so... "within 2 years" appears to be the runway for which we can expect a more conclusive verdict on AI investment?
No one’s networking is worth hundreds of millions. A reasonable approach is to cap CEO salaries at $2-5M, and invest everything else in the business. How is this already not being done, AI has rendered a lot of their “analysis” common place.
- CEO salaries, even when they are high, are still a drop in the ocean compared to the company’s revenue or profits.
- Elon Musk and the CEO of Starbucks have very famously been able to boost stocks just by existing…
I was with you until here. Honestly, some networking is actually just that good. I dont believe this to be the case universally, but some people can connect people so efficiently that its mind blowing.
>A reasonable approach is to cap CEO salaries at $2-5M
Why is this reasonable?
Only if we include corruption. Any deal that takes special consider is arguably corrupt.
Like theres plenty of information thats not public, but isnt confidential either. Accessing that information can be exceptionally lucrative. Now if that information is used to trade publicly traded shares, that can be criminal (the ethics of the situation are another matter) but if you were to instead, provide product or service on that basis, you are in fact just a well connected salesman. Not everything that doesn't go to a public tender is corruption.
as opposed to saying CEO’s should be compensated less
a muuuuch greater amount of dilution would come with that though, I’m for it
It could (??) also lead to comp going up for CEOs, as I imagine (??) they would demand more in stock because stock is riskier than cash.
However, only months ago Anthropic published about their autonomous vending machine experiment which had both humorous and interesting but not financially successful. I think maybe we're not to the point where AIs can run businesses. https://www.anthropic.com/research/project-vend-1
Now, actually getting in a room and talking to those people and getting in on the execution of that idea... that's a different story.
It's quite possible that an AI would do a better job at CEO than running a vending machine.
And it's quite possible that successful CEOs would not succeed at managing a vending machine. It's almost guaranteed that many will fail, given that some C-level people are selected for risk appetite and social skills over actual business acumen.
But it also kind of reminds me of the DAO hype a few years ago where people suggested decentralized, blockchain-based organizations could be the org structure of the future. (Not realizing that they were just putting a co-op on a blockchain :) )
Now that I think about it, I do hope someone tests this theory out. It would be very interesting to see. You could probably start by saying "You are the CEO of a lemonade stand. Your goal is to maximize revenue while [constraints]..."
- hallucinating conversations with fictitious people
- role playing as a real person and telling them they are present in meatspace, or claimed to have visited a fictional location
- hallucinating details of an account involved in payments
- both ignoring lucrative opportunities and accepting loss-making deals
- making commitments before having done any research
While yes, the skills to run a vending machine and lead a company are not exactly the same, I think the nature of failures discussed means they would likely affect both roles.
I think there's every possibility that a present-day AI allowed to act as CEO would make mediocre choices for some period and then decide that it was a character in a scifi novel selling bespoke space-yachts to comet-mining magnates.
maybe thats why you're not a CEO
IMO, the role of government this is to establish societal baselines in terms of criminality of certain actions and to ensure a baseline of competition. Actually acting on the former and the latter... the former by affecting the financial wealth of executives and board members when a company's neglegence harms society. The latter by not giving outsized govt contracts to a single provider for anything and to establish that if there are restrictions (medication, for example( that multiple providers/manufacturers must be present in the market, and that a certain portion is produced domestically (for security).
You said it like nobody hasn't attempted it. Many people are trying to solve this. They've replaced small mechanical works like scheduling meetings and reaching out already.
The majority of CEO job is excellent judgement and motivating people. It's difficult to replace that function.
A lot of CEOs are founders and owners. You can bet they want the business to run by itself, so it would free up their time to go do something else. That's the dream. You do realize that these people will earn regardless who runs the business, right? because they are significant shareholders. In a way, you wish for CEOs/founders to be much much much richer.
Average CEO in a nutshell. CEOs should be reviled.
I kept reading and quickly realized this entire report reads like a bunch of bull,
"94% of CEOs believe AI agents could provide equal or better counsel than human board members."
Board members hold CEOs accountable. How is an AI agent going to do that? "hey chatgpt, can I have a raise this year?"
AI automation is a high risk and the rewards are questionable at best.
Nobody ever got fired for running a profitable business with real employees, creating value for their customers and executing well. Note the word AI is NOWHERE in that sentence.
My company just spend months on several AI initiatives. They created several power point presentations about how we can use AI to automate all of the teams repetitive tasks. Our senior leadership didn't even know that we've been doing automation for almost ten years now.
They think automation IS AI.
Q: What did he do to solve the problem?
A: pasted the raw react code and asked an Llm to translate it to html.
The Llm did a terrible job. For example, it converted unordered lists to paragraphs with break tags. When I asked why he didn’t just copy the rendered output from the browser, he didn’t know how to do that.
When he was shown how to do that, he said it would be too much work because the rendered output still contained css tags. So he would have to get the Llm to remove the tags for him.
This is an engineer who’s been developing software for at least 10 years, well before Llms even existed. Yet he cannot figure out how to do a simple regex find + replace.
Even if your buddies don't get you into a cushy position, a multi million salary for a decade or so sets you up. No tears shed.
pfdietz•1h ago