21.11: Notes on Yuval Noah Harari
Any system that demands perfection, won’t work.
Adapt — we don’t know which skill will be useful in 10 years.
Know thyself — for the first time, algorithms could potentially know us better than ourselves.
- This was not a problem in the XV century, because an external agent could not know more than we did about ourselves — no matter how little we did know. Today, however, an algorithm can potentially know us better than we do.
- We trust algorithms to tell us what to watch next, who to be friends with, which mode of transport to pick, what music to listen to…
- The tipping point for algorithm-driven products won’t come when these products make perfect decisions, rather when they make just better decisions than we do. The bar is not high.
- For the first time, the combination of biology, data, and computing power makes it possible.
- Desire vs. truth — we fool ourselves and say we want something yet do the opposite.
- 🤏 Tip: know you better, through meditation for example.
To hack humans at scale, two things are required:
- First, a lot of data. This route was previously blocked because you couldn’t get countless “KGB agents” wandering around writing down reports on each citizen. Second, a mechanism to analyze all this data. Once the data was gathered, you needed clerks to make sense of those reports.
- Nowadays, the first is called “smartphone”, the second machine learning and AI.
A society built on top of incompatible ideas rooted into the 18th century — assumptions like “the customer is always right” or “voters know best”.
- Our society is founded on top of the assumption that our “heart’s feelings” are the ultimate authority. But what happens when these feelings are driven by an external agent?
- Free will — human choices are not this sacred arena that can’t be hacked. This has never been true, and before nobody had the technology to hack human beings at scale. Nowadays, corporations — sitting on top of a vast corpus of data — can predict and manipulate our will (see the previous point).
- We invented these illusions we all agree upon money, nations, religion… they are not bad per se. They are “good” if they serve us and have a positive outcome, but we have to discard them the moment we become their slaves.
The best way to unite people is to find a common enemy.
For the first time we face the problem of becoming irrelevant — which is worse than slavery. Because under slavery, you still required somebody to do the work. But irrelevant means you are no longer needed.
- Maybe philosophers will become much valuable than engineers.
💡 Scary idea: ubiquitous AI that monitors everything and creates a real-time profile of you that can be used as a resume, grades for education…
- This can even be magnified with COVID and classes taken over Zoom. Problems Yuval thinks are relevant: nuclear warfare, technological disruption (AI), climate change.
- They need global solutions, there is no way we can solve this problem but in a global scope.
UBI has a problem with the ‘U’ and with the ‘B’, the I is fine.
For major impact, join an organization — fifty people working together towards a single goal will have way more impact than fifty scattered individuals.