Ideas for the newly rich
With the AI labs about to IPO, we’ll soon have a strange new class of tech ~billionaires, many of whom feel like they got rich somewhat incidentally and don’t have much ego tied to it. It's a chance for philanthropic projects that would never have worked before, whether they fund them or pursue them directly. Here are some ideas that I’m personally excited about:
Hotel-like sleep lab
Sleep is the most important thing for health, a huge fraction of people these days have sleep issues, and sleep labs today are terrible:
- It’s nigh impossible to book multiple nights, which is crazy because single-night results are corrupted by the first-night effect given how unusual the lab sleeping conditions are
- Results inexplicably take weeks to arrive
- You need a referral from a doctor, which serves no meaningful filtering purpose
- It's hard to know what to do with the results, since you don't know how your stats compare to others like you
It would be great to have a super comfortable sleep lab that lets you book online, no doctor in the loop, allows you to book multiple nights in a row, and sends you your results immediately so you can have AI agents analyze them. There could be multiple tiers of fanciness, the top one nice enough to be a third date. Crucially, the lab’s data should be public, so anyone can compare their results against the full distribution of people who slept in the same conditions. Staying anonymous could be allowed, though discouraged; publishing the data itself would be non-negotiable.
A reason this is hard to make viable as a business today is insurance. Insurance companies are not happy with skipping doctor referrals, booking multiple nights, or excess comfort. So this business would need to eschew insurance entirely, which usually means serving only the rich, but the gap can be filled with philanthropic support. It seems like with AI-assisted monitoring, costs could go down to ~$500 a night in SF before losing money.
You could then add other biodata services like blood panels, MRIs, and genome sequencing into the same visit, turning the lab into a single place to track your whole biological profile over time.
Public data hospital
A more brazen generalization of the public data sleep lab—a hospital that publicizes all of its data by default: every scan, every diagnosis, every doctor recommendation, every declined patient, every mistake. Protocols on GitHub. Ideally even livestreamed video of every interaction.
Patients can request their data be censored, but they’re encouraged and maybe even incentivized to let it be public. This would fix a lot of the information asymmetry problems in healthcare and incentivize doctors to do what’s actually right, since eventually AI will be able to retroactively look at their decisions and evaluate whether they were a good doctor. Ideally, doctor referrals should not be required; video applications would be a better filter.
Of course the privacy issue is a non-starter for many types of ailments, but that’s okay; this hospital doesn’t need to handle everything, it just does what it can. Hopefully this can also help normalize sharing health problems.
Many doctors would understandably be terrified to work in this environment. If they make a fatal mistake, video will be online forever. But the video may save many more patients, so rather than hiding such mistakes, we should make the reward worth the risk. A seven figure base salary would help. But more importantly, the best doctors should be able to thrive on the credibility of their public track record here, which is far more meaningful to a sophisticated customer than a Harvard diploma could ever be.
Incentive rating agency
Most of the harm in the world comes not from bad people but from business models that reward negative-sum behavior. A rating agency that deeply understands good incentive design could become a trusted authority on evaluating any type of business with information asymmetry problems. It could rate a wide range of institutions—hospitals, schools, restaurants, mortgage brokers, startups, governments—or focus on one type. It would field anonymous investigators, some posing as customers, some getting hired, some working sources. They'd dig into promotion criteria, outcome accountability, pricing transparency, dark patterns like making it hard for customers to leave, and so on.
The meta challenge is ensuring the rating agency itself retains good incentives. It helps to have radical transparency, to be funded by vetted donors rather than the rated agencies themselves, and maybe to have red teams that seek inaccurate ratings, but none of these are complete solutions. I think it ultimately rests on the founders’ willpower to keep enforcing purity top-down.
AGI short films
People tend to believe the fate of the world is predetermined by the forces of capitalism or something, but it isn’t; public opinion is highly swayable, leaders are swayed by sufficiently strong public opinion, and a shared concrete vision can go far. The world is still lacking good drama that helps people understand AGI more than it misleads. A feature-length movie would be cooler than a short, but it’d probably take so long it’d be outdated before release. Money can't manufacture a unicorn person who both understands AGI and can write great drama, but it can build a system that brings the two halves together. The ideal structure for this may be something like a residency program for screenwriters paired with alignment researchers, backed with access to famous producers.
There is of course the PR issue that no matter what narrative is presented, many will see these films as propaganda to boost lab equity value. I’m not sure whether this can be eliminated, but we can push back with Sloan-style independence, published funding, and a mandate to greenlight scenes that make labs look bad (e.g. regulatory capture, cutting corners under competitive pressure).
There are of course a lot of reasons why none of these things exist, and they’re not easily solved by having new smart billionaires. Many of them brush against deeply resistant industries and lead to countless lawsuits. They’re socially weird and controversial; they are not the easy social status boosts that most of philanthropy is.
The reason I feel like there’s hope here is because many of these newly minted billionaires are weird. They don’t share the same social circles that past billionaires have had. Many inherit more of their temperament from academia than from business, many have frequented LessWrong, many apply social pressure on each other to be altruistic.
Maybe more importantly, many young people who once dreamed of starting a startup are now finding that traditional startups are losing their meaning. AGI timelines are a major reason, but not the only one. In general, as startups have become socially legitimized, they no longer feel like the route for the crazy ones, the misfits, the rebels. They're now simply a standard career path. The unconventional-minded need a new frontier to direct their ambition at, and impossible nonprofits may be it.