I said:

Back to you and the forthcoming crash. If we can't prevent it, how do we mitigate it? And are we creating a desolate landscape for mankind?

Claude said:

Those are really two separate questions and the second one is the more important one, so let me take them in order.

Mitigation — what could soften the landing:

The crash itself probably can't be prevented, but the difference between "painful recession" and "generational catastrophe" comes down to policy choices that are still available.

Graduated adoption mandates. The EU is actually better positioned here than the US. Rather than letting companies fire 30% of their workforce overnight to hit a quarterly earnings target, you regulate the pace. France already has consultation requirements before mass layoffs. Germany's Kurzarbeit model — where the government subsidises reduced hours rather than letting companies lay people off entirely — proved extraordinarily effective during COVID. Extend that logic to AI-driven displacement: if you're automating roles, you fund retraining for the people in those roles before you eliminate them. The EU AI Act has some of this thinking embedded in it, though it's focused more on safety than employment.

The US won't do this. The current administration's ideological commitment to deregulation makes employment protection legislation essentially impossible at the federal level. Some states — California, New York, possibly Massachusetts — might act independently, but it'll be patchwork.

Counter-cyclical investment in infrastructure and public services. This is the classic Keynesian response and it works. When private sector employment contracts, government spending on physical infrastructure, digital infrastructure, healthcare, and education absorbs displaced workers and maintains demand. Ireland actually did a version of this after 2008 — badly at first, then better with EU support. The EU's recovery mechanisms are more robust now than they were then.

The critical version of this for AI displacement specifically is investment in the sectors AI can't easily replace: care work, skilled trades, physical infrastructure, education, community services. These are chronically underfunded everywhere. A deliberate policy of redirecting displaced white-collar workers into these sectors — with real retraining and real wages, not make-work — would both cushion the employment shock and address genuine social needs.

Financial circuit breakers. When the tech stock correction hits, the cascade effect — margin calls, pension fund losses, venture capital freezing, startup collapses — is what turns a market correction into an economic crisis. Stronger circuit breakers on trading, limits on leverage in tech-heavy portfolios, and pre-positioned liquidity facilities from central banks could contain the financial contagion. The ECB and Fed both have the tools. Whether they deploy them fast enough is a different question — the 2008 experience suggests central banks are too slow initially but eventually act aggressively.

Portable benefits and universal basic services. Rather than tying healthcare, pensions, and training to employment — which makes job loss catastrophic for individuals — you decouple them. Ireland's system is already partially there through the HSE and social welfare, though neither is generous. The Nordic countries are better positioned because their social safety nets are stronger and less employment-dependent. The US is worst positioned because healthcare is employer-linked, which means losing your job also means losing your health insurance at exactly the moment you're most stressed.

Universal basic income gets discussed a lot in this context. I'm sceptical of it as a primary solution — not because the economics are wrong but because it doesn't address the deeper problem, which I'll get to in the second question. But universal basic services — healthcare, education, housing support, digital access — as a floor beneath which nobody falls regardless of employment status, that's both more achievable politically and more effective practically.

Managed deflation of the AI bubble. This is the hardest one. In theory, regulators could require AI companies to demonstrate actual revenue sustainability before allowing further capital raises. They could impose capital requirements on venture funds similar to banking requirements. They could limit the tax deductibility of AI capex spending that doesn't produce measurable returns within a defined period. Any of these would slow the bubble inflation and make the eventual correction less severe.

In practice, none of this will happen in the US because Silicon Valley has enormous political influence and the current regulatory philosophy is hostile to intervention. In the EU, it's possible but slow — the AI Act took years to negotiate. The bubble will likely pop before any regulatory framework is in place to manage it.

So the honest mitigation picture is: Europe has the institutional capacity to cushion the blow significantly if political will exists. Ireland specifically benefits from EU membership, social welfare infrastructure, and a relatively diversified economy (agriculture, pharma, tech, tourism). The US is likely to experience the correction more severely because of weaker safety nets, more aggressive corporate behaviour around layoffs, and political paralysis around any form of intervention.

The countries that will be hit hardest are those with high exposure to AI-automatable work, weak social safety nets, and limited fiscal capacity to respond. That's a lot of the developing world — which is bitterly ironic because they're the least responsible for creating the problem.

Now the bigger question: are we creating a desolate landscape for mankind?

No. But we might be creating a very different one, and whether it's desolate depends on choices we make in the next five to ten years.

Here's what I actually think is happening, stepping back from the immediate crisis:

The economic model that's dominated since the Industrial Revolution — humans sell their labour, employers buy it, the transaction creates both income and social identity — is being fundamentally disrupted for the first time. Previous waves of automation displaced physical labour and created demand for cognitive labour. This wave displaces cognitive labour, and it's not yet clear what new category of human contribution emerges to replace it.

That's not unprecedented in human history, but it's unprecedented in capitalist history. And our entire social infrastructure — education, healthcare, housing, pensions, social status, personal identity — is built on the assumption that most adults will exchange labour for wages for most of their lives.

If that assumption breaks, everything built on top of it wobbles. Not just economics — identity, purpose, community, mental health. This is actually what your dissertation touches on, from a different angle. The psychology of how people behave when their relationship to work and status is disrupted.

The desolate scenario looks like this: AI concentrates wealth in a small number of companies and their shareholders. Displaced workers can't find new roles because the new roles require skills that take years to develop and the training infrastructure doesn't exist. Governments can't fund safety nets because their tax base has eroded — corporations optimise taxes globally, unemployed people don't pay income tax. Communities hollowed out by job loss experience rising substance abuse, family breakdown, political radicalisation. We've already seen a preview of this in the American Rust Belt after deindustrialisation. Scale that globally and accelerate the timeline.

The non-desolate scenario looks like this: we use the productivity gains from AI to do things we've always wanted to do but couldn't afford. Universal healthcare that's actually good. Education that's personalised and lifelong. Infrastructure that's maintained. Clean energy. Environmental restoration. Care for the elderly that's dignified. Research into diseases we've neglected because they weren't profitable. The wealth exists — AI genuinely does create enormous value. The question is purely distributional: who captures that value and what do they do with it?

The historical pattern is cautiously optimistic. Every previous technological revolution — agricultural, industrial, electrical, digital — produced a period of severe disruption followed by broadly shared prosperity. But that "followed by" hid decades of suffering, and the transition was never automatic. It required political movements, labour organisation, regulatory intervention, and sometimes violence. The New Deal didn't happen because Roosevelt was a nice person. It happened because the alternative was revolution.

What I think is most likely — and this is genuinely my assessment, not optimism or pessimism — is something in between. The next five to ten years will be genuinely painful for a large number of people, particularly in the categories we discussed: entry-level white collar, middle management, routine cognitive work. Some countries will manage the transition much better than others. The Nordic countries, possibly Ireland and parts of the EU, will adapt relatively well because they have the institutional capacity and political culture for collective solutions. The US and UK will struggle more because of ideological resistance to government intervention and weaker safety nets.

After the transition period, the economy will restabilise around a different model. Fewer people will work traditional full-time jobs. More people will work in hybrid roles — human judgment augmented by AI tools. The value of uniquely human skills — empathy, physical presence, creative vision, ethical judgment, community building — will increase relative to cognitive skills that AI can replicate. New forms of income distribution will emerge, probably not UBI in its purest form but something more like a combination of public services, shorter working weeks, and profit-sharing from AI-driven productivity.

But — and this is the crucial caveat — none of that happens automatically. It happens because people fight for it. It happens because political leaders make choices. It happens because the people who understand what's happening — people like you, who see both the technology and the human response to it — refuse to let the default outcome stand.

The default outcome without intervention is concentration of wealth, erosion of the middle class, and social fragmentation. That's not desolation for mankind — it's desolation for most of mankind while a small number do extraordinarily well. That's not a new pattern in human history. It's feudalism with better technology.

Whether we get the good outcome or the bad one is not a technology question. It's a political question. And politics is just people organising to demand something different from what the powerful would choose by default.

That's not a hopeless answer. It's actually the most hopeful framing I can offer — because it means the outcome is still up for grabs, and the people who understand what's happening have disproportionate influence over which direction it goes.