Yesterday when I wrote about the new Chinese-produced LLM, I started thinking about an earlier topic from the All-In Podcast. In it, they talked about quantum computing figuring out the security that safeguards cryptocurrency, specifically Bitcoin.
It is hypothetical at this point, but it looks like quantum could crack the security code very quickly.
The answer would be to create a new level of security that quantum couldn’t break. Except, quantum will get better and better and so the security will have to be updated consistently.
Crypto is an open network. Anyone can enter and there are limited costs to entry along with limited costs to exit. It’s a free market. Most LLMs are that way too. The new Chinese DeepRake is an open network. AI needs openness because the incentive is to suck up as much data as it can. More data means better decisions.
Our fiat currency system has problems to be sure. But, it is a closed centralized network. The Federal Reserve controls it. There are barriers to entry, transfer, and exit. Of course, those barriers and transfers allow gatekeepers to set up shop and charge outsize rents for the services they provide. The system has fraud in it as well which stinks since every time I log in to any financial provider I have to do two-factor notifications and maybe other keystrokes. It becomes very hard to wire money or do other things.
But, is paying the rent to the gatekeepers worth it compared to the costs of running an open network in a quantum computing world?
I don’t know the answer to that question.
Calling the new Chynese “AI” product “DeepRake” vs “DeepSeek” may be the best Spoonerism yet, as it certainly seems, at second glance, a voluntary spyware system.
The “spread” trades off of this bubble (can we say that yet?) are quite the example of “Overdone”. [The “reports” this AM from GS & JPM on DeepSeek and future chip spend sure seem defensive of the original bet on “AI”.]
I’m thinking maybe I should learn to code