AGI has been 20 years away for the last 50 years

People flock to the next value-unlock, from PMFit to the metaverse

To me, it feels like the hype around product-market fit is over. Product-market fit surged in popularity when SAAS was not new anymore. For a brief moment, everyone realised that the big next value unlock was at product-market fit.

AI seems like a simple value unlock

However, when AI came along, attention for value unlock shifted towards AI. People were never really interested in PMFit; they were attracted to the potential money behind it.

That explains the superficial understanding of most people, and the lack of desire to go deeper, just like most AI evangelists don't understand neural networks or cognition, yet they rave on about AGI.

If it’s your first tech bubble, you can be easily persuaded

I guess this is the circle of news and hype, which for some reason keeps repeating itself. Still, my thesis is that it's mostly new gunslingers that jump on the hype train.

If you haven't been part of or invested in any tech hype, everything seems new and filtering out false promises can be challenging. After doing two or three, you get the drill.

It's like a 16-year-old girl falling for the wrong guy time after time, believing the promises of the next one that says 'he's really different'. After some breakups, she learns to avoid such guys.

We are really bad at predicting technology

Basically, tech bubbles, including AI, are like predatory bad boyfriends you learn to avoid in your teens and twenties. After big data, blockchain, NFTs, the metaverse, AR, AI is the true one that 'really matters'. Do you believe him?

We generally suck at predicting when technology will come about. See this graph with predictions of AGI. It’s always 10-20 years away or something. Sounds familiar?

“Fusion power has been one of those things that's been "only 20 years away" for about 50 years now.” - Source

In research reviewing tech forecasts, up to 80% of the predictions are off (Source + Source). Why, all of a sudden, did we get better at predicting AI’s future?

Technological innovation is non-linear, and a sudden jump in performance we’ve seen in the past years doesn’t mean anything for the future. It might plateau for 25 years, who knows? Nobody.