Happy Shiny People
Part 2 Development, Work, the Search for Meaning in the Age of AI In part 1 , we explored Dario Amodei s most convincing predictions from Machines of Loving ...
Part 2 – Development, Work, & the Search for Meaning in the Age of AI
In part 1, we explored Dario Amodei’s most convincing predictions from Machines of Loving Grace – the health revolution aiming to be AI’s lowest friction win. In part 2, we leave the lab and step into daily life to ask what happens to society when AI becomes capable enough to handle most of our work? This is where Dario’s sunny optimism starts to meet some very messy realities about human psychology, economics, and politics.
Promises of Prosperity
To his credit, Dario tackles whether AI’s benefits will reach the developing world and how they may affect poverty and inequality. Notably he calls out that this section is fuzzier than predicting technological breakthroughs as it involves some very human constraints, like our old friends’ corruption and political complexity. Ah, the fun stuff.
However, he sees reasons for optimism. The health interventions described in part one could benefit the developing world disproportionally, providing increased human labour into a booming economy enhanced by AI decision making, all while climate change mitigation technologies reduce the burdens on poorer countries.
How much growth are we talking about? A bold target of 20% GDP growth per annum. Even if you’re bullish on AI, that’s ambitious. For context, China’s reform‑era, world-historical boom averaged around 9-10% a year. Some other East Asian economies matched this – though they have slowed once they approached the levels of the industrialised countries. It’s double the most successful economic transformation in history.
Serious economists project far smaller global AI uplifts. Recent joint Nobel winner Daron Acemoglu estimates ~1.1-1.6% total growth over ten years, while Tyler Cowen has higher but still modest numbers. They do not expect economically transformational AI on these timlines. I think these estimations are too low, but Dario’s 20% is too high – especially for the developing world. This unprecedented growth would allow sub-Saharan Africa to reach the prosperity of modern-day China inside the 5-10 years post “powerful AI“.
I am not confident that the health benefits would amount to 10% annual GDP growth. I am far less convinced that the vaguely described AI economic guidance will deliver another 10% – for both technical & human reasons. The binding constraints in development are rarely ideas or intelligence. Many developing countries face a lack of institutions, low social trust, and harsh politics.
While AI can assist with ideas and strategy, it can’t force anyone to implement them (in his vision). I don’t see how it will convince a warlord to undergo an audit, or manufacture public trust in a place that has never had any. We already know how to prevent malaria and improve crop yields, yet millions suffer and go hungry.
Polio almost became the second disease humanity eradicated after smallpox. However, the good folks in the Taliban & Boko Haram prevented vaccination programs in Afghanistan and Nigeria. Violently. Will AI be able to persuade the type of people who murder vaccination workers to change their minds and enjoy 20% growth rates instead? Better governance & social trust won’t come at the end of a chat with Claude. Our messy, tribal politics may prove to be more problematic than expected. This may shock you, but not all people in politics are in it for the social good. Never mind LLM models, humanity is not aligned.
While almost everyone will be at least on board with the idea of assisting developing nations, overall inequality will surely rise. No one is suggesting the developed nations slow down – does anyone think that the marginal returns will be greater in the developing world? I didn’t think so, though some form of global democratic access to AI could help, for example global healthcare advice would provide greater returns to those who have none today.
Inequality within countries? Messier still. Dario is more optimistic here, especially in the developed world, but I think the politics of this would take superintelligence to work out! We have enough issues sharing what we have today within countries; it’s not clear that having more to share makes it easier.
As with other global transformations, the benefits will likely go to the rich first and disproportionately to the richest economies. Superstar firms have been driving a reduction in labour’s share of GDP globally; “Powerful AI” controlled by a few (or a single) companys would push this dynamic to extremes.
Politically, there will be many failures and missteps from all parties and electorates. Driven by human psychology and tribalism, we will continue producing uneven, unfair, and unexpected outcomes within the new dynamic. Some places may stick the landing. And while almost everyone’s lives will improve on many metrics, it may not feel like that. Feelings drive politics, not intelligence. That’s maybe one thing that will be the same as today.
When the Work Stops – Meaning Beyond Productivity
Dario picks up from the story on work & meaning after the preceding four waves break over humanity perfectly. We are healthy, wealthy, and peaceful. It is the dawn of a golden age, a post-scarcity economy, a world of abundance with our core needs catered for. Some of us would be freed from danger, boredom, and drudgery. Some may be delivered into it.
What will we do all week? It’s not work because very powerful AI will be able to outcompete all humans on all intellectual tasks. While comparative advantage may keep humans relevant initially – massively improving productivity – eventually AI becomes so capable and cheap that current economic structures won’t make sense. Many manual roles would not succumb to this fate without a robotics revolution that is not discussed but is also underway.
What would I do day to day? If you’ve ever dreamed of winning the lottery, now’s the time to daydream with me. Holidays, time with family & friends, physical fitness, embarking on projects even if AI can do them better, learning even if AI already knows the answers, and lots of new experiences. Good times. But I’d need meaningful things to do (or I’d drive my wife insane).
How will society respond once necessary work is removed? Temporary joy (and parties) I’m sure, but after a few months? Years?
A lack of meaning in life, even if meaning is self-created, can be dangerous. Its presence is essential.
I’ve been thinking about this for 30 years. My mum bought me a book of 20th-century history when I was about ten. On the page for 1943, there was a two-sentence entry noting the release of Jean-Paul Sartre’s Being & Nothingness. The brief passage described the core of his existentialism – explaining there is no meaning outside what you yourself create, giving individuals total freedom but huge responsibility. I re-read and wrestled with this for years – it is the proximate cause of why I would study philosophy at university. I have settled on a universe & existence with no inherent meaning but lots of happiness. Being & Happiness if you will.
On meaning, Dario argues that people already find purpose in activities where they’re not the best, and that meaning primarily comes from human relationships rather than economic productivity. In the same vein, he notes we spend lots of time on leisure activities that are not economically valuable. We do things because we like them, for a sense of accomplishment, or the thrill of competition – even if an AI could do them better.
I would expect interest in religion, immersive video games, and sports teams to explode as people took on deeper identities outside work. This would have positive and negative effects. I would expect more gambling, drinking, and loneliness at a societal level. This despite, or because of, the general abundance as the lack of meaning for those unable to sell themselves a self-directed story suffered. Forty hours a week will be a lot of time to fill for many people.
More politics too – we are social creatures and there are several intersting theories around how the development of launguage through gossip affected humanity. No longer needeing to work full time gave space for advanced launguage & status games, we could maybe expect more not less politcs unfortunatly.
I would hope we would see more children, larger families. It’s the oldest and best way to create meaning in your life! It could contribute to the looming population crisis in the developed world (to be another blog post). Aged 40, another child is not on my menu. Though with no worries and free time would it shift my likelihood? Yeah, maybe. Although, one important counterpoint to a world without work, it’s where I met my wife. People seem to hate online dating – leaving a lot of pressure on these new leisure activities 🙂
The Fragile Transition
There are deeper questions to answer around work & meaning. Dario suggests we will require “societal conversations about new economic models”, potentially involving universal basic income (I agree this can’t be the only solution), AI-mediated resource distribution based on “human values”, or entirely novel economic structures. A few of these ideas are explored well in AI 2041 (recomeded) – especially UBI alternatives and an actual role for “bullshit jobs” – genuinely fake work to make humans feel better while AIs do the economically important stuff.
He states historical precedent showing that civilisations have navigated major economic transitions before, but that this one may require unprecedented innovation in social organisation.
Yeah, no kidding. I feel this section is ever so slightly underplaying what that would mean – in footnote 2:
“I think there’s only so much change people can handle at once, and the pace I’m describing is probably close to the limits of what society can absorb without extreme turbulence.”
Indeed. We can ask how civilisations navigated major changes in the past. Were those transitions ever painful, deadly, or explosive? If this change is as big, or bigger, than going from hunter-gathering to farming, or feudalism to capitalism – and if it happens globally in a decade instead of over centuries what would that look like? I have grave reservations about the ability of our misaligned human politics to make that work even in a dream world of AI-enhanced health & abundance.
Next: Timelines, Geopolitics & Sci‑Fi
In Part 3, we will find even bigger questions to ponder and more to disagree on, as we discuss timelines, geopolitics, and Sci‑Fi (and yes, human annihilation, where I’ll reveal my p(doom).)
P.S. – All Watched over by Machines of Loving Grace is also an exellent 3 part documentary from 2011 by Adam Curtis – https://machines.cargo.site/