How to prepare yourself for AGI

What if artificial general intelligence (AGI) arrives in just a couple of years, triggering an explosion in science and technology that transforms life as we know it?

Whether you find this timeline plausible or not, it’s worth thinking through: what can an ordinary person do to prepare?

It’s a scary situation, because one possibility is we all die. In that case, all you can do is try to enjoy your remaining years.

But let’s suppose we don’t die. How can you maximise your chances of surviving the transition and flourishing in whatever happens after?

In many scenarios, the world will be so radically transformed that any efforts to prepare will be rendered moot. But there’s no point worrying about these scenarios. Focus on the scenarios where you can influence the outcome.

One approach is to try to make the transition go well for society. That’s what I write about most of the time (e.g. see here and here), and given the insane stakes, should be where we focus. But this post isn’t about that; it’s about how you can personally protect yourself.

Here are the best ideas I’ve heard so far:

1. Seek out people who have some clue what’s going on. During COVID, life was suddenly upended, and every week brought confusing, high-stakes developments. Now imagine that, but lasting for a decade. And nothing goes back to how it was. In COVID, it was very helpful to know people who were ahead of the curve, and could provide reasonable advice under uncertainty. It was also very helpful to have your own basic model of what’s going on. Find the same but for AI. Some of the people and sources I’ve found most prescient and useful are Carl Shulman, Paul Christiano, Leopold Aschenbrenner, Ajeya Cotra, Daniel Kokotajlo, The 80,000 Hours Podcast, the Dwarkesh Podcast, Zvi Mowshowitz, The Cognitive Revolution, and Epoch AI, though it’s also useful to have people you personally know and trust for rapid updates. Join a team that’s plugged in and be useful, if you can. In the meantime, learn as much as you can about what’s happening. Here’s a starting point. Use the most advanced models so you develop intuitions for what they can and can’t do.

2. Save money. AGI probably causes wages to increase initially, but it’s possible they later collapse. Once AI models can deploy energy and other capital more efficiently to do useful things, there’s no reason to employ most humans anymore. At that point, you have to live on basic income provided by the state, plus whatever you’ve saved for the rest of your (maybe longer) life. Money also has option value, which is helpful in navigating a radically uncertain future, and investment returns could be very high. Most likely everyone is going to become rich by today’s standards. Many goods and services will become extremely cheap. The government today taxes about a third of GDP, and if GDP is 100x higher, then welfare would likely also grow around 100x too. And redistribution would become the number one political issue (so it seems likely, if not guaranteed). And you might decide you don’t care whether you have $1 billion or $10 billion. But if any goods remain scarce (like land, computing power, or political influence), it could still matter. The good news is that you have one last chance in history to make bank in the upcoming boom…

3. Invest the money in things that do well in the lead up to AGI. (Or at least that don’t do terribly.) I can’t discuss the specific investment implications, am not a financial advisor, and this is not financial advice, but it seems possible to do a lot better than a standard 60:40 portfolio. Here’s some other writing on the topic (which I don’t fully agree with). If you have a safe way to take long-dated fixed rate debt, consider that it’ll be cheap to pay back if there’s been a lot of growth (more here).

4. Learn skills that are likely to increase in value in the lead up to AGI. Human labour remains a bottleneck until AI can do basically *everything* a human can do and has been fully deployed. Until then, the skills AI can’t do that are complementary to AI deployment increase in value. Small teams will be able to do what would have taken hundreds of people in the past. I’ve written a separate guide to what those skills are, but in brief: applying AI to solve real problems; personal effectiveness; social skills; learning how to learn; entrepreneurship, management, and strategy; expertise in complementary fields like ML, robotics, AI hardware, and cybersecurity; communications; getting things done in government; and complex physical skills, such as data centre or robotics construction.

5. Get citizenship of a country that’ll have lots of AI wealth. The US seems by far the best positioned country for AGI, so is the best place to be a citizen. But I expect its close allies, and especially those with parts of the AI supply chain or significant military power, get cut into a deal with the US to share the benefits. This might include the EU (especially the Netherlands due to ASML), UK, Japan, South Korea, Australia, and Canada. To be clear, I wouldn’t be surprised if the gap between the US and the rest increased, but I also don’t expect them to be totally left behind. China is also in a good position, especially if AGI comes later in the 2030s, but naturalization is extremely rare in practice. The rest of the world will also still benefit from scientific innovation, the cost of superhuman AI services being reduced to near zero, far cheaper robot-produced goods, taxing AI models deployed in their country, and having energy & raw materials that are needed in the supply chain.

6. Make yourself more resilient to crazy times. I think going ‘off grid’ is overrated as a way to protect yourself from AGI risk (as with going to Mars, it won’t protect you). However, I expect an intelligence explosion will destabilise the world order, maybe leading to threats of nukes, bioweapons, cyberattacks, or newly invented weapons of mass destruction, maybe through the 2030s. So I think there’s some benefit of making a plan to go somewhere outside of major cities with several months of supplies, where you can at least feel less anxious, or in the worst case, wait out a pandemic, internet shutoff, or fallout cloud. In COVID, it was very helpful to be part of a group, so seek out people you could live with and who are willing to take non-common-sense actions. Investing in your mental health, groundedness, and ability to deal with stress over long periods could also be worth it. That might include trying to help.

7. Prioritise things you want to be in place before AGI. The intelligence explosion will likely either start in the next six years, or take much longer. And we’ll know a lot more in just three. So delay anything you can delay three years, and focus only on things you’d want to do before an explosion starts. For example, if you’d like to live in both the US and Argentina at some point in your life, do the US now. If the intelligence explosion starts, you’re already three years into getting citizenship. If it doesn’t, you can still tango in Argentina in the 2030s. More controversially, if it’s not a big cost to you, it could make sense to delay having children by three years, and have them later when uncertainty is reduced. Likewise, if medical technology will be a lot better in 10–20 years, then it becomes more important to try to stay alive until then (e.g. avoiding dangerous hobbies like scuba diving), but you can worry less about health problems that only bite 20+ years in the future. Whether to focus on enjoying ticking off your bucket list the next few years or going all out trying to help make the transition to go well seems a difficult personal decision, depending on the likelihood of different scenarios, your altruism and ambition, and what options are open to you.

You might think this all sounds far fetched, and of course AGI might still be decades away. That means it’s important to look for courses of action that are also fine if timelines turn out to be much longer. I think the actions above mostly are, provided you don’t take them to the extreme. (While spending all your money on a final three-year binge is not.)

You still have some years left – so use them to prepare as best you can. (And if you want to help society as a whole navigate the transition, see here.)

I’m keen to hear more ideas in the comments. I also hope to write more about them, so subscribe to get notified.