10 Controversial Things I Believe About AI That I Shouldn't Say Out Loud

Having an open mind is a superpower but at some point, one has to start living

This article is a repost from November 2023. There’s an ongoing vibe shift in AI. In a way, the controversial things I share below (not so controversial anymore) hinted at it back then. Since many of you have subscribed to TAB after I published this article, I thought it’d be a good idea to resurface it. It was paywalled, now it’s not. Short and sweet, unlike my usual writing. Enjoy!

Thomas Carlyle said:

A man lives by believing something; not by debating and arguing about many things.

G. K. Chesterton said:

The object of opening the mind, as of opening the mouth, is to shut it again on something solid.

These two great quotes are meant for me. The way I usually think about stuff is I tend to hover over topics to understand distinct points of view (often failing to meet my own standards). To do that satisfactorily, I have to pay the price: I fail to get too close to any particular stance for fear of becoming one with it and losing perspective of the value existing elsewhere.

I believe there’s virtue in trying to find something worthy where those around me have lost hope even though sometimes, eventually, I find their approach to be the wiser one. To me, having a nuanced — often contradictory — set of views on world matters is critical to be able to make sense of a reality that will try its hardest to be absolute nonsense.

But it’s just as important to not hide in the safety of abstract ideas, thought experiments, and endless debates and settle down on a specific attitude about things. Otherwise, I am just a witness to a world that unfolds without me. Otherwise, as Carlyle says, I am not really living.

I want to live. That’s why, as a first step, I’ve compiled ten of my most controversial opinions about AI. I never write about these topics in such blunt terms, but if I were not forced — and tempted — by a complex and multifaceted reality to step out of my bubble and consider alternatives (which more often than not provide me with a healthier perspective and with a window of opportunity to change my mind), that’s what I’d do.

I still ascribe value to contrarian views — or at least partially differing — to the ones below, but if I glance deep into my soul, this is what I see. So, here you go, 10 controversial things I believe about AI, in no particular order.

1. AI could be a more impactful technological innovation than the Internet, the printing press, or fire have been, but to prematurely label it a revolution is damaging insofar as it unnecessarily raises expectations for those who don’t know how it works and how it could fail to work in the future, while leaving no room for absorption of the failures typical of arduous scientific endeavors — that’s how AI winters are born.

2. As a general rule, people who have developed or want to develop a relationship with an AI (e.g., friendship, partner) would be better off working on their emotional — both intrapersonal and interpersonal — skills to improve their relationships with other humans instead.

3. Large language models (LLMs), as they stand today, are a net cost for the world: socially (where does the data come from? who does the RLHF? who is going to lose their jobs?), economically (are those millions better invested there than somewhere else?), and scientifically (are other approaches to AI being stifled by the LLM gold rush?). The purpose LLMs serve doesn’t come anywhere close to compensating for the toll they create.

4. It makes no sense to force AI into one’s life; if using it doesn’t come naturally to you, then you’re better off without it. Perhaps you don’t have any problem where AI can help you and that’s fine (AGI is a completely different question, though).

5. Throughout history, many technological innovations have indisputably improved the world from the onset; I think generative AI (understood as AI systems intended to create data of some kind) is not one of them. So far, it is being used mostly for trivial tasks (e.g., cheat on homework, do marketing copy) and that probably won’t change anytime soon because humans won’t change.

6. The world would gain more from AI being only open source than from it being only closed source, including the possibility of malicious actors using it to do harm. Centralization and private control are better in specific but not super common situations (e.g., high-tech weaponry).

7. People who are concerned exclusively (or mostly) with studying the medium-term (e.g., disinformation) and long-term problems (e.g., existential risk) of AI do so from the privilege of not being affected by the short-term problems (e.g., bias and discrimination, copyright violations, job losses), which includes the vast majority of people studying AI in the first place, creating an inherent imbalance between what matters to most people now and what the people who have the power to decide focus on.

8. Everyone who hypes AI to some extent — companies, researchers, journalists, influencers, bloggers like me — has a vested interest (e.g., personal, professional, financial, reputational) in the hype existing as an end in itself. This applies just as well, or even more, to anti-hypers.

9. Even if AI allowed humanity to reach the stars (metaphorically speaking) most people would only use the tech to fulfill their most basic drives and needs: make money (e.g., spam sites), minimize effort (e.g., homework cheating), and get off (e.g., deepfake porn).

10. AI enthusiasts live in an echo chamber: Most people don’t care about AI. Most people won’t ever care about AI, not even once it is ubiquitously integrated into their daily lives or once AGI is achieved and it redefines society’s rules.

That’s it. Come at me in the comment section with your anger (or your support, you crazy ones). Both are welcome as long as you do it constructively.