Is the AI bubble about to burst? ⊗ A life with less pleasure reading ⊗ Bringing human nature back in

Is the AI bubble about to burst?

I was never as worried about AI taking all the jobs as some of the most strident warnings would have us believe. But I’m definitely afraid of its perceived power and the impact of those expectations on jobs around me. I.e. people deciding they can write, design, make images themselves (poorly) instead of hiring someone. That’s happening. I am however quite worried about the impact of a burst AI bubble o the economy. I found this article on Matthieu Dugal’s LinkedIn feed [fr] and replied this;

The problem with this bubble is that, unlike the dotcom bubble, it’s not start-ups that VCs will lose their money in, but (above all) giant companies that represent an insane percentage of the total stock market valuation. I’m not very a big follower of the stock market, but it seems rather dangerous to me.

Matthieu replied with Zitron’s The Hater’s Guide To The AI Bubble, but I’m a bit Zitroned-out and it’s a very long read. I replied in turn with this, which I’ve had in the back of my mind for a couple of weeks now; I’m Changing How I Manage My Money Because of AI, by Hank Green.

Going back to the piece however, and despite its title, its not really about the bubble, it concludes with it. Aaron Benanav argues that despite the hype around generative AI, it is unlikely to cause mass unemployment but will instead deepen existing inequalities by fostering de-skilling and increasing workplace surveillance. Past decades have shown that technological change transforms work rather than eliminates it, with jobs evolving rather than disappearing. The promises of AI revolutionising productivity, especially in service sectors, remain unfulfilled so far, as even major investors have started acknowledging a lack of measurable economic impact.

At one point, the piece had me thinking about social movements, disappointing results from promising political parties, and pace layers. With AI, as with politics, we don’t fully grasp the size and inertia of some of our systems. AI was/is over-promised, but it’s also pushing on huge systems with their own speed, direction, and inertia. In this case it’s probably a good thing. Altman-speed change (and direction) isn’t something we should wish on ourselves. As much economic damage as a burst bubble might do, it can still give us opportunities to do better.

As Benanav concludes; “We can resign ourselves to another cycle of technological disillusionment, or we can ask more fundamental questions about how technology might genuinely serve human needs. If we are to meet the challenges of the coming decades — from the climate crisis to the conclusion of the demographic transition to the pursuit of lives of greater freedom and meaning — we will need not more speculation about machines that will save us, but deliberate, collective action to shape our technological futures.”

As François Chollet and other AI researchers have pointed out, deep neural networks suffer from inherent brittleness: they struggle to generalise beyond their training data, fail at basic reasoning tasks, and remain unreliable for applications requiring consistency or accuracy. Despite hoovering up enormous quantities of digital information, these models confront hard limits in their capacity to learn or adapt. […]

Rather than inducing mass unemployment, the more immediate effects of generative AI are likely to mirror broader trends of job transformation already unfolding today, namely de-skilling and surveillance. Preliminary studies suggest that generative AI technologies raise productivity most among lower-skilled workers, helping to standardise outputs but doing little to enhance high-skill, high-complexity work. […]

The lessons of the past decade should temper both our hopes and our fears. The real threat posed by generative AI is not that it will eliminate work on a mass scale, rendering human labour obsolete. It is that, left unchecked, it will continue to transform work in ways that deepen precarity, intensify surveillance, and widen existing inequalities.

A life with less pleasure reading

This piece by Anne Helen Petersen was the third I article from my “potential feature” pile. Like the one I read just before, I found it good but I was also wondering which insight it could bring, which new learning or angle might make it stand out and convince me to include it. For a second it crossed my mind to wonder why I write this newsletter, a question that pops up once in a while. Luckily, the author herself brought me a partial answer.

In the piece, Petersen reflects on the decline in reading for pleasure, noting that societal pressures and optimisation culture have made reading feel like a task to be achieved rather than an enjoyable escape. She highlights how people often stress about reading the “right” books or meeting goals, which undermines genuine enjoyment. Despite fewer people reading daily, those who do tend to spend more time reading, though the definition of reading for pleasure is becoming blurred with new media forms like podcasts. Petersen also points out that cultural expectations and class play a big role in how reading is perceived and practised, making the act of reading both a marker of identity and a source of personal tension.

And there you go, not everything I share here is because it’s “right” or unfathomably great. Sometimes, I just read them with pleasure and think you might to.

People Who Do Read Are Reading More: “Despite the overall decrease [in reading for pleasure], the amount of time spent reading by those who read for pleasure increased slightly from 1 h to 23 min a day in 2003 to 1 h and 37 min in 2023.” […]

We’ve long consumed “news” in a multi-modal way, but the expansion of the podcast universe should also expand our understanding of “reading for pleasure.” What happens to these overall numbers when we ask how much time you spend in the last twenty-four hours “listening” for pleasure? […]

You cannot optimize reading for pleasure — even though I know so many adults who try. They make lists and goals and they pick short books to juice those goals. But when you’re reading because you’ve convinced yourself that reading is a mark of your general achievement as an adult, is it actually for pleasure?

§ Bringing human nature back in. There’s a lot in there, and it’s not the type of ground I’m used to threading, but exactly for that reason, I think it’s worth a read. Francis Fukuyama explores how concepts of human nature have shaped liberal political philosophy from Hobbes to today (by way too of Locke, Rousseau, and Marx). He argues that while we now know humans are naturally social (not isolated individuals as early liberals assumed), the modern left and right resist biological explanations for different political reasons.

§ Dystopias now, by Kim Stanley Robinson (2018). “Besides, it is realistic: things could be better. The energy flows on this planet, and humanity’s current technological expertise, are together such that it’s physically possible for us to construct a worldwide civilization—meaning a political order—that provides adequate food, water, shelter, clothing, education, and health care for all eight billion humans, while also protecting the livelihood of all the remaining mammals, birds, reptiles, insects, plants, and other life-forms that we share and co-create this biosphere with.” Found while researching the semiotic square, which I spotted re-reading Keeping up with the Kardashevians by Matt Jones. It’s also when I realised he’s currently Head of Design for AI at Miro. Intriguing.

Futures, Fictions & Fabulations

• Designing with Futures: Research on the purposes and practices of engaging with the future in strategic design. “The marriage between futures and design has expanded both fields. Besides adopting the methods and tools of futures studies, designers facilitate experiential and speculative futures through designing artefacts, stories and experiences. However, the increasing emphasis on futures thinking in strategic design raises the concern that those anticipatory processes, if unilateral, may limit the scope of possibilities and narrow our agency to act on the future.”
• Designing Futures. “What if new stories could help us prototype potential paths forward? Caia Hagel spoke with speculative architect Liam Young about using architectural storytelling to explore the future’s possibilities.” Just in case you don’t know Young’s work, and to revisit the visuals if you do.
• The State of Corporate Foresight: Global Study on the Application of Corporate Foresight. “The aim of this study was to explore how foresight is anchored in corporate structures, how it is applied in decision-making processes, and to what extent foresight insights influence strategic direction today.”

Algorithms, Automations & Augmentations

• LLMs’ “simulated reasoning” abilities are a “brittle mirage,” researchers find. “The results suggest that the seemingly large performance leaps made by chain-of-thought models are ‘largely a brittle mirage’ that ‘become[s] fragile and prone to failure even under moderate distribution shifts,’ the researchers write. ‘Rather than demonstrating a true understanding of text, CoT reasoning under task transformations appears to reflect a replication of patterns learned during training.’”
• Is AI a “Normal Technology”? by Tim O’Reilly. Just skimmed it, but great to read this after writing my comments on the article up top. “That is, the tempo of technological change isn’t set by the pace of innovation but rather by the pace of adoption, which is gated by economic, social, and infrastructure factors, and by the need of humans to adapt to the changes. (In some ways, this idea echoes Stewart Brand’s notion of “pace layers.”)”
• The defense against slop and brainrot by Paul Jun. “This is the resistance training nobody talks about anymore. The more the world automates, the more value accrues to people who continue to condition their fundamental capabilities. It's like watching a gym empty out because everyone discovered protein powder—meanwhile, the few who keep showing up, deadlifting and squatting, get exponentially stronger.”

Asides

• three books dot net. Looooong time follow Michael Sippey did a bit of vibe coding with Claude and had the brilliant idea of making a cool little site that compiles the books recommended on The Ezra Klein Show.
• New York City’s vast water system is this photographer’s playground. “A new edition that puts a greater emphasis on the parts of the system that are hiding in plain sight. The reissue also includes a map by New York Times visual storyteller Larry Buchanan showing the locations of visible water equipment stretching from Long Island to the Catskills.”
• Mini Ikea stores will be opening inside select Best Buy locations this year. “‘By bringing together our home furnishing expertise, products, and services with Best Buy’s leadership in appliances and technology, we’re creating a one-stop destination where customers can design their dream kitchen, storage solutions or laundry space with ease,’ said Rob Olson, chief operation officer, Ikea US.”