The closer I start working with and on AI stuff, the more I start seeing the disconnect between doomsday predictions of what AI will replace vs. what it is actually capable of. Yes, it can do stuff, and yes, it's getting better. But the closer you look the more clear it becomes that the enthusiast vision of completely independent AI systems is unrealistic as of today. Yes, all the tech companies are pushing for exactly this, but reliability and accuracy is all over the place. Plus, many of the technologies that everyone is talking about in the wider public (e.g. image recognition) have been quite well-developed and widely applied for years, before the current LLM boom, but they now get more attention as part of the overall AI hype.
Many indistries are changing, but in most cases the new tools will be more akin to cars that still need drivers, rather than robots who take over the whole job. Yes, jobs might be lost, or shifted to others, but it's not like suddenly 90% of people will have nothing to do. There were similar shifts in the past with new technologies, and we made it past them.
"ultimately accountable for the success or failure of a specific project, initiative, or activity"
I think that role should be reserved for a human, who can then use all the agents they like but has to take accountability for what is ultimately delivered.
You know, as an IC, if I get get the level of introspection you can get from building distributed systems, and that with reasoning/thinking traces from LLMs, I think I'd prefer the entire level of middle management made out of LLMs rather than humans. It'd be helpful to be able to see exactly where their logic suddenly took a skip out the window.
Are their any handicap for measuring Baboon middle management capabilities and capacities to account for the unique challenges of interpersonal challenges Baboons face with each other vice homo sapiens? I imagine smiling and showing ones teeth can have unintended hurdles if not accounted for.
What showing teeth does depends on their body language. I don't know
"The famous primatologist Robert Sapolsky has spent decades studying baboons and has described how status, alliances, conflict resolution, and coalition-building resemble politics inside human organizations"
Baboons are a good example for studying mammal social management strategies, which humans also do.
They are organized by dominance hierrarchy, like humans, but baboons have a distributed leadership.
I think we could learn a lot from baboons when it comes to management
The author is completely right about the AI Lab's promised vision of the world: They claim to want to create superhuman intelligence, which will produce vast abundance. But superhuman intelligence would be extremely dangerous, so it needs to be controlled by a tiny "priesthood" of trusted people, or somehow designed so that the superhuman intelligence could be trusted. (We have no idea how to do that.)
But the author's vision is also suspect, if you assume that the models will become much more intelligent:
1. Hypothetically, we can't give every human their own personal SkyNet to command. That would, uh, probably end very badly. If everyone gets an agent, those agents can't be too capable?
2. If you do somehow build a model that's much smarter than you, what do you contribute by managing it? How many people here have ever worked for a well-intentioned manager who couldn't understand the people they managed? So in this scenario, human management would be mostly displaced by agent management. Most companies could lay almost everyone off and let the agents manage each other. We only need humans to manage models now because the models are still pretty broken.
3. If we create models that can genuinely replace humans at almost any task, you won't be able to buy those on the API. At that point, the billionaires and the politicians wouldn't need human workers any more, because everything can be done better using their pet agents. Just have the robots build stuff for the billionaires directly. And if any of the former human peons get upset about being locked out of the economy to starve, then have the agents pilot the drones, too.
Basically, almost none of the people imagining a future of superhuman intelligences have actually though through how it would actually work in the real world. We're going to spend trillions of dollars and vast amounts of resources chasing the goal of making ordinary humans obsolete. Now, that goal might be unobtainable, I hope. But I'm deeply alarmed at how much we're spending pursuing it.
Many indistries are changing, but in most cases the new tools will be more akin to cars that still need drivers, rather than robots who take over the whole job. Yes, jobs might be lost, or shifted to others, but it's not like suddenly 90% of people will have nothing to do. There were similar shifts in the past with new technologies, and we made it past them.
"ultimately accountable for the success or failure of a specific project, initiative, or activity"
I think that role should be reserved for a human, who can then use all the agents they like but has to take accountability for what is ultimately delivered.
The middle management in companies is one of the worst inventions ever. I think baboons have better middle management structure than us.
Might as well replace all that.
You know, as an IC, if I get get the level of introspection you can get from building distributed systems, and that with reasoning/thinking traces from LLMs, I think I'd prefer the entire level of middle management made out of LLMs rather than humans. It'd be helpful to be able to see exactly where their logic suddenly took a skip out the window.
"The famous primatologist Robert Sapolsky has spent decades studying baboons and has described how status, alliances, conflict resolution, and coalition-building resemble politics inside human organizations"
Baboons are a good example for studying mammal social management strategies, which humans also do.
They are organized by dominance hierrarchy, like humans, but baboons have a distributed leadership.
I think we could learn a lot from baboons when it comes to management
But the author's vision is also suspect, if you assume that the models will become much more intelligent:
1. Hypothetically, we can't give every human their own personal SkyNet to command. That would, uh, probably end very badly. If everyone gets an agent, those agents can't be too capable?
2. If you do somehow build a model that's much smarter than you, what do you contribute by managing it? How many people here have ever worked for a well-intentioned manager who couldn't understand the people they managed? So in this scenario, human management would be mostly displaced by agent management. Most companies could lay almost everyone off and let the agents manage each other. We only need humans to manage models now because the models are still pretty broken.
3. If we create models that can genuinely replace humans at almost any task, you won't be able to buy those on the API. At that point, the billionaires and the politicians wouldn't need human workers any more, because everything can be done better using their pet agents. Just have the robots build stuff for the billionaires directly. And if any of the former human peons get upset about being locked out of the economy to starve, then have the agents pilot the drones, too.
Basically, almost none of the people imagining a future of superhuman intelligences have actually though through how it would actually work in the real world. We're going to spend trillions of dollars and vast amounts of resources chasing the goal of making ordinary humans obsolete. Now, that goal might be unobtainable, I hope. But I'm deeply alarmed at how much we're spending pursuing it.