Whether AI leads to a catastrophic 40% unemployment rate or a desirable three-day workweek is fundamentally the same in terms of total hours worked. The outcome depends entirely on policy and wealth distribution choices, such as creating more public holidays or an 'AI dividend,' rather than the technology's inherent effect.

The confident belief that AI's impact on jobs will "just work out" is dangerously naive. A more responsible approach, advocated by groups like Windfall Trust, is to use scenario planning. Just as governments plan for pandemics or cyber attacks despite their uncertainty, we must plan for worst-case economic outcomes from AI.

The discourse on AI is overly focused on preventing harms like existential risk. A more productive approach is to also define a public agenda for what we want AI to *achieve*—the public 'goods' it can create, such as solving orphan diseases or simplifying government services.

Calls to regulate AI based on speculative futures like Artificial General Intelligence (AGI) are a flawed basis for policy. These predictions have a poor track record and are often self-serving arguments used by incumbents to justify regulations that entrench their market position today.

As AI makes the future radically unpredictable, the traditional human calculus for decision-making will change. Instead of optimizing for probable outcomes based on risk, people will shift to minimizing potential regret, a fundamentally different psychological framework for navigating an uncertain world.

Citing historical failures like David Ricardo's on automation, individual AGI forecasts are deemed useless. A better approach is to model potential scenarios (e.g., labor share collapses) and then identify the crucial, currently missing data (like consumer demand elasticities) needed to determine which scenario is likely.

The most likely future is a "weird" state we can't easily classify as good or bad. Rather than comparing today to a hypothetical endpoint, we should focus on evaluating the desirability of the path, or trajectory, we are on.

The AI 2027 team prioritized creating a highly detailed scenario, even if specific predictions were individually unlikely. This trade-off is valuable because it allows for a more thorough 'gaming out' of possibilities, providing specific hypotheses that can be debated and built upon, which is more useful than vague, high-level statements.

Instead of betting on a single AI timeline, plan your career across three plausible scenarios: 1) A short, fast-takeoff scenario where AI automates R&D by ~2027. 2) A medium timeline where this takes until the 2030s. 3) A long timeline where the current paradigm plateaus. This portfolio approach makes career strategy more robust.