F2F #61: AI Hype - Breadth vs depth

Every new cycle brings the same dynamics, with new players and new products. But we can reduce them to people who talk and people who build. Follow the builders.

F2F #61: AI Hype - Breadth vs depth
Photo by NEOM / Unsplash

Evaluating new technologies is always complicated. Hype cycles create perilous dynamics disguised as dopamine hits with each new release, product announcement or new shiny thing. I like new shiny things. I can't help it.

The pursuit of understanding AI as a whole has led me to want to try every new thing out there and to know everything about every single tool out there and left me with knowing nothing about anything. I even felt the AI burnout of More AI, more problems.

I have been long enough in business to spot the trend here: every new trend (passing or not) follows the same logic: there are people who talk about it and people who build with it.

We tend to follow those who talk about it (mostly threadbois) because they post shiny new things every day, with "use cases you can't miss" and the mandatory "this is not a paid post, I just like this product". Of course. And we also follow them because builders are busy doing builder things, like, for instance, building.

Once you get trapped in the vicious circle of following threadbois, you get sucked into oblivion by the incommensurable feeling of not being able to catch up with everything. The FOMO dread is growing every day, and you jump from trial to trial, from demo to demo, wasting resources, time and brain cells in the process.

In doing so, we miss out on the opportunity cost that builders are capitalising on. By betting on depth, they're able to milk every single drop of value from the tool they heavily invested in, instead of jumping from flower to flower. Instead of doing this search by breadth.

Personally, after almost two years of trying too many apps, I've decided to go back to basics and just use ChatGPT. I have cancelled my accounts (and zombie subscriptions) in Poe, Riffusion, Suno, Midjourney, Claude, Elevenlabs, Descript, etc.

In an ever-changing world where every app releases new features every week, change scope every month and update models daily, you're more likely missing out on what's happening if you don't master at least one of them.

While I don't actually like all the shenanigans Sam Altman is pulling, and I always shudder when I have to recommend using a big tech company, I think most people out there would be better off just using ChatGPT, because, right now, as things stand, most people are using only 10% of ChatGPT and squander precious time and resources trying to make other tools provide a fraction of the value they'd get off ChatGPT alone.