Use your rationality or it’s useless
Infinite rationality is worthless if you never act. An omniscient entity which can never act has a net progress score of 0.
Likewise, an entity with totally random actions is going to fare poorly. Given that the possibility-space of state improving actions is much less than half, I see it as in fact likely the chaotic entity will worsen its situation over time.
If progress requires both rationality and action We are left with 3 possibilities.
An entity has disproportionately better decision-making then level of activity, and should focus its energy on using its knowledge.
An entity is disproportionately more active than rational, and would benefit from becoming more rational instead of continuing to be busy, or
An entity is somehow balanced, in which case good for them.
Looking at LessWrong, I’ll give you three guesses which most of us are..
So for the people who read LessWrong, myself included, it would almost be more helpful to have a LessPassive and let our tendency towards epistemological optimisation play out on its own.
We need to increase our level of action. The upvotes on actually “Making Vaccine” are our belated acknowledgement of that.
The problem is arguably making a DIY vaccine is less rational, if rational is defined by “optimizing our personal survival and wellbeing”. The best we can do presently seems to be:
a. Work in tech for the high personal income and decent job satisfaction
b. Work in tech on AI to nudge the needle to the extent we can as individuals (and also increase income and job satisfaction)
c. Make the decisions we can make—what college to accept, what car to buy, what job offer to accept—based on quantifiable, expressible metrics that factor in probabilities of failure
d. Sign up for cryonics, use any other method of life extension that has credible evidence behind it
That’s all we can do. We individually are motes of dust, it took the collective efforts of millions of people to develop the infrastructure where rational thought and discussions are even sorta possible. We need to invent artificial intelligence to determine what the actual unbiased, rational decision for a complex real world decision even is. I would say that AlphaGo Zero and other efforts, where an AI agent trained with no starting human priors learned to outperform humans, is one of the first proof of concepts that shows such a machine is even practical, and that was 2017! (and it only tells you the approximate rational answer in the restricted game of Go, we do not yet have an AI that can tell you the right legislation to pass to say minimize childhood hunger or some other agreeable worthy cause)
The vaccines that were made by Big Pharma had the collective infrastructure and efforts of thousands of people who specialized in related subfields their entire lives. Unless we have a tool like an AI agent available to us that we can expect to outperform the experts (and if we were playing against every Go player in all of Asia we now could) we can’t do better. Yet.