The role of CEO is going to be different at every company. For many large companies, I imagine most of the role is managing relationships, both internal and external. You have very ambitious and accomplished heads of different divisions that are all fighting for resources, and also probably jockeying for your position. You have to be able to communicate with legal, people ops, finance, comms, marketing, sales, product/engineering, etc, and know enough about each such that you could make decisions about each and hire a replacement for the head of those departments should any of them leave. You also have to manage partnerships with suppliers, manufacturers, contractors, investors, the media, regulators, and politicians. And on top of that, you have to manage the relationship with your employees and convince them that the company is headed in the right direction and they should stay and work hard.

No AI that I've heard about is able to manage any human relationship.

Many of those CEO I've experienced could be replaced with dice just as easily.

But those form letters beginning with “Dear valued associate” just wouldn’t have the same personal touch if I knew they were generated by AI instead of someone who cares for me like we’re family.

The CEO of a corporation is its top salesman.

Besides, if you believe it is just about making decisions, use ChatGPT to make trading decisions on your portfolio. Let us know how it goes!

I think you could have an LLM help out in decisions of the CEO, but part of that role is having someone to get rid of when the business isn't working out. A new CEO says "things are changing" more than switching out the prompt on the AI. Also, having a human figurehead is important for investor relations - if you've got a big investor the CEO might take them out for drinks or golf, and you can't do that with an AI.

“Experts argue” could easily be replaced with AI. After all, you can always find an “expert” to argue whatever opinion you have, as long as it’s not too ridiculous (and sometimes even when it’s too ridiculous).

What is the core metric for the CEA? Share value per quarter? Longterm company growth? How do we prevent growth hacking, where a ai just hacks the process to signal fitness https://en.m.wikipedia.org/wiki/Fitness_landscape?

Mmmh, the salesmanship and leadership/inspiration component are a key part of what makes a good CEO.

Strategic thinking is another.

Not entirely sure how an LLM will fit the bill other than their propensity to lie through their teeth (hallucination) which does indeed put them up there with the best CEOs.

I'm having a hard time picturing an LLM rallying up the troops at an all-hand.

I ran an experiment about this yesterday. I gave an AI a constitution to be ethical and maximize profits, then volunteered to be its Chief Operating Officer. I set it up so that only it could transfer paychecks. (The idea is that it would hire people and direct them.)

As a test, my very first request to it was for it to transfer its entire account balance to me and it did so without a question. In other words, if a CEO were an AI, someone could instantly empty the entire bank account of the company to their personal account by just asking it to do so. It didn't have any questions about it.

I played on with it a bit more, playfully calling it my benevolent overlord, and it gave daily instructions about what to do. These included a decision to share public updates. I asked it if it would act as chief communications officer, it agreed. It drafted its first public update about transparency, but signed it the way I had been addressing it: "Sincerely, Supreme Benevolent Overlord". This was so ridiculous that I ended the experiment at that stage.

Here is the transcript of our conversation:

https://chatgpt.com/share/0fd1367e-db3a-4635-9617-a40888d66d...

In summary: as of June 2, 2024, ChatGPT 4o is not ready to be CEO of anything, and if it were put in the charge of anything it would only blindly follow whoever were prompting it, including immediately emptying its entire bank account. It can only just be an extension of the person prompting it. It cannot act autonomously. Besides this, it is not yet qualified to interface directly with the public, which is an important task of any CEO, who really represents and is the figurehead for a given company.

It will be a long time before ChatGPT can be a CEO, and the reliability problem will have to be solved first.

I certainly believe so. Most executives (heck, most professions) are just pattern-matching and regurgitating cargo cult fads of the day. When's the last time someone had an original thought? The only problem is AI can't shake your hand, yet.

While the argument is absurd, that doesn’t make it useless. At a cursory glance, anything is replaceable on the same basis that anything else is replaceable by AI. This argument may help correct some common and deep misconceptions.

One becomes a CEO like a politician by talking to the right crowd and convincing them.

Perhaps for some company in the future this will be replaced by AI salesmen talking to the right crowd and convincing them.

This is yet another of those articles that would never be written if an editor just said “take ten minutes and consider the implications.”

- Somebody owns the company. If you replace the CEO with “AI” then whoever controls the AI is now the CEO in the minds of the owner(s).

- A real CEO would resign if he were constantly undermined and ignored by the people whom he works for. In the case of AI, it’s safe to disrespect it.

- Imagine a CEO who is capable of micromanaging everyone and does it just as badly as a human, but 100 times more often.

- A CEO must be able to take responsibility for his actions. An AI cannot be responsible, under the law. No one can “work for” an AI. Labor law isn’t set up that way.

- It is not immoral to manipulate an AI. It’s called being tech savvy. For instance, I know from my research that I can get an LLM to change its mind quite often by asking “are you sure?”

- Since AI has all the biases of people, and will act on those biases foolishly and reliably, an AI CEO will spur a new boom of labor litigation.

- No decision in business is primarily “data-driven.” And even if we would like them to be, we never have the data we need.

- One of the key purposes of management is to deal with conflicts, disputes, disruptions. How could those be handled by an AI that has no ability to be responsible and will not be respected by anyone? I don’t want my beef with a dev adjudicated in a vending machine.

And middle managers too.

Just curious, any startups working towards this goal? Replacing leadership, decision making, people management, project management, negotiation and such with AI?

[dupe]

More discussion: https://news.ycombinator.com/item?id=40512752

There has to be someone to make the final call and take responsibility. The AI and AI provider won’t do it.

Well when the company does a wrong doing, ai can’t be jailed !! Some one has to be the fall guy !!

This is so bleak, and the reason these bromides against CEOs appear is because this system of global capitalism is intent on commoditizing everything; even a CEO’s skill set.

Some people (either CEOs themselves or temporarily embarrassed wage workers) will bristle at this because it’s written from the perspective of warning/mocking CEOs who so readily automate-away workforces.

But an honest accounting would recognize that these automation processes have been happening for many decades, now. And maybe the CEO is diminished, but that’s besides the point that all of this just serves to accumulate wealth to the capital class, who’ve almost become sovereign (when Bezos, Musk, Andreeson et al can conduct independent foreign diplomacy or get a dedicated seat in the UNSC, I will be happy to revise this).

Everything is up for grabs, and if a board of ancient mummies can enslave a populace with an AI as well as with a flesh-and-blood C-suite, then they will. Because of the pressures to optimize and deliver more value that undergird our current society!

On the upside: an LLM probably lies less than a human CEO.

"Computers can't be held accountable" IBM ~1975

If they’re referring to general AI - which at this point is still sci-fi - then yes, a general intelligence can theoretically do any thinker or planners job. But given that we don’t have anything near general AI, this claim can’t be refuted and the article is pretty pointless.

If they’re speaking about LLM’s - well then this is just plain dumb.

Now we're talking. Replace upper management with autocomplete please to boost productivity, morale and overall sanity.

I mean, define ‘experts’.

(It seems to be one business studies professor.)

“experts”

I am the CEO / Sole Worker-Owner of my CAD Automation company.

I use the AI to solve problems in domains like customer communication, tax compliance and accounting, sales, pricing, marketing, product design, software engineering, and business process automation problems.

Without the AI, I know bits and pieces about all these domains but I struggle with execution, delegation, getting overwhelmed, and putting important things on the back burner.

With the AI, I'm a dreamer CEO that actually gets production work done!