison, or worse, if they lose, so they will stop at NOTHING to maintain control.
This whole ICE thing is about building a private army, because the actual military are unlikely to follow illegal orders, “immigration” is just the excuse they’re using, in my opinion. It remains to be seen how many traitors he can find to join them. A mass peaceful mobilization MIGHT do it, IF we can get 35 million in the streets, IF. Idk, I still think it would be wise to have a backup plan, cause remember the Boston “massacre”, it was, relative to polulation, only a couple of guys, but it was enough to touch off a war.
We should also bear in mind, at that time, the King’s army was severely depleted from other wars, or the outcome might have been much different. My point anyway, is that they tried every peaceful means at their disposal, and STILL ended up with a war. Don’t get me wrong, we should absolutely not be going off half cocked, and should do everything to end it peacefully, but we also need to be clear eyed and realistic about it, that is, I for one ain’t gonna be allowing myself to be mowed down without a fight.
I just think it’s better to be prepared for the possibility and not need it, than need it and not be prepared, if you see what I mean. At any rate, I don’t see it ever going back to the way it was, not without some major disruption, because all of it stays in place, clearly both sides ain’t on OUR side, they’re all on the side of the $$$, that is never gonna change in my humble opinion, ergo, revolutionary change is probably gonna be necessary in the end. But again, just my opinion, and I’m only one guy so, unlikely to change much by myself, time’ll tell, hope I’m wrong…..
Lot’s of iffy stuff, Lux, AND some wild-ass shit is brewing. But the 1%ers—who own us all, right and left wing, no matter—have too much invested in quantum computing, the mars spaceship hype, AND the 1%ers of China with a standing military of 2million are not about to let some lardgutted old dotard fuck with them—The PLA just snickers at this dipshit’s so called trade wars and proceeds full speed at techno development. So, lots of iffy stuff. I would suggest that you watch videos of what a the US military’s old Ma Duce 50cal machine guns can do in 1 minute. AND, don’t miss out! Today, 11/16/2025 Ken Burn’s incredible film, The American Revolution starts on PBS! pres47 hasn’t fully gutted it, not by a longshot! KEEP ON TRUCKIN—
You're right. If you remember that donnie little dick doesn't have any alleigence to anyone. Except the almighty buck. He will sell all you "patriots" out, just watch and see. He sold out Ukraine to Russia, he's sold out Palestine to Israel and for what MONEY! Your country needs a revolution but not of the violent kind but of the intelligence kind. You have lived in this bubble of holly-weird where all the problems are "over there" yet YOU have been part of the cause. The fact that you won't accept that fact is going to dog you into the failure of your so-called democracy. I say "so-called" because it was built on shifting sand. You stole the country and called it yours by defeating the real protectors of it. Then you made yourselves the owners by keeping her races "in their place", did you really think it would last forever? You are only 250 years old and already your society is crumbling because you thought "patriotism" was the all you needed. Sorry to break the bad news but it's about time you faced facts. Your president may rewrite history but he can't change it. Greed is your legacy!
I agree. If this isn't stopped soon America will lose. Now it's becoming apparent that your president was installed by the Russians and obviously donnie little dick is in panic mode because if the Epstein files are released it will alway show how Jeff was controlling d.l.d. way back when! This I think scares him more because he can bullshit the pedo stuff to his base but being found to be a TRAITOR to the country is a whole other ball of wax. I suspect the GOP's in Congress will see the danger of being tarred by the same brush and may come overwhelmingly to release the files so that he can't Veto the vote, which I think he is planning to do. As he has all the oligarchs onside they may try to reassure an idiot like him but surely the Senators are a tad smarter? They thought they could control such an idiot for those own purposes so now the question is, will they continue with their Project 2025 agenda or will they realise the gig is up?
Why not a Russian one? Do you think that when he asked the Russians to loan him money in the 1990's when the American banks got fed up with his bankruptcies that they gave it out of the goodness of their heart? He and JE were there together! Wake up! He was laundering Russian money and he was a man without scruples. He fucked children and beat his wife. So because he wraps himself in a piece of material with a depiction of an American flag on it you all think he's a Patriot! Give me strength! He has sold out your country for money! Anybody's money. All his life!
I guess you're an insider. You seem to know a lot of inside dope. Luckily, Biden, Clinton, and Obama didn't work for foreign interests or hang our with Jeffery Epstien. It's good you're with them on the moral high ground.
I live in Charlotte. I was born here. This has turned into shit, especially in the East side of the city. And our crime rate is 20% lower than last year. Fuck ICE.
ICE got out of Chicago after a little snow and temps just under 32 degrees. That's why those sissies are in Charlotte. Give them some Chicago style Hell!
Watch on PBS this week: Ken Burns' documentary series, "The American Revolution," premiered on PBS on November 16, 2025, and airs for six consecutive nights through November 21st.
It's an open secret in the tech world that Large Language Models are a dead end on the path to greater intelligence. For all their brilliant features on pattern matching, they have hit the limit of text availability on the web and will not advance beyond the world view of a four year old. Other approaches (and ones that several places are well advanced on) are needed. The hungry hippo pursuit of extra money for compute is blinding a lot of people from politicians to fund managers who haven't picked up the nuances.
The four gaps are real world exposure, reasoning, learning and memory (to apply the other three). Listen to Demis Hassabis, Yann Lecun, Andrew Ng and Fei-Fei Li - they are the adults in the room. I think you'll find that Google Deepmind (Demis) is probably the key person here.
Quite a claim: “It's an open secret in the tech world that Large Language Models are a dead end on the path to greater intelligence.” Is this true in India and China as well? are the tech labs there also abandoning LLMs as is claimed here that the US tech world is? Is Google’s new Gemini 3.0 just a feint? I have no insider connections, so I would appreciate those of you on the inside, like Ian I’m assuming, have some answers or good guesses about India and China. Thank.s
I don't think anyone is abandoning LLMs and am sure research on the four main gaps is widespread. The four gaps being real world experience, reasoning, learning (with immediate feedback) and memory to retain the first three.
Fundamentally, LLM models hallucinate by design. They are numerical pattern probablility machines, but which don't understand the meaning of the words they are outputting. They accept all input as unchallenged gospel. You can still ask for the top 10 books in any arcane subject area and get served with 2-3 titles that don't exist. You can ask any to draw a watch face showing any time, only to find that most models are trained on pictures from the watch industry that have the hands at 10 to 2. You can ask them to solve impossible "white to move and checkmate black in 2 moves" Chess puzzle use cases and find them solving them by changing the inputs; a bit dangerous if it were a legal case. And so forth. You do have to treat all output as if it were generated by an enthusiastic intern.
LLM-only vendors do things like adding system prompts alongside your own to give limited "if the user asks this, say that" and/or to slant the positivity of responses given. One of the Anthropic models runs to the equivalent of 50 pages of text. So solace is to invest in research to close off those four areas.
Giving the impression that LLM vendors can get to AGI by throwing ever more compute at them is a lie. No amount of money is going to help them fund the purchase orders they are PR'ing outside of large (temporary) bailouts. Their business model shows zero sign of giving them the cashflow they need.
Google have a real world model (Aria and in a practical sense, Waymo), they have experience with reasoning in specific niches (maths, protein folding) and they have learning (AlphaGo, AlphaGo Zero) in specific niches. Plus other research. Some of that work has gone into Gemini 3, and the world has noticed. The other vendors will need to step up too.
ps: I have no knowledge of work in India. Chinese ones tend be following open weight models, but there AI is being treated as a complement to the other two legs of their industrial strategy - power generation/storage (ie: rare earth materials, batteries, etc) and (robotic) movement. US vendors tend to focus on just the brain part.
I got more out of this one comment than the entire article, tbh. Thanks for the adult leads. I agree it's likely that LLMs will prove to be the pets.com of the AI era. But its also just as likely that there will be AI versions of Google, Amazon, etc. Any predictions on where the FAANGs of AI will bite hardest?
The notion that LLMs won't lead to AGI--an empirical prediction that the next few years should show us to be true or false, likely.
The notion that LLMs aren't transformational--implied here I think, at least it's what Gary Marcus keeps saying, and Peter Ochs above, if not Ian. But as an academic researcher and college teacher I'd say LLMs have already torpedoed research and teaching. Hardly pets.com. And AI companions--talk to people under 30--LLMs are transforming their personal relationships, too.
Some radical new forms of how research is credited and how teaching is done will have to be developed. Now we're just in chaos and I don't see any replacement institutions/strategies emerging. It feels very disturbing, feels like a revolution. If AGI is giant steps beyond this--God help us. LLMs are radical enough, I think, to carry us through many societal changes even if AGI proves illusive.
Houston, I really appreciate you bringing in this dimension because it’s easy to get caught in the architecture debates and forget that the biggest shockwaves are already happening in how people think, relate, and learn. You’re seeing that firsthand in academia. I’m hearing it from folks in teaching and research too — old norms are collapsing overnight with no new ones in sight, and everyone is scrambling to adapt. That kind of institutional freefall is its own form of revolution. Spot on.
For me, this is why the “dot-com bubble” analogy feels off. The dot-com crash wiped out bad business models, but it didn’t rewrite how people made meaning or how institutions functioned at a foundational level. What we’re living through now feels more like one of those inflection points you get in earlier industrial revolutions — where there’s upheaval in the material base and upheaval in the cultural superstructure at the same time.
On the material side, investors are pouring billions into compute, data centers, chips, energy — almost like building out the railroads before anyone knew exactly how society would reorganize around them. On the cultural side, as you describe, the basic architecture of meaning-making is shifting: research norms, teaching norms, relationship norms, even how young people form attachment and identity. And both layers are feeding each other.
So even if AGI ends up being a mirage or a distant horizon, I agree with you that the transformation already feels deep enough to be civilizational. Maybe the more important question isn’t “Will LLMs get to AGI?” but “What happens to the human systems we’ve relied on for coherence when the ground underneath them is already moving?”
I’d be curious how you see those two revolutions — the infrastructural one and the meaning-making one — interacting. Do you think one is driving the other, or are they converging from different directions?
There are also a few of us that have spotted that Nano Banana Pro can generate infographics. I’ve put a few of mine on LinkedIn but there are some really wild ones - from summarising text, to doing the same of an hour long video, and one that did a wall chart summarising what happened in the world in 2024.
We’re going through the “let’s ban calculators in maths exams” phase with AI in education. Main concern is the ability to use tools lead to an atrophy of critical thinking skills. But it also questions the structure of further education establishments and what/how they deliver too. I’ll leave that to my ex-colleague at Jisc, Michael Webb.
I think we're way past the calculators phase--more into tech swallows entire cultures phase! I think what Edmund Carpenter found when new media came to traditional Indigenous cultures a good way of thinking about the transformation coming for us. Confusion; bewilderment, and then it gets bad.
Thanks for the links. Off to look at them now. Very thoughtful.
Ps: I got laid off in March and my ability to get meaningful income is getting hit for six with all these changes. I manage data engineering, analytics, integration and decommission teams - with both eyes on industry changes that we can use to improve productivity and efficiency. Market is brutal at the moment in the south UK.
If hyperscalers become the de facto carriers of AI infrastructure — like the railroad barons of the 19th century — what does that mean for the next phase? Regulation? Nationalization? Open alternatives? Or do we enter a world where AI capacity is as strategically vital as oil fields?
personally unsure how this plays out. At the moment, pieces of AI are called out like APIs, and it's not clear if more limited or niche models with real world experience, reasoning, learning and/or memory will emerge. At least away from the "knows everything" versions that need hyperscaler platforms to run.
APIs will be deterministic as today or have agency with more intelligence. That's the piece I agonise over professionally with Databricks in Enterprise processes.
I've seen something today that suggests Gemini 3 has real world exposure and reasoning inbuilt. Gravity, physics and working back from a result to deduce the value of an ambiguous piece of ancient writing,
How this rewires data flows and business processes - and where economic abstraction of value relocates to - is a big case by case unknown.
That was a good read. Yikes, never thought of connecting the whole Epstein saga with the broligarchy and the ai frenzy but here we are. Just to note, that graph is missing some key players, notably Google, who also seem to be putting out the best ai tools as of late. Not sure what it means. Hopefully they're being slightly more clever than the rest of the herd.
The problem, Carole, is that both things can be true
1. This is a massive bubble
2. AI will have vast impact
Look at the big bubbles. The South Sea Company was a precursor to 150 years of colonialism. The railways bubble of 1840s was a precursor to massive industrial change and total domination of railways. The Wall St Crash was a precursor to the modern industrial age and driven by investment on its stocks. The Internet bubble was a precursor to the entire transformation of society by the digital.
Bubbles are real. Vast change is real. Both can be true. A journalist needs to navigate that.
Putting aside the hype cycles - which I'm very well acquainted - my own industry, software engineering, is, long term, entirely transformed by AI. Good developers today, equipped with the right mix of automated AI, can implement in a day what used to take 5, 10 or 50 days to do. Companies are offering automated app development for £100K in two weeks to match what the big consulting companies used to charge £1M for over 6mo. There is no doubt this industry is undergoing major change.
The big question with world-changing technologies is whether they are world-changing like the pharmaceutical industry (life-changing drugs at a life-changing price tag for the patient, resulting in a concentration of economic and political power) or world-changing like the solar panel industry (literally decentralizing electrical power, as the dramatic drops in prices over the last 20 years have suddenly made it feasible for people to slap a power source on their roof, even in relatively poor countries).
AI is a strange mix of centralised and decentralised.
It is built on a bedrock of open source software and incredibly open research communities stemming from grad schools. Nearly all the software, model architectures, training processes are pretty much open or widely understood and disseminated. This is one reason the field has been so fast, so energised, so many grad students making new breakthroughs.
The hardware is of course capital intensive and centralised in fabs and NVIDIA. That's a incredible concentration of power.
Another great point. And I think that's where AI investors see a value difference. They're not just buying websites. They're buying the most powerful chips on earth, with massive resource and supply chain implications. NSA and DOD didn't care about pets.com, but they sure as hell care about OpenAI. This is a different game, different level.
The other thing that makes it hard to judge is its applications are so widespread - this is more the steam engine than a specific discovery with specific effects - a new core capability for humanity. Occasionally for better and often much for worse.
I shed no real tear for the coders affected by changes in the software industry, for example, who have been automating other people out of their jobs for 60 years. But we're not really used to white collar automation.
The problem is that all of the A.I. developers are feeding the same data into their LLRs and only one of them is going to survive. Just like the Google competition in the early days, it will soon become apparent which system has the best options and all the others will just vanish, like morning mist. The Chinese AI that was released last year proved more efficient at less cost than the monsters being created by Altman and Co. Just think what will happen if they win the race.
I don't think the structural economics of the training industry will play out that way. I used to think one model would rule them all, but having seen it up close that's not how it's going to shake out. There is too much variation and too many niches - code models, fast models, large models, small models, compliant models, sovereign models, dodgy models, reasoning models - and applications and delivery stacks to match these. If anything models are looking surprisingly commoditised. Facebook and Mistral's move to open their weights is a major part of that. The commoditising effect of openness is very strong.
Don't get me wrong - there are huge economic and structural forces at work here, effectively adding 'foundation models" to the overall platform equations at work - just like "computer hardware', "operating system", "hyoerscale cloud", "search index" and "social information platform" all created matching dominating platforms.
Oh, how I wish this could be on the front page of every newspaper and the headline of every post on every social media platform for the foreseeable future. And I personally like to read about political change more than economics. But I am also a lifelong intuitive, one of those people who got the heebie jeebies the second she saw a photo of Donald Trump, who almost threw up every time she saw one of Richard Nixon, who broke into tears at the raping of the world when she heard about Agent Orange. Can't help it. I felt a similar vibe when I started to hear about AI, so I did a lot of reading--intelligent reading, as intuition can sometimes be tainted by personal bias--and the feeling only became stronger. The word "bubble" emerged. And it fit. I'm old enough to have lived through quite a few of these, the equivalent of pet rocks and what were those little stuffed bears people were trading a while back. All, all driven by greed and powered by hype. These tech Masters of the Universe know they're full of it. It's criminal. Maybe they are greedy and psychopathic enough to love the fact the the Trump Sh** Show is distracting the general public for the great risks their untried unstable product pose not only for our psychological wellbeing but our economy. OMG.
Thanks for this Carole. I'm no computer whizz but the idea that AI is the the best thing since sliced bread has always left me super sceptical. If for instance, it's true that would mean that millions of people all over the world would not have a job! What happens to them if the governments can't afford to take care of them because they have to supply the AI monster with his daily feeding schedule of money? Also, once they have all your data from birth to death who's to say they can't use it in nefarious ways. I've also heard that they cannot (or will not) take their algorithms out out of the system they installed it into, if the government decides belatedly that they don't want it! So where does that leave us? I do not want to be controlled in an Orwellian way by anybody, wether it's human or AI. I hope you can keep finding out what's coming up before it hits us bad and really hurts us.
It’s a strange hybrid of both! Made by Asif Kapadia it takes current tech trends to build a narrative about where we’re heading and what that year could look like on the current trajectory.
Carole have you actually seen what AI agents can ALREADY do? Without any further expansion, the tech is capable of replacing millions of jobs. It has already started. Once bosses of all businesses catch up with what’s available to them, they will use it, because it saves them money. Enough of the obsession with where AI will go - look at where it is NOW and adapt quickly.
Then Govt needs to make a choice. Let AI take the jobs and provide universal income to the jobless. Or tax the use of AI so that employing humans remains cheaper. Or to put it another way - as humans do we want paid employment or not?
It's much bigger than that unfortunately. Intelligence is powerful. It can be used for good - and all our lives improve, and ethical guardrails are placed on AI to protect people, perhaps with a universal basic income included. Or it can be used for bad, and the wealth hoarders make life even more difficult for those who have lost their jobs and have nowhere to turn - not to mention the control they will have over food systems, weapons etc. Do any of us trust our governments, of whatever colour, in whichever country we live, to make the right decision for everyone on AI?
Guardrails are very weak and currently really unreliable ! AI goes bad very quickly . The Broligarchs don’t care Meta child protections are still largely ineffective . Do you want your SNAP qualification to be reliant on AI which frequently makes mistakes with no oversight or accountability . No oversight in a fully automated Trump regime ! The Broligarchy wants real slavery back to achieve their Golden Age😱
I agree with all your concerns, but AI is not going away. It’s a weapon that can be used for good or evil. It will undoubtedly be used for both. So the more that people can engage, and shape it in the ‘good’ direction, the better. It’s another battle but it has to be taken on.
I picked a simple example. Protein folding is far more complex and above my head, but AI has shaved off decades of work and thousands of dollars. Even at a basic level, it is more efficient than millions of humans. Just look at how many thousands of jobs Amazon has cut for example. There’s no point anyone denying that this technology will be incredibly disruptive.
I have no doubt AI will replace jobs, move science along, and cause disruption. I always think the finance and investment makes absolutely no sense. It's just that picking it apart needs a bit more nuance - from your Amazon example: https://edition.cnn.com/2025/10/30/tech/amazon-layoffs-andy-jassy-ai-culture
Klarna is an example of how "AI" actually failed to replace humans. (For those not familiar with the subject, it was widely covered in press in May-June this year.)
Amazon is an example of how fake claims of "AI" are used to "whitewash" layoffs (where in the past a company laying people off would look bad, now they can just say "AI" and everything suddently looks great.) But their claims actually have no substance:
AlphaFold is a different technology from the one involved in the current bubble (the bubble is specifically about LLMs and not ML/DL in general) and is a tool, not intended to replace humans, but to aid scientists in their work. (And it's not without significant limitations.)
Ok I may not have quoted the best examples. But the facts remain, AI has taken hundreds of thousands of jobs, and will continue to claim the place of millions of people worldwide. I have seen AI in positive action through optimising microgrids. I have also seen it replacing people in my own industry of media/PR/content creation. And let’s face it, AI has been used since before 2010 for such things as social media and Google advertising. This isn’t new tech, it’s just finally reaching the level that disrupts. So it’s going nowhere.
Do you have any real evidence to back up the claims of "AI has taken hundreds of thousands of jobs"? Real, independently verified evidence that such replacement has actually happenned and was successful (not rolled back after a couple of months), and not someone's self-serving but unsubstantiated claims.
Not everything is being rolled back. Here’s just one article about just one sector in just one country, but it doesn’t take much research to see the big picture, sadly.
There was always a belief by employers that they needed fresh people who would learn the trade and fill more senior positions as they became vacent. A.I. is replacing all of them. Eventually the newcomers will find a niche, perhaps by coming in already proficient in technology, and ready to do the work required. How long this interregnum will last is anyone’s business but it is starting to show.
I think we humans also have to be more honest with ourselves. We are not the perfect working models we think we are. Plenty of us fall short, even into incompetence. Hopefully not myself or yourself though!
Definitely true, we are not perfect working models. But neither is AI or any other automated system. Just the fact that so many like my boss think his employees are unnecessary is disheartening (and he's running a 100-year-old company).
I suspect, neighbor, that what this'll all mean, is that we'll get even more incredibly manipulative and utterly useless 'customer service' from 'Ai' driven responses to our requests to correct or amend the umpteen 'subscription services' that 'maintain' our 'internet presence'. I'm changing my visa card number and canceling all of them. Have I mentioned the pleasure of typing on my 1946 Remington Office Manual Deluxe? And getting actual letters in the mail? (until The Man totally enshittifies THAT, too).
Customer service is actually a good example of what I'm talking about. Klarna’s AI customer-service agent performs the work of roughly 700 full-time human employees. It resolves issues with a similar customer satisfaction to humans, cuts average resolution time from 11 minutes to under 2 minutes per case, and works 24/7 in 23 markets and more than 35 languages. Which is why Klarna’s CEO admitted they could halve the overall workforce while still growing the business. That's the AI world we live in NOW - not a vision of a dystopian future.
If AI could really replace so many jobs, the service providers would be earning giant profits from it. (Productivity gains are split between buyer and seller.)
So much of the hype has people believing that LLMs generating AI slop is adding value. Only in very technical fields do I see gains from AI, or rather call it machine learning.
It does all seem very male-dominated as you say Carole, with bravado, a gung-ho attitude, phenomenal wealth, hubris, and toxic masculinity all in the mix. No one really knows how it will end, but compared to the juggernaut coming down the road of climate change, it is a sideshow!
Curious as to how Michael Wolff will emerge from this. He knows so much and is a ubiquitous Talking Head, but can he come out without a liberal coating of the faecal matter, even if what he provides us with is information treasure? And does it ultimately matter if it brings the whole ghastly infrastructure down?
That Altman interview isn't the meltdown you describe though. A bit of edgy peevishness that wouldn't be remembered after 5 minutes at a dinner party, I don't see it featuring in Netflix docs any time soon. None of which means you're wrong in your prognostications, and if we've got the possibility of two global disasters ahead of us, either a spectacular collapse of the AI circus or a totalitarian domination by AI, I would certainly choose the former (it is recoverable from). You're always heartening to read Carole, but I don't think you hit the nail here.
Glad to read your common sense!
I am particularly worried that our politicians are buying so heavily into the idea of these new clothes, and setting up our economy to rely on their production.
It was Starmer doing deals with the tech companies when Trump came over. I'm clueless as to what it's about. We give them all our data and they consume our energy and cooling systems - yea, but what do we get? I feel out of my depth but increasingly relying on Carol C to work it out before anyone else.
Yep. Absolutely spot on. I find the lack of scrutiny around that totally alarming & grossly naive
American here; It’s your strength over years that is giving me my strength now, to have faith that huge effort will succeed. The good and strong have assuredly enable a cosmic strength to do the human job. I’m not exaggerating. I needed the example of your brutal journey, to know what efforts would be needed for our whole country. Now I believe in the possibility of great things and have a realistic view of what long term and what’s needed to accomplish…you had people across this pond thinking of and cheering for you in the past. And grateful.
Absolutely Tim. Carol C has been vocal about this, and I have echoed her in trying to raise awareness of Starmer’s deals with Palantir and others. Letting them harvest our data to control and manage us, and to cosy up to Trump, is inexcusable.
I agree with Tim. “ I feel out of my depth but increasingly rely on Carol C to work it out before anyone else” . She always does……..
"Gave them?" DOGE and Musk stole it all. Just thinking about that almost gives me an anxiety attack. I'm waiting for the other shoe to fall.
I agree and each time I hear “jobs”. I think “what jobs?” Construction I guess but all the compute comes from elsewhere and how how many people are needed to babysit warehouses full of banks and banks of computers?
You get above inflation increases to your utility bills.
Those clothes are straightaway destined for the charity shop.
What a great read, Carole--BRILLIANT! Are the suity-fucks getting it all plotted now, to do it again as the Quantum Computers roll in for them to pitch. AI ain't seen nothun yet! Are you kidding me: data sorting at Quantum Speed! Holy shit!
The suity-fucks are trying there damnedest to steer us away from the nitty-gritty like this:
THE RED HATS ARE COMING TO MOW YOU DOWN WITH AUTOMATIC WEAPONS
THE RED HATS ARE COMING TO TAKE YOU TO ALIGATOR CAMP
Just like the Red Coats. They came day and night until Sam Adams and a group of Patriots took serious action against the brutalization of people from the Colonies.
The red hats swept into Charlotte today, 11/16/2025,
to beat up on WE THE PEOPLE. We need 10% of the population—39 million people—in the streets for serious non-violent resistance just like Thoreau called for.
Don’t get violent cause the red hats are just looking for an excuse to murder you in your own country! They will mow you down with automatic weapons the first chance they get.
Charlotte today. Your city is next. Their goal is to cover the voting poles to scare WE THE PEOPLE AWAY!
NOTES FROM PATRIOTS
Love the ‘suity-fucks’! Spot on!!
One way or another, my prediction is that this ends in revolution, they violate the Constitution daily, either all of it applies, or none of it does. This is to say, we no longer live in a democracy, that is gone, they didn’t bother rigging the last election, they WILL rig the midterms, or simply try to cancel them. Either way, it is my bet they will not be going peacefully, it’s do or die for them now, they KNOW they’ll be going to prison, or worse, if they lose, so they will stop at NOTHING to maintain control.
This whole ICE thing is about building a private army, because the actual military are unlikely to follow illegal orders, “immigration” is just the excuse they’re using, in my opinion. It remains to be seen how many traitors he can find to join them. A mass peaceful mobilization MIGHT do it, IF we can get 35 million in the streets, IF. Idk, I still think it would be wise to have a backup plan, cause remember the Boston “massacre”, it was, relative to polulation, only a couple of guys, but it was enough to touch off a war.
We should also bear in mind, at that time, the King’s army was severely depleted from other wars, or the outcome might have been much different. My point anyway, is that they tried every peaceful means at their disposal, and STILL ended up with a war. Don’t get me wrong, we should absolutely not be going off half cocked, and should do everything to end it peacefully, but we also need to be clear eyed and realistic about it, that is, I for one ain’t gonna be allowing myself to be mowed down without a fight.
I just think it’s better to be prepared for the possibility and not need it, than need it and not be prepared, if you see what I mean. At any rate, I don’t see it ever going back to the way it was, not without some major disruption, because all of it stays in place, clearly both sides ain’t on OUR side, they’re all on the side of the $$$, that is never gonna change in my humble opinion, ergo, revolutionary change is probably gonna be necessary in the end. But again, just my opinion, and I’m only one guy so, unlikely to change much by myself, time’ll tell, hope I’m wrong…..
Lot’s of iffy stuff, Lux, AND some wild-ass shit is brewing. But the 1%ers—who own us all, right and left wing, no matter—have too much invested in quantum computing, the mars spaceship hype, AND the 1%ers of China with a standing military of 2million are not about to let some lardgutted old dotard fuck with them—The PLA just snickers at this dipshit’s so called trade wars and proceeds full speed at techno development. So, lots of iffy stuff. I would suggest that you watch videos of what a the US military’s old Ma Duce 50cal machine guns can do in 1 minute. AND, don’t miss out! Today, 11/16/2025 Ken Burn’s incredible film, The American Revolution starts on PBS! pres47 hasn’t fully gutted it, not by a longshot! KEEP ON TRUCKIN—
You're right. If you remember that donnie little dick doesn't have any alleigence to anyone. Except the almighty buck. He will sell all you "patriots" out, just watch and see. He sold out Ukraine to Russia, he's sold out Palestine to Israel and for what MONEY! Your country needs a revolution but not of the violent kind but of the intelligence kind. You have lived in this bubble of holly-weird where all the problems are "over there" yet YOU have been part of the cause. The fact that you won't accept that fact is going to dog you into the failure of your so-called democracy. I say "so-called" because it was built on shifting sand. You stole the country and called it yours by defeating the real protectors of it. Then you made yourselves the owners by keeping her races "in their place", did you really think it would last forever? You are only 250 years old and already your society is crumbling because you thought "patriotism" was the all you needed. Sorry to break the bad news but it's about time you faced facts. Your president may rewrite history but he can't change it. Greed is your legacy!
I agree. If this isn't stopped soon America will lose. Now it's becoming apparent that your president was installed by the Russians and obviously donnie little dick is in panic mode because if the Epstein files are released it will alway show how Jeff was controlling d.l.d. way back when! This I think scares him more because he can bullshit the pedo stuff to his base but being found to be a TRAITOR to the country is a whole other ball of wax. I suspect the GOP's in Congress will see the danger of being tarred by the same brush and may come overwhelmingly to release the files so that he can't Veto the vote, which I think he is planning to do. As he has all the oligarchs onside they may try to reassure an idiot like him but surely the Senators are a tad smarter? They thought they could control such an idiot for those own purposes so now the question is, will they continue with their Project 2025 agenda or will they realise the gig is up?
Pretty sure if Trump is a traitor he's an Israeli operator not a Russian one.
Why not a Russian one? Do you think that when he asked the Russians to loan him money in the 1990's when the American banks got fed up with his bankruptcies that they gave it out of the goodness of their heart? He and JE were there together! Wake up! He was laundering Russian money and he was a man without scruples. He fucked children and beat his wife. So because he wraps himself in a piece of material with a depiction of an American flag on it you all think he's a Patriot! Give me strength! He has sold out your country for money! Anybody's money. All his life!
I guess you're an insider. You seem to know a lot of inside dope. Luckily, Biden, Clinton, and Obama didn't work for foreign interests or hang our with Jeffery Epstien. It's good you're with them on the moral high ground.
Are you being facetious? Why do you think it's necessary to be an "insider" to read and listen and have a functioning brain?
I live in Charlotte. I was born here. This has turned into shit, especially in the East side of the city. And our crime rate is 20% lower than last year. Fuck ICE.
I’m sorry for Charlotte’s invasion. Sending strength and support to you all. 🫶🏼
ICE got out of Chicago after a little snow and temps just under 32 degrees. That's why those sissies are in Charlotte. Give them some Chicago style Hell!
Watch on PBS this week: Ken Burns' documentary series, "The American Revolution," premiered on PBS on November 16, 2025, and airs for six consecutive nights through November 21st.
It's an open secret in the tech world that Large Language Models are a dead end on the path to greater intelligence. For all their brilliant features on pattern matching, they have hit the limit of text availability on the web and will not advance beyond the world view of a four year old. Other approaches (and ones that several places are well advanced on) are needed. The hungry hippo pursuit of extra money for compute is blinding a lot of people from politicians to fund managers who haven't picked up the nuances.
The four gaps are real world exposure, reasoning, learning and memory (to apply the other three). Listen to Demis Hassabis, Yann Lecun, Andrew Ng and Fei-Fei Li - they are the adults in the room. I think you'll find that Google Deepmind (Demis) is probably the key person here.
Quite a claim: “It's an open secret in the tech world that Large Language Models are a dead end on the path to greater intelligence.” Is this true in India and China as well? are the tech labs there also abandoning LLMs as is claimed here that the US tech world is? Is Google’s new Gemini 3.0 just a feint? I have no insider connections, so I would appreciate those of you on the inside, like Ian I’m assuming, have some answers or good guesses about India and China. Thank.s
I don't think anyone is abandoning LLMs and am sure research on the four main gaps is widespread. The four gaps being real world experience, reasoning, learning (with immediate feedback) and memory to retain the first three.
Fundamentally, LLM models hallucinate by design. They are numerical pattern probablility machines, but which don't understand the meaning of the words they are outputting. They accept all input as unchallenged gospel. You can still ask for the top 10 books in any arcane subject area and get served with 2-3 titles that don't exist. You can ask any to draw a watch face showing any time, only to find that most models are trained on pictures from the watch industry that have the hands at 10 to 2. You can ask them to solve impossible "white to move and checkmate black in 2 moves" Chess puzzle use cases and find them solving them by changing the inputs; a bit dangerous if it were a legal case. And so forth. You do have to treat all output as if it were generated by an enthusiastic intern.
LLM-only vendors do things like adding system prompts alongside your own to give limited "if the user asks this, say that" and/or to slant the positivity of responses given. One of the Anthropic models runs to the equivalent of 50 pages of text. So solace is to invest in research to close off those four areas.
Giving the impression that LLM vendors can get to AGI by throwing ever more compute at them is a lie. No amount of money is going to help them fund the purchase orders they are PR'ing outside of large (temporary) bailouts. Their business model shows zero sign of giving them the cashflow they need.
Google have a real world model (Aria and in a practical sense, Waymo), they have experience with reasoning in specific niches (maths, protein folding) and they have learning (AlphaGo, AlphaGo Zero) in specific niches. Plus other research. Some of that work has gone into Gemini 3, and the world has noticed. The other vendors will need to step up too.
ps: I have no knowledge of work in India. Chinese ones tend be following open weight models, but there AI is being treated as a complement to the other two legs of their industrial strategy - power generation/storage (ie: rare earth materials, batteries, etc) and (robotic) movement. US vendors tend to focus on just the brain part.
There's a very long article somewhere about the Electric Stack.I had summarised in a NoteBookLM presentation. See: https://drive.google.com/file/d/1hUwbUKsY9dQe43vpUjfKh74daHPL9C_M/view?usp=sharing
I got more out of this one comment than the entire article, tbh. Thanks for the adult leads. I agree it's likely that LLMs will prove to be the pets.com of the AI era. But its also just as likely that there will be AI versions of Google, Amazon, etc. Any predictions on where the FAANGs of AI will bite hardest?
My guess is that LLMs will be eventually subsumed into the hyperscalers.
The notion that LLMs won't lead to AGI--an empirical prediction that the next few years should show us to be true or false, likely.
The notion that LLMs aren't transformational--implied here I think, at least it's what Gary Marcus keeps saying, and Peter Ochs above, if not Ian. But as an academic researcher and college teacher I'd say LLMs have already torpedoed research and teaching. Hardly pets.com. And AI companions--talk to people under 30--LLMs are transforming their personal relationships, too.
Some radical new forms of how research is credited and how teaching is done will have to be developed. Now we're just in chaos and I don't see any replacement institutions/strategies emerging. It feels very disturbing, feels like a revolution. If AGI is giant steps beyond this--God help us. LLMs are radical enough, I think, to carry us through many societal changes even if AGI proves illusive.
Houston, I really appreciate you bringing in this dimension because it’s easy to get caught in the architecture debates and forget that the biggest shockwaves are already happening in how people think, relate, and learn. You’re seeing that firsthand in academia. I’m hearing it from folks in teaching and research too — old norms are collapsing overnight with no new ones in sight, and everyone is scrambling to adapt. That kind of institutional freefall is its own form of revolution. Spot on.
For me, this is why the “dot-com bubble” analogy feels off. The dot-com crash wiped out bad business models, but it didn’t rewrite how people made meaning or how institutions functioned at a foundational level. What we’re living through now feels more like one of those inflection points you get in earlier industrial revolutions — where there’s upheaval in the material base and upheaval in the cultural superstructure at the same time.
On the material side, investors are pouring billions into compute, data centers, chips, energy — almost like building out the railroads before anyone knew exactly how society would reorganize around them. On the cultural side, as you describe, the basic architecture of meaning-making is shifting: research norms, teaching norms, relationship norms, even how young people form attachment and identity. And both layers are feeding each other.
So even if AGI ends up being a mirage or a distant horizon, I agree with you that the transformation already feels deep enough to be civilizational. Maybe the more important question isn’t “Will LLMs get to AGI?” but “What happens to the human systems we’ve relied on for coherence when the ground underneath them is already moving?”
I’d be curious how you see those two revolutions — the infrastructural one and the meaning-making one — interacting. Do you think one is driving the other, or are they converging from different directions?
Have a look at this: https://www.reddit.com/r/copywriting/s/0CpYQj1lSv
There are also a few of us that have spotted that Nano Banana Pro can generate infographics. I’ve put a few of mine on LinkedIn but there are some really wild ones - from summarising text, to doing the same of an hour long video, and one that did a wall chart summarising what happened in the world in 2024.
See: https://www.linkedin.com/posts/ianwaring_another-gemini-jaw-drop-i-have-a-history-activity-7397736448747278337-WiOW
Also: https://www.reddit.com/r/nanobanana/s/2MEu4MeWxY
We’re going through the “let’s ban calculators in maths exams” phase with AI in education. Main concern is the ability to use tools lead to an atrophy of critical thinking skills. But it also questions the structure of further education establishments and what/how they deliver too. I’ll leave that to my ex-colleague at Jisc, Michael Webb.
I think we're way past the calculators phase--more into tech swallows entire cultures phase! I think what Edmund Carpenter found when new media came to traditional Indigenous cultures a good way of thinking about the transformation coming for us. Confusion; bewilderment, and then it gets bad.
Thanks for the links. Off to look at them now. Very thoughtful.
Here's Carpenter: https://mindrevolution.substack.com/p/blow-by-blow?r=2cybp
Ps: I got laid off in March and my ability to get meaningful income is getting hit for six with all these changes. I manage data engineering, analytics, integration and decommission teams - with both eyes on industry changes that we can use to improve productivity and efficiency. Market is brutal at the moment in the south UK.
If hyperscalers become the de facto carriers of AI infrastructure — like the railroad barons of the 19th century — what does that mean for the next phase? Regulation? Nationalization? Open alternatives? Or do we enter a world where AI capacity is as strategically vital as oil fields?
personally unsure how this plays out. At the moment, pieces of AI are called out like APIs, and it's not clear if more limited or niche models with real world experience, reasoning, learning and/or memory will emerge. At least away from the "knows everything" versions that need hyperscaler platforms to run.
APIs will be deterministic as today or have agency with more intelligence. That's the piece I agonise over professionally with Databricks in Enterprise processes.
I've seen something today that suggests Gemini 3 has real world exposure and reasoning inbuilt. Gravity, physics and working back from a result to deduce the value of an ambiguous piece of ancient writing,
How this rewires data flows and business processes - and where economic abstraction of value relocates to - is a big case by case unknown.
The portion of the sentence, "...it'll be too soon if we never hear from him again.", is brilliant. As is your reporting.
That was a good read. Yikes, never thought of connecting the whole Epstein saga with the broligarchy and the ai frenzy but here we are. Just to note, that graph is missing some key players, notably Google, who also seem to be putting out the best ai tools as of late. Not sure what it means. Hopefully they're being slightly more clever than the rest of the herd.
The problem, Carole, is that both things can be true
1. This is a massive bubble
2. AI will have vast impact
Look at the big bubbles. The South Sea Company was a precursor to 150 years of colonialism. The railways bubble of 1840s was a precursor to massive industrial change and total domination of railways. The Wall St Crash was a precursor to the modern industrial age and driven by investment on its stocks. The Internet bubble was a precursor to the entire transformation of society by the digital.
Bubbles are real. Vast change is real. Both can be true. A journalist needs to navigate that.
Putting aside the hype cycles - which I'm very well acquainted - my own industry, software engineering, is, long term, entirely transformed by AI. Good developers today, equipped with the right mix of automated AI, can implement in a day what used to take 5, 10 or 50 days to do. Companies are offering automated app development for £100K in two weeks to match what the big consulting companies used to charge £1M for over 6mo. There is no doubt this industry is undergoing major change.
Yes. I think that’s not an unlikely outcome. These are tools not God & if you look at dot com bubble, it was a bubble & a world-changing technology
The big question with world-changing technologies is whether they are world-changing like the pharmaceutical industry (life-changing drugs at a life-changing price tag for the patient, resulting in a concentration of economic and political power) or world-changing like the solar panel industry (literally decentralizing electrical power, as the dramatic drops in prices over the last 20 years have suddenly made it feasible for people to slap a power source on their roof, even in relatively poor countries).
AI is a strange mix of centralised and decentralised.
It is built on a bedrock of open source software and incredibly open research communities stemming from grad schools. Nearly all the software, model architectures, training processes are pretty much open or widely understood and disseminated. This is one reason the field has been so fast, so energised, so many grad students making new breakthroughs.
The hardware is of course capital intensive and centralised in fabs and NVIDIA. That's a incredible concentration of power.
Another great point. And I think that's where AI investors see a value difference. They're not just buying websites. They're buying the most powerful chips on earth, with massive resource and supply chain implications. NSA and DOD didn't care about pets.com, but they sure as hell care about OpenAI. This is a different game, different level.
The other thing that makes it hard to judge is its applications are so widespread - this is more the steam engine than a specific discovery with specific effects - a new core capability for humanity. Occasionally for better and often much for worse.
I shed no real tear for the coders affected by changes in the software industry, for example, who have been automating other people out of their jobs for 60 years. But we're not really used to white collar automation.
The problem is that all of the A.I. developers are feeding the same data into their LLRs and only one of them is going to survive. Just like the Google competition in the early days, it will soon become apparent which system has the best options and all the others will just vanish, like morning mist. The Chinese AI that was released last year proved more efficient at less cost than the monsters being created by Altman and Co. Just think what will happen if they win the race.
I don't think the structural economics of the training industry will play out that way. I used to think one model would rule them all, but having seen it up close that's not how it's going to shake out. There is too much variation and too many niches - code models, fast models, large models, small models, compliant models, sovereign models, dodgy models, reasoning models - and applications and delivery stacks to match these. If anything models are looking surprisingly commoditised. Facebook and Mistral's move to open their weights is a major part of that. The commoditising effect of openness is very strong.
Don't get me wrong - there are huge economic and structural forces at work here, effectively adding 'foundation models" to the overall platform equations at work - just like "computer hardware', "operating system", "hyoerscale cloud", "search index" and "social information platform" all created matching dominating platforms.
Oh, how I wish this could be on the front page of every newspaper and the headline of every post on every social media platform for the foreseeable future. And I personally like to read about political change more than economics. But I am also a lifelong intuitive, one of those people who got the heebie jeebies the second she saw a photo of Donald Trump, who almost threw up every time she saw one of Richard Nixon, who broke into tears at the raping of the world when she heard about Agent Orange. Can't help it. I felt a similar vibe when I started to hear about AI, so I did a lot of reading--intelligent reading, as intuition can sometimes be tainted by personal bias--and the feeling only became stronger. The word "bubble" emerged. And it fit. I'm old enough to have lived through quite a few of these, the equivalent of pet rocks and what were those little stuffed bears people were trading a while back. All, all driven by greed and powered by hype. These tech Masters of the Universe know they're full of it. It's criminal. Maybe they are greedy and psychopathic enough to love the fact the the Trump Sh** Show is distracting the general public for the great risks their untried unstable product pose not only for our psychological wellbeing but our economy. OMG.
Ed Zitron is also excellent on the bubblenomics of it all.
Having just finished reading "Enshittification", by Cory Doctorow, I look forward to hearing about your conversation with Audrey Tang.
Carole, you are talkin' some SHIT in this piece, and I loved every word of it.
Hahaha! Thank you. Great compliment
Thanks for this Carole. I'm no computer whizz but the idea that AI is the the best thing since sliced bread has always left me super sceptical. If for instance, it's true that would mean that millions of people all over the world would not have a job! What happens to them if the governments can't afford to take care of them because they have to supply the AI monster with his daily feeding schedule of money? Also, once they have all your data from birth to death who's to say they can't use it in nefarious ways. I've also heard that they cannot (or will not) take their algorithms out out of the system they installed it into, if the government decides belatedly that they don't want it! So where does that leave us? I do not want to be controlled in an Orwellian way by anybody, wether it's human or AI. I hope you can keep finding out what's coming up before it hits us bad and really hurts us.
AI drones that can target an apartment /person or a specific car…. Frightening. Watch 2073.
Is 2073 a movie or documentary or what Andrea? Aren't they using that already from Russia with hate on Ukraine?
It’s a strange hybrid of both! Made by Asif Kapadia it takes current tech trends to build a narrative about where we’re heading and what that year could look like on the current trajectory.
Thanks. I'll check it out. Even at 76 I can never know enough!
Carole have you actually seen what AI agents can ALREADY do? Without any further expansion, the tech is capable of replacing millions of jobs. It has already started. Once bosses of all businesses catch up with what’s available to them, they will use it, because it saves them money. Enough of the obsession with where AI will go - look at where it is NOW and adapt quickly.
Then Govt needs to make a choice. Let AI take the jobs and provide universal income to the jobless. Or tax the use of AI so that employing humans remains cheaper. Or to put it another way - as humans do we want paid employment or not?
It's much bigger than that unfortunately. Intelligence is powerful. It can be used for good - and all our lives improve, and ethical guardrails are placed on AI to protect people, perhaps with a universal basic income included. Or it can be used for bad, and the wealth hoarders make life even more difficult for those who have lost their jobs and have nowhere to turn - not to mention the control they will have over food systems, weapons etc. Do any of us trust our governments, of whatever colour, in whichever country we live, to make the right decision for everyone on AI?
Guardrails are very weak and currently really unreliable ! AI goes bad very quickly . The Broligarchs don’t care Meta child protections are still largely ineffective . Do you want your SNAP qualification to be reliant on AI which frequently makes mistakes with no oversight or accountability . No oversight in a fully automated Trump regime ! The Broligarchy wants real slavery back to achieve their Golden Age😱
I agree with all your concerns, but AI is not going away. It’s a weapon that can be used for good or evil. It will undoubtedly be used for both. So the more that people can engage, and shape it in the ‘good’ direction, the better. It’s another battle but it has to be taken on.
Definitely not
We may not be the ones making that decision.
That was almost 2 years' ago. And Klarna is starting to find out that AI isn't quite what it was sold as: https://tech.co/news/klarna-reverses-ai-overhaul
I picked a simple example. Protein folding is far more complex and above my head, but AI has shaved off decades of work and thousands of dollars. Even at a basic level, it is more efficient than millions of humans. Just look at how many thousands of jobs Amazon has cut for example. There’s no point anyone denying that this technology will be incredibly disruptive.
I have no doubt AI will replace jobs, move science along, and cause disruption. I always think the finance and investment makes absolutely no sense. It's just that picking it apart needs a bit more nuance - from your Amazon example: https://edition.cnn.com/2025/10/30/tech/amazon-layoffs-andy-jassy-ai-culture
Klarna is an example of how "AI" actually failed to replace humans. (For those not familiar with the subject, it was widely covered in press in May-June this year.)
Amazon is an example of how fake claims of "AI" are used to "whitewash" layoffs (where in the past a company laying people off would look bad, now they can just say "AI" and everything suddently looks great.) But their claims actually have no substance:
https://edition.cnn.com/2025/10/28/business/amazons-layoffs-ai-nightcap
AlphaFold is a different technology from the one involved in the current bubble (the bubble is specifically about LLMs and not ML/DL in general) and is a tool, not intended to replace humans, but to aid scientists in their work. (And it's not without significant limitations.)
Ok I may not have quoted the best examples. But the facts remain, AI has taken hundreds of thousands of jobs, and will continue to claim the place of millions of people worldwide. I have seen AI in positive action through optimising microgrids. I have also seen it replacing people in my own industry of media/PR/content creation. And let’s face it, AI has been used since before 2010 for such things as social media and Google advertising. This isn’t new tech, it’s just finally reaching the level that disrupts. So it’s going nowhere.
Do you have any real evidence to back up the claims of "AI has taken hundreds of thousands of jobs"? Real, independently verified evidence that such replacement has actually happenned and was successful (not rolled back after a couple of months), and not someone's self-serving but unsubstantiated claims.
https://futurism.com/companies-hiring-humans-fix-ai
Not everything is being rolled back. Here’s just one article about just one sector in just one country, but it doesn’t take much research to see the big picture, sadly.
https://www.independent.co.uk/news/world/americas/artificial-intelligence-replacing-jobs-report-b2800709.html
There was always a belief by employers that they needed fresh people who would learn the trade and fill more senior positions as they became vacent. A.I. is replacing all of them. Eventually the newcomers will find a niche, perhaps by coming in already proficient in technology, and ready to do the work required. How long this interregnum will last is anyone’s business but it is starting to show.
Like how? Adaptation…..?
Individuals have to adapt and focus on skills and services which AI is not yet superior at delivering.
We have (writing well, speaking clearly, and editing complicated texts), but my boss still thinks he can replace us with AI. He's fallen for it.
I think we humans also have to be more honest with ourselves. We are not the perfect working models we think we are. Plenty of us fall short, even into incompetence. Hopefully not myself or yourself though!
Definitely true, we are not perfect working models. But neither is AI or any other automated system. Just the fact that so many like my boss think his employees are unnecessary is disheartening (and he's running a 100-year-old company).
I suspect, neighbor, that what this'll all mean, is that we'll get even more incredibly manipulative and utterly useless 'customer service' from 'Ai' driven responses to our requests to correct or amend the umpteen 'subscription services' that 'maintain' our 'internet presence'. I'm changing my visa card number and canceling all of them. Have I mentioned the pleasure of typing on my 1946 Remington Office Manual Deluxe? And getting actual letters in the mail? (until The Man totally enshittifies THAT, too).
Tim Long, Just Up the Hill from Lock 15.
Customer service is actually a good example of what I'm talking about. Klarna’s AI customer-service agent performs the work of roughly 700 full-time human employees. It resolves issues with a similar customer satisfaction to humans, cuts average resolution time from 11 minutes to under 2 minutes per case, and works 24/7 in 23 markets and more than 35 languages. Which is why Klarna’s CEO admitted they could halve the overall workforce while still growing the business. That's the AI world we live in NOW - not a vision of a dystopian future.
Amazing. And then, just what is it that people are for when they're deprived of income and meaning? Count me out.
It really is going to be a rediscovery of self, of what it means to be human. It's far, far bigger and more disruptive than most people realise.
If AI could really replace so many jobs, the service providers would be earning giant profits from it. (Productivity gains are split between buyer and seller.)
So much of the hype has people believing that LLMs generating AI slop is adding value. Only in very technical fields do I see gains from AI, or rather call it machine learning.
Great reading Carol. Keep up the fantastic work you do
It does all seem very male-dominated as you say Carole, with bravado, a gung-ho attitude, phenomenal wealth, hubris, and toxic masculinity all in the mix. No one really knows how it will end, but compared to the juggernaut coming down the road of climate change, it is a sideshow!
Not exactly a sideshow because AI will hasten climate change and water shortages.
Curious as to how Michael Wolff will emerge from this. He knows so much and is a ubiquitous Talking Head, but can he come out without a liberal coating of the faecal matter, even if what he provides us with is information treasure? And does it ultimately matter if it brings the whole ghastly infrastructure down?
That Altman interview isn't the meltdown you describe though. A bit of edgy peevishness that wouldn't be remembered after 5 minutes at a dinner party, I don't see it featuring in Netflix docs any time soon. None of which means you're wrong in your prognostications, and if we've got the possibility of two global disasters ahead of us, either a spectacular collapse of the AI circus or a totalitarian domination by AI, I would certainly choose the former (it is recoverable from). You're always heartening to read Carole, but I don't think you hit the nail here.
I think it might be both those things
Ouch.