Marc Andreessen Says AI Can’t Replace His Job: VC Tech Investing

Marc Andreessen is undoubtedly the most famous venture capitalist on earth. The co-founder of the legendary venture capital company Andreessen Horowitz, inventor of the first popular web browser, and by a reputation an intellectual grandeur so largely read that his colleagues call him “Marcgpt”. And as he beats his nickname, Andreessen is a great believer in the future Propelled by artificial intelligence. His company – “A16z” at the Silicon Valley Sophistate – Invested in the Openai of Elon Musk and Sam Altman. Andreessen called ai “Our alchemy, our philosopher stone” and “a resolver of universal problems” which “increases the capacities of our machines and ourselves”.

But for Andreessen, there is a job that AI will never do as well as a living and breathable human being: his.

Do you think I’m kidding? On one Podcast A16Z Last week, Andreessen considered that being a venture capital Perhaps a profession that is “literally timeless”. “When the ISs do everything else,” he continued, “it may be one of the last remaining areas that people always do.”

Here is the logic. Andreessen begins by talking about all the things that people thought could disrupt the functioning of VCs – as the approach of Craigslist in Angellist, or crowdfunding. “The other form of structural change, of course, is AI”, explains Andreessen. Then, he challenges the crowd of AI: “Very well, intelligent guys. You are sitting on this analysis, and you have all these intelligent people who do all this modeling and all this research and so on. Why don’t you connect this to Claude or Chatgpt or Gemini and have said what to invest?”

The reason, explains Andreessen, is that it takes a VC like him to know how Choose a winner. He throws a lot of examples, going back to the whale hunting industry 500 years ago: book publishers, film studio leaders, talent scouts at Music Labels. (I’m going to save you the details here, but I spoke with an economist who analyzed the whale industry, and he says that Marcgpt is roughly wrong with each account.) Andreessen insists that these are key jobs that arise “each time you have a part of the economy in which you have a entrepreneur.”

Here, supports Andreessen, this is the place where the human element is irreplaceable. “You are not only funding them,” he said. “You must really work with them to execute the whole project. It is art. It is not science. It is art. We would like it to be science, but, as it is art.”

Now correct me if I’m wrong, but it seems that many people of AI have tried to tell us that IA can make art. Last year, even Andreessen said that AI had enough sense of humor. “save the comedy. “But apparently it can’t do her art.

Which brings us to the crazier. Andreessen says he knows that business investment is an ineffable, intuitive and intrinsically human competence precisely because the venture capital are Very bad. “The big VCs have a success rate, I do not know, two of the 10 of large companies in the decade, right?” he said. “If it was science, you could possibly have someone who is simply composed and gets eight out of 10. But in the real world, it’s not like that. You are in the sector of dynamism, and there is an intangibility. There is a taste aspect.”

Even by accepting the premise of Andreessen – that the VCs provide better advice than AI on how to manage a business – it seems that it is wrong to know if it is replaceable. In a recent survey carried out by the corporate software SAP, 75% of level C managers in companies of $ 1 billion said that AI already gave better commercial advice than their friends and colleagues. And 38% said they trust AI to make business decisions. The kind of people on which Andreessen counts for his livelihood already begins to think that he is obsolete.

But the portrait of Andreessen on the way in which the work of venture capital does not really agree with reality. VC investors say they are looking for a disruptive innovation. In practice, they worked roughly like any good network of good old people, constantly funding many more white men than women or people of color – often the same white men as they knew previous startups, whether they succeed or not. And economists say that this is an open question if the VCs really add value or are only the most basic type of “pickers”, identifying companies It would have succeeded Without them. This is the kind of operation in which large languages ​​are quite good. The identification of models in large data sets is, like their whole.

I actually agree with Andreessen. I am skeptical about any form of AI – even less generative, Openai chatbot type products Or Google – will never be able to make a high -level critical and creative thinking as well as a human. But the reality is that the quality AI work may not have any importance. History is essentially a large cemetery of handicrafts which ended up being automated, even if automation has produced a Objectively lower product. Andreessen, like many of us, wants to think that he is special – that no machine can never do what he does. But he cannot have both ways. If he is right that artificial intelligence cannot perform the types of skills his work needs, then he is wrong to invest in companies that promise him.

In any case, when Andreessen says “AI cannot do my job”, the work he describes is not a venture capital. It is not even an ordinary investor, which buys actions in a company. The main function that the Endesen is lionizing is Be a porter – The power on which can join the influence club.

An AI that has formed each decision that each venture capital has ever taken and tried to align them with various definitions of success could choose very different types of businesses. In a world of VCAI, startups could finally rely on their merit, instead of knowing how much their founders resemble and speak like Andreessen. Who knows? Heuristic with the black box of the VC can promote things as if an idea is “good for humanity” or “promotes the mobility of the classes”. What if it was starting to hallucinate something on the “wealth of redistribution”?

At least at least, Andreessen is probably right. No machine could never replace him.

Adam Rogers is a main correspondent at Business Insider.