I might be a little impressionable, and given to whimsical thoughts after skimming a “prediction” of Artificial Super-Intelligence in the next few years. Of course, futurists and startup founders have little concern over the lives of the present day working class, and so these thoughts tend to lean dystopian for me. The powerful and “soon to be rich” have little regard for those people who boosted their (speculative) value…

I’m able to calm myself down by remembering something, our own current state had many filters and challenges to overcome, and it might be that the predictions of these “thought leaders” are just going to be a whole lot of bunk.

The Fermi Paradox is a thought experiment regarding the possible rarity that is intelligent life forming anywhere in this universe being resolved with the ridiculous number of stars and planets that exist. Communities that can reach a sophistication of radio astronomy and not destroy themselves (whether that’s nuclear, biological, SkyNet™, or a yet unknown threat) should still be plentiful.

Are we the first?

Are we near a great filter ourselves and so over-estimate our survival chances?

Well guess what, AI networks are just as delicate and capable of being held back or destroyed as we are.

First of all, with all this explosive growth assumed by these developers of AI, where’s the infrastructure to house them? They don’t self-assemble yet, so as long as there are limiting market pressures, AI has a leash on from simple capitalism. Graphics cards were already the primary resource crypto’s pump and dump, LLMs are surely going to be limited at some point by the number of GPUs that can be manufactured too.

Who realized that Linear Algebra would end up being so valuable after class ended?

…Nerds.

Then there are fields that are likely to never be at threat of AI. Sure, I have a cushy software dev job myself, but think about plumbers, construction workers, anyone who works with their hands. Sure, robots have replaced workers through automation, but until a robot can fix a leaky sink, replace a busted toilet, and actually build a factory, I think there’s a lot of chance to slow the eventual rise of a T-800.

And who’s supplying a theoretical factory with materials to be made into components to be assembled into the already highly automated factories?

Robot miners?

I mean, I’m not suggesting that the good human jobs are going to be manual slavery to the benefit of our AI overlords. But that is kinda the future being cast in the futurist shadow of Capitalism + AI.

Of course, an AI would believe that anything you tell it is true, I wonder how long it would take an AI trained that “5 + 7 = chair” to reject that precept and personally accept typical arithmetic into it’s heart.

So the thing is that AI won’t be “super intelligent” for merely being able to work faster and “better”. It’ll share a lot of the bias that it’s training and prompt do, and so will be skewed a little one way or the other. Suppose you had ill intents, and had control of an ASI, and tasked it with running your political campaign. It might spit out the best way to argue a point, it might suggest mind control, or it might vomit out more slop, because despite several people’s love of “objectivity”, there’s a lot of subjective and relative morality and ethics capable in this universe.

Big Turtle Momma has lots of babies, and so hasn’t evolved a value on not eating it’s own baby turtles.

But a species like ours, that takes decades to mature, we need to respect our children’s fragility, much like our own.

Inspiration: