jinxey

You guys are all worried about the wrong thing. Everyone is afraid of cryptography being broken when the real threat is that break being made public. Those are two very different threat scenarios, and the difference is clearly underappreciated on here. Governments have routinely kept world-breaking secrets for decades: - US and Israeli intelligence share access to a technology that lets you pull all of the data from a cell phone merely by typing in its number. It's called Pegasus, and it is still in use today. - The NSA discovered the cryptography we are talking about and used it for a decade before the public caught on independently. - The NSA and MI5 learned in the '50s that you could decode pretty much any data from electronics just by listening to them. They launched a research program to exploit this vulnerability and to defend against it, but didn't make any of that research public for decades. For those decades, the world was an open window to them. Per the above, a scenario where only US labs have this knowledge is not a huge threat. The real threat is the potential game-theory scenarios between US labs and Chinese labs. To think about that, I think these are the relevant considerations: (1) Breaking crypto would break most of the global economy, which is the government's source of income, as well as government infrastructure. It would be incredibly detrimental to government. (2) Because of 1, labs that are de facto government-controlled will strongly prefer not to release a way to break crypto without also releasing a way to fix it. Add these up and the real threat is not the existence of the knowledge but the game-theoretic conditions under which it's optimal for one lab to release the break before the solution. This would not happen between two US labs, which are ultimately restrained by the USG. It might happen with a Chinese lab, however, if the Chinese feel such a break would weaken them far less than their adversary: kicking themselves in the gut while kicking their opponent in the teeth. I'm frankly not knowledgeable enough to know who gets hurt worse in that scenario, but I think people aren't taking seriously the likelihood that governments will have a gentlemen's agreement not to break the world unless some greater conflict escalates and forces them into it. It's essentially the logic of nuclear weapons. This is not a threat for the next few months, but for the next few years. Moreover, if one party would be hurt much worse by cryptography breaking, you can expect that party to throw massive resources behind the scenes at finding a solution as quickly as possible. So with every passing month, at least one party will be getting closer to a solution that lets it defang its opponent. The real story here is the concentration of power within governments. AI labs have more and more negotiating power with the US government. If they hold all these secrets, and have the ability to produce more, you can expect that they would not release them, but instead start to exert more and more soft power within the US government. They will accrue enormous political power and conversely the US government will gain enormous power through them. The next chapter has nothing to do with crypto. It will be an all-out assault on open-source AI because open source is the only source of this breakthrough that would threaten existing power structures. At the end of the day, the fact that cryptocurrency shares its primitives with most of the global economy makes it a much less significant part of a much bigger power struggle.