Hacker News Reader

Comment by mxstbr | Hacker News Reader

…writes a person who's never been a CEO before.

Which is not to say that AI won't be immensely helpful for running companies; just that once you're in the role for two days you realize that replacing that role entirely isn't going to happen.

> AI can win at chess now. You think being a CEO is harder than chess? Come on.

Being a CEO isn't like playing chess. Being a CEO is more like a really complex version of self-driving cars. Why? Because it involves interacting with other humans, all the time. It's all people problems.

Chess isn't. Chess is decision-making with a finite (if immense) set of possibilities. Very different set of problems.

View on HN

>just that once you're in the role for two days you realize that replacing that role entirely isn't going to happen.

Journalists might very well say the same thing (when people who've never been journalists talk about replacing them with AI).

The kind of journalism that journalists want you to think they do - going out into war zones or directly interviewing people to find out new things of public interest - isn't replaceable by AI. The kind of journalism they actually do - hanging out on Twitter and repeating whatever goes trending there - is easily replaceable. They did this to themselves.

That is certainly the view of many people who aren't writers or journalists. We'll see if it plays out. Jobs do tend to appear simpler than they really are when looking in from the outside.

>They did this to themselves.

And who set this bad direction? The CEO. This hardly adds to the case for not replacing him with an AI.

It is also the view of many working journalists, or at least the ones I know.

And it's not like some CEO decided "hey, let's pivot to churning out low-quality clickbait because I really like that stuff." They were forced to by the shift of ad budgets from print to more targeted advertising and the consolidation of news readership into a small number of well-known publications. Though I do agree that it's equally unfair to blame journalists for this.

> They were forced to by the shift of ad budgets from print to more targeted advertising

Indeed. As with almost everything else terrible about the world today, the root of the problem is targeted-advertising-based business models.

It really has caused a race to the bottom anywhere it becomes the primary business model. Everyone hates it but no one’s got a solve to get people paying for things once they are “free”.

We'll see how successful the autogenerated clickbait farms end up being. But nothing you're saying is really making a case for the company needing a visionary human intelligence for its CEO. If the entire business model of the company is strictly determined by external conditions, then the whole thing seems ripe for automation.

> forced

Only if you consider it to be inevitable that their primary goal is to make themselves and their owners preposterously rich.

Chess also has a rigid, well-defined set of rules. Business does not. Being a CEO requires combining a lot of (sometimes nebulous) mental models, each with a degree of uncertainty, to some overall business strategy. I think it was Sam Harris who pointed out this is why there are child prodigies in subjects like math and chess, but none in something like political science.

"Chess also has a rigid, well-defined set of rules. Business does not"

I believe every MBA Program, Management Department and Executive Coach would beg to differ

If that's the case, can you define those succinctly here?

My MBA program did not teach a rigid, well-defined set of business rules. Even something rules based like GAAP is somewhat open to interpretation. And when it comes to something like strategy or finance or organizational structure the options are endless.

Only the dumbest tik-tok positive grindset twitter thread version of the above suggests that doing x, y & z will gaurantee a win in business.

> mental models, each with a degree of uncertainty, to some overall business strategy

That actually sounds a lot like model based reinforcement learning. AlphaGo still had an explicit model of a perfect information game, but newer algorithms are capable of learning the rules of games like Minecraft with very little prior information.

The problem is that the real world is still much more complicated and hard to access for machines. However, I wouldn't say it is a fundamentally different problem, but rather a matter of scale. That's part of the reason why AI research spent so much effort on building game playing systems.

I'm not so sure it is a matter of scale. Humans have already lost the scalability race to machines. But I still don't think there is AI developing political policy. Or, if they are, like in the case of setting sentences/bail for criminals/suspects, they show how bad current AI is at it, despite being able to scale more effectively.

'Scaling' IMO is one of the things SWEs harp on a lot, to the extent that they see it as a solution to every problem. This is in part because they recognize that software scales well. When you're a hammer, every problem becomes a nail.

There's also some question about the fragility of some of these models, compared to humans [1]. I think there have been some other examples where changing one rule makes their aptitude crash. The implication being they are still hard-rule based, but scale well. But that scale advantage erodes quickly with a rule change because of the lack of contextual understanding.

[1] https://arstechnica.com/information-technology/2022/11/new-g...

Reply on news.ycombinator.com