Artificial Intelligence and Brain Drain » Enrique Dance

OpenAI’s generous $6.6 billion funding round, which values ​​the company at no less than $157 billion, allows it to temporarily hide a problem that will be recurring for quite some time in all companies developing artificial intelligence products: the brain. . drain.

In fact, all indications are that a significant portion of the billions of dollars raised in the round will be directed toward a goal that investors typically don’t think too much about: buying back employee shares that were, at the time, partially paid for with stock options companies and who have been passionate about “life changing” for some time.

The question is complex: on the one hand, as we have already seen, working at OpenAI or in general in an artificial intelligence company these days is without a doubt a guarantee of the opportunity to go into the chosen industry. and demand astronomical salaries for the development and implementation of what he can do. Obviously, it is not that simple: launching a successful AI initiative requires a very large skill set, and many of the variables involved are not even in the hands or control of the manager who intends to do so. , which is also usually an expert in only a certain subset of these skills. Even so, few would complain about their professional expectations if they were working in this field today.

On the other hand, the fact of employee stock buybacks, in addition to irritating investors who see that their money is not being used to acquire the resources that the company needs to make more money and justify its astronomical valuation (expectations are reasonable, yes, but if they do not seek to be converted into money at some point, they have a limit), this does not guarantee that these employees will continue to remain committed to the company and their work, because when you receive an amount sufficient to think about making changes in life , your priorities often tend to change, and simply working on managing your money begins to play a much larger role.

In this context, we are no longer talking about the departure of some co-founders or the most famous employees of the company who had packages stock options more importantly, but from other levels who see that their bosses are already multi-millionaires, that their talent can be better paid in other companies or that they are even looking to create their own and in some way repeat the move they saw in the period not so for many years now.

The concern extends to the fate of Sam Altman himself, who has seen rumors linking him to a huge stake in the company that will be transferred to him to “align his interests” with those of OpenAI, rumors he has denied. On the other hand, a funding round includes a huge line of credit that can allow the company to pay a lot of expenses, but at a price that is anything but cheap.

If one thing is clear, it is that no matter how much Altman argues that creating artificial general intelligence alone is enough to justify any investment, OpenAI will have to demonstrate the ability to do more than just generate billions of dollars in revenue. , but also to put an end to the excessively growing consumption of computing resources. Something that requires not just specific conviction and will, but also management, negotiation and optimization skills, which is not at all easy.

OpenAI’s problem for the first time is no longer one of applied science, but of nasty metal: how to prevent the people who work there, blinded by huge investments, from wanting to benefit from their experience and talent elsewhere, feeding competitors, or even devoting themselves to enjoyment with your own money, because life is short and you have to live it. We’ll see how he solves this problem, and how he does it in such a way that his investors don’t lose patience as they see their money being funneled not into buying more computing power or more storage, but into paying for luxuries for employees who are after seeing the numbers of the investment round, they already know its value.

It won’t be easy. But in practice, when we encounter innovations that can cause dimensional changes, this almost never happens.