Understanding the AI ​​Mental Illness Debate

Box founder Aaron Levie brought us this week with a social media contribution in which he pointed out that Tech CEOs are “especially vulnerable to AI psychosis”. In the latest episode of TechCrunch's equity podcast, Kirsten Korosec, Sean O’Kane and I have done our best to decrypt Levie's comment. On the one hand, we have found that he does not reject AI tools, but only insists that CEOs must actually use these tools to understand them. This is a relatively gentle sound of scepticism compared to other signs of a broader counterreaction, be it when you look at the outgoing college students who bounce out any mention of AI, the bad mood around dismissals in the tech industry or the obvious rise in installations at the search engine DuckDuck after the announcement of Google, integrate more AI into the search experience. Kirsten suggested that Google is facing a dilemma where it “follows what it seems to have to do to stop, but it bothers what people connect most to the brand and does not improve it.” In the further sense, she wondered, “whether this anti-AI moment represents a chance for startups or other business areas.”

Read on to get a preview of our conversation, which has been edited for reasons of length and clarity. Anthony Ha: AI polarizes incredible. And that is part of what it is difficult to talk about; You can feel a bit crazy because [at the same time] everyone uses it and everyone loves it, but no one also uses it and all hates it at the same time. There are large quotas for both. On the user's page was something very noticeable: we have spoken about Google's announcements to search and about how AI becomes a larger part of the search – although it was interesting to see how Google has tried to take that a little back or at least add a few nuances if you want the 10-Blue Links experience, there are still ways to get it. It won't disappear. But I think many people aren't excited about the direction Google goes into. And so you can see, for example, that DuckDuckGo says that the installations have increased by 30%, which is a huge jump. Now DuckDuckGo is of course a much smaller product than Google. I don't think Google is in immediate trouble, but I think that is a sign that there is a very important target group that doesn't like the current AI direction. Sean O’Kane: I would like to say one thing that I always look out for when I look at all these leading AI labs or technology companies that really drive AI functions and products forward – for me it seems simply a collapse towards the approach of Anthropic to give this idea to really understand what people want to offer and adhere to it. And Google is one of those I would say that they are still driving the other direction. They try to do many different things, but they don't do a favor when they stay so vague. So I mean, if Google enters the stage on the IO and talks about the way in which it will change the search, they mostly talk about purchases or things that end in a commercial transaction. And I think that Google is, on the whole, an information retrieval system, especially for those who have been using it for two or three decades. Google can often struggle with this because they get reactive fears that they might damage the information call and their answer is: “Yes, but that will still be there. Let’s focus on how to help you book a flight or the like.”

And then they go and shoot themselves in the leg with letting go, so to speak – it must be a great challenge to underlie these systems in a stress test, but they go out and let go of this stuff and they encounter the same problems as for years. Kirsten Korosec: We have just published a great article about Google not knowing how to spell its own name. If you ask: “How many ps are there in Google?” there are two. It is this tension between: Google chases what it seems to have to do to keep up, but it causes that in conflict what people connect most to the brand, and it does not improve it. What I ask is that we have already seen some early evidence that people's fingers vote or go for them by literally going to another worship. But I wonder if there are opportunities for other startups or culturally, whether this anti-AI moment is a chance for startups or other business areas we have not really thought about. Anthony: Definitely. Here too, it is probably a challenge because the opinions are so different. And if you develop something that is tailored to a group that is skeptical towards AI, you will likely irritate other users who are much more evangelistic or enthusiastic about it. But I think that's exactly the moment we live in. And you can see by the way DuckDuckGo advertises for himself that they strongly emphasize this anti-AI version, which I find very noticeable because I have already mentioned that I have removed myself from Google and tried out other search engines. And I would say that a year ago, when I started this exploration, even those alternative search engines still tried to experiment with AI functions, stressing the AI to a certain degree, because they thought they had to do it. And now I believe that they realize that there is actually a way to say: “No, we were simply not interested in this stuff at all. Or by doing it, we put it in a separate sandbox that will not affect your central search experience.”

Kirsten: I think we categorize wrongly sometimes all technology CEOs as people who force-feed AI. And there’s at least one tech CEO who said: “I think that there’s a certain psychosis with other tech CEOs regarding AI.”

I'm talking about box founder Aaron Levie, who has often come to disrupt and is certainly a friend of TechCrunch. He said that CEOs are particularly vulnerable to AI psychoses, because they – and I read this – are “sufficiently removed from the last mile of work that needs to be done to generate the greatest possible added value with AI.”

I really thought that was interesting. And I wonder if there are other CEOs who agree. I am also wondering whether, in the context of this rethinking of what has to happen in order to generate the greatest possible added value, they also think about how their workforce changes, what is our other topic today – it is not just about the AI gap, but also about how AI changes work. And we have certainly seen some of the bad news, namely many dismissals. But I also think that we see great changes in the way people work. I wonder if you see evidence in the areas you both cover, because I don't think it's just the "KI startup sector" or the big technology companies. Sean: As far as the companies I report about, so many of them are concerned, if not with physical transportation, then with adjacent things. And it seemed much slower there than on the software page, which is not surprising. We begin to see that something changes. We talked a little bit about Mind Robotics in the show, the spinout of Rivian CEO RJ Scaringe. And you know it will certainly use more AI in physical infrastructure and manufacturing as well as in robotics and autonomous driving. I think on the software page things are really changed where there are people whose job is directly related to code production. Anthony: Some of the question concerns, in my opinion, both the introduction of AI into companies and AI-related redundancies – to what extent do they take place from top to bottom or from bottom to top? Because I think that many other changes in the workforce have taken place in recent decades at least to a certain degree from the bottom to the top: these are tools that people really like to use, they bring them in, and from a certain point on, the executives and IT managers accept that. There is the feeling that a large part of [convincing that there will be] AI productivity gains seem to be accepted by the executives – or if you are in a startup, probably by the VCs you finance – who love this dream that you can be as effective with just a small team as a company with a much bigger team. And I don't think that's absolutely impossible, but I think that Aaron's argument is essentially that you don't really touch the end work, how should you know that? He is also not someone who says we should just throw away all AI tools, but he says that you really need to use and understand what they are doing. You can’t just look at a slide and say: “Yes, incredible efficiency, let’s go.”

Kirsten: Well, I think there are many real evidence that these companies use these tools and that this affects workers directly in the form of redundancies and in the way they work. The two truths are true here.