@stuartmills
Assistant Professor of Economics at the University of Leeds. Visiting Fellow of Behavioural Science at the LSE. Interested in Digital Economy + Behaviour + Political Economy. All views are my own. Website: siu2lh.com
There is definitely a strategic play here for China and Russia to exploit the fuel starved Asian tigers and weaken America's position in the Pacific. Russia must surely be mulling over how to exert pain on Europe and strengthen their Ukraine position, too.
The long twentieth century is over.
The UK's permission for the US to use its bases in attacking Iran amounts to complicity in the crime of aggression. It drags us into what could be an endless war, without democratic or even parliamentary consent.
Allowing British bases to be used in an illegal war of aggression is a catastrophic and historic mistake.
Britain has been dragged into another war because our Prime Minister would rather appease Donald Trump than stand up for international law.
Why is everything now gambling? Because alternative ways of 'getting ahead' no longer exist.
Why are public holidays now such big deals? Because there are fewer opportunities to achieve (and celebrate) personal successes.
And yes, why do (some) people obsess about the gym? Well...
Besides the obvious, horrible, systemic risk that embedding AI into critical systems demonstrates...
...dude, is organisaing your desktop really something you need AI to do? Really?
Higher house prices not only add to household debt but also increase the need to work for longer. We can’t imagine a future of less work while housing costs keep on rising www.theguardian.com/business/202...
Ivan Illich's radical monopoly: You can buy the product from many different vendors, but you must buy it from someone.
For those I've been discussing this with earlier - a new version of the AI/AGI and the social contract paper. Feedback very much appreciated. papers.ssrn.com/sol3/papers....
Thrilled to share a new paper with @stuartmills.bsky.social &
@casssunstein.bsky.social. The paper discusses ManipulationDetect (previously NudgeDetect), a tool I launched a few months ago that uses LLMs to highlight 'manipulative' practices online.
ideas.repec.org/p/not/notcdx...
New podcast episode: AI Bubble Economics 101! (W/Stuart Mills)
A word that is increasingly being discussed is "bubble". But what is a bubble really, are we in an AI bubble now, and if we are: what’s next?
Have a listen, share widely!
podcasts.apple.com/us/podcast/a...
2025 is coming to end and some people who claim not to be fascists are still associating themselves with fascist outlets like X.
I claim, perhaps charitably, that it's because they don't understand how the internet works. Either way, this needs to change.
berjon.com/fascintern-m...
I've been thinking about three categories of what I'm calling AI Epistemicide. Basically, how AI destroys human ways of knowing. I'll check out the book, and I think the idea of epistemia is interesting, too. Thanks!
The worse one I've had was a reviewer asking for additions because 'in the future, so and so difficult problem with AI might be solved.'
In the past couple of years, both editors and reviewers have been pushing for 'bothsides' additions to my papers which are critical of AI. For editors, I get it's a hedging strategy. I think reviewers should be more attentive of when they're doing it, though.
Interesting! We are getting better at doing the wrong things! Drucker would have a few few things to say.... "There is nothing quite so useless, as doing with great efficiency, something that should not be done at all."
I have a new paper in the BMJ Open. We find 71% of GP surgeries do not follow NHS guidance around patient registration. The guidance says patients do not need documents like photo ID, but many surgeries demand them. This will reduce people's access to care.
bmjopen.bmj.com/content/15/1...
"One feature of capitalism is its ability to create work for people to do. A world without work is incompatible with capitalism," says me in this piece on AI and job loss - heartening to see a link to my book here www.dazeddigital.com/life-culture...
Average percentage of US firms using any kind of AI to produce goods and services, plotted against average percentage of US firms anticipating they will use AI to produce goods and services in the next six months. Expectations are consistently higher than reality. Data from the US Census Bureau Trends and Outlooks Survey. Calculations my own.
Something I'm working on:
1) On average, only 9% of US firms were using any form of AI to produce products in September
2) US firms are consistently overestimate the rates of AI adoption (dotted line)
3) Expectations are maybe(?) starting to outpace reality...
This is precisely it. We are living in an era of commoditised conflict, where local conflicts can now have global effects, conflict itself can be more ambiguous, and expensive prestige weapons are increasingly for show.
Nice to be quoted alongside stalwarts like Gary Marcus and Carl Benedikt Frey in this article about the AI bubble. I'm not sure I think AI companies *should* raise prices, only that I think they will *have* to.
www.dw.com/en/will-the-...
Also, I am sure at the time all those politicians were using his work to justify austerity, he was *definitely* explaining how his work does not imply austerity economics. Heaven forbid he be misunderstood!
This reads more like Rogoff advertising his books than him making some profound comment about the state of economics discourse.
Data suggest generative AI is being used for backroom firm activities, rather than on the frontline. Adoption rates imply low productivity returns, too, which is also consistent with backroom rather than frontline applications. Some thoughts from my ongoing research:
siu2lh.com/articles/wha...
When people talk about AI boosting productivity, do they ever account for the time lost now that every website on the internet has to have an 'are you human?' check which depends entirely on whether Cloudlfare decided to switch their servers on today?
The US military took 15 years to restructure itself around IT. When they did, they followed Soviet ideas.
The Soviets lacked IT, so needed to predict how the US would use IT, to learn how to counter them.
Jumping head-first into AI means you're likely to get concussion.
siu2lh.com/articles/how...
Right now, a smart government would be massively investing in publicly owned renewables, supported by commitments to domestic manufacturing and nationalised steel production. They'd be talking about driving the price of electricity in the UK to zero, and labelling anti net zero an energy tax.
Our article on the politics of AI prompts is out now @econsocjournal.bsky.social www.tandfonline.com/doi/full/10.... 1/n
Is generative AI profitable? Probably not.
Would businesses pay more for it? Also no.
Is it politically useful? Yes.
Would politicians pay a lot for it? Also yes.
New paper with @richardrwhittle.bsky.social
papers.ssrn.com/sol3/papers....
Simon called it the "Information Fallacy." We call it "efficient inefficiency"--more efficiently doing things which aren't worth doing:
www.sciencedirect.com/science/arti...
This would be an enormous mistake for the UK and an error of judgement which will only compound as climate change bites. Gross negligence on the part of our politicians to bet the house on this stuff. Selling out our future for today's naive delusions.