3 ms·

> And yet, no organization has come to dominate the world. Why would an AI? I think the general point is valid (that AI is an intelligent partner rather than S

> And yet, no organization has come to dominate the world. Why would an AI? I think the general point is valid (that AI is an intelligent partner rather than Skynet) but organisations do dominate the world, especially the parts of it (including older organisations) that don't employ smart people or newer technologies. There is a view that organisations are a form of slow AI executing complex business / legal processes (traditionally based on people power).

sshine 3y ago
> There is a view that organisations are a form of slow AI executing complex business Multi-national companies do act a lot like hiveminds. There are individual human beings directing at the top, but only because they comply with the hive's interests. Boards act as an averaging mechanism that seeks to maximize self-preservation and growth. The anti-capitalist incentive to not understand why a company would want to grow beyond its positive impact on society think that these companies exist to benefit society. But once something has got big enough to become self-sustaining, and not depending on individual patches of land, it's only natural to see this pattern repeat. I don't really see what's artificial about it, though. Thousands of human beings clicking away on interconnected computers just seems pretty intelligent. Adding LLMs to the mix seems like a way to boost the people who already have a function in an organisation without having to scale with people; for any task where LLMs can make a person better, your organisation will grow in proportion.
- rbanffy 3y ago
  > Multi-national companies do act a lot like hiveminds. Corporations are AIs. They are just much slower at making decisions than single humans. They also limit competition against themselves by consuming human time. We are at the stage where we give an AI a limited, carefully described, goal and it fulfills it based on patterns it learned from human codified knowledge. At some point, not that far from now, there will be AIs that are capable of selecting goals based on more generic directives and optimize from there better than any human could do. At that point, where motivations become incomprehensible to humans, we can say we have something like a superintelligence. It'll be interesting.

sshine 3y ago

> There is a view that organisations are a form of slow AI executing complex business Multi-national companies do act a lot like hiveminds. There are individual human beings directing at the top, but only because they comply with the hive's interests. Boards act as an averaging mechanism that seeks to maximize self-preservation and growth. The anti-capitalist incentive to not understand why a company would want to grow beyond its positive impact on society think that these companies exist to benefit society. But once something has got big enough to become self-sustaining, and not depending on individual patches of land, it's only natural to see this pattern repeat. I don't really see what's artificial about it, though. Thousands of human beings clicking away on interconnected computers just seems pretty intelligent. Adding LLMs to the mix seems like a way to boost the people who already have a function in an organisation without having to scale with people; for any task where LLMs can make a person better, your organisation will grow in proportion.

rbanffy 3y ago
> Multi-national companies do act a lot like hiveminds. Corporations are AIs. They are just much slower at making decisions than single humans. They also limit competition against themselves by consuming human time. We are at the stage where we give an AI a limited, carefully described, goal and it fulfills it based on patterns it learned from human codified knowledge. At some point, not that far from now, there will be AIs that are capable of selecting goals based on more generic directives and optimize from there better than any human could do. At that point, where motivations become incomprehensible to humans, we can say we have something like a superintelligence. It'll be interesting.