Prediction: AI will collapse

Prediction: AI will collapse

The dot com bubble got some things right. It noticed that the internet was going to be big. It noticed that the value generated by the internet would roughly scale in proportion with the amount of fiber laid. It recognized a latent need in all humans to use technology for self-expression.

There are a few things that it missed.

First: it thought that html and javascript were the means by which common people would express themselves. The dot com bubble thought everyone would have a website. As it turns out, normal people prefer expressing themselves with text, images, and video.

The first 1% of internet adopters expressed themselves by creating personal websites. But as the internet grew, we didn’t get more and more websites. The internet “collapsed” into a much simpler form of distribution: social media, which only let people express themselves in an incredibly limited fashion. Among social media, 90% is text, image, or video.

Html is not a good form of self-expression. It’s hard to secure. It’s hard to distribute. It’s hard to reason about. You need to learn how to code.

Images are an incredible form of self-expression. Non-technical people love editing photos of themselves. Non-technical people hate making websites.

The TAM of Instagram is 100x larger than the TAM of Godaddy.

Second: it missed that 90% of the internet’s traffic would come through video. As the total usage of fiber increased, we saw the internet once again “collapsed” into a simpler form of distribution: video streaming. Now, Netflix, YouTube, Tik Tok, Instagram, and so on take up the majority of data passed through internet fiber.

Roughly correct:

• GDP growth would scale with fiber usage
• GDP growth would scale with people in network

Not correct:

• Fiber usage would scale with number of websites
• People in the network would scale with number of websites

It would be more true to say:

• Fiber usage would scale with video streaming
• People in the network would scale with social media adoption

Why were social media and video streaming so popular? If I had to guess, I think the reason is that they are 100% sandboxable. Compared to html and javascript, which are only 99% sandboxable. You can ensure best practices, beware untrusted sites, and take a course in cybersecurity to keep a personal website secure. But there is still a 1% risk. You have to worry about phishing, cross-site scripting, malware, viruses, fake website impersonations, scams, and so on and so forth. For consumers, who are generally lazy, this 1% is too much to ask. Consumers want to completely turn their brain off. Consumers prefer social media and video streaming to “browsing the web” because they can do the former while drunk, in bed, having sex, on the toilet, during their lunch break, in the background while working on homework, or while their brain is 99% preoccupied with a difficult task at work.

The three big cornerstones of social media — text, image, and video — also behave much more predictably than html and javascript. You can build recommendation algorithms. You can design content delivery networks to store and cache media. You can hire engineers to design scalable systems. It is difficult to design recommendation algorithms, CDNs, and scalable systems for html and javascript. And so, the user experience of Netflix, YouTube, Instagram, and Tik Tok gradually improved over time, while browsing websites became worse and worse.

The modern AI discourse is making a similar mistake to saying “everyone will have a website” back in 1995.

I hear stuff like:

• People want personal software
• Everyone will build a mobile app
• The market for software is infinite

No.

You are directionally correct, but you are missing something big.

Every human will use agents. But less than 1% will use coding agents. (In the same way every human has a “personal page” through social media, but less than 1% have a personal website.)

You need an “idiotproof” sandbox. Your consumer agent, as well as agents for enterprise use cases, must have a 0% chance of failure. Your consumer agent must be usable:

• While drunk
• In bed
• Having sex
• On the toilet
• During lunch break
• In the background while working on homework
• While your brain is 99% preoccupied with a difficult task at work

It must be impossible for AI to: make purchases you did not intend, send messages you did not intend, and access information you did not want to share. It cannot be “difficult” for these to happen. It must be impossible. (In the same way that YouTube videos have a 0% chance of phishing, cross-site scripting, malware, and viruses.)

No amount of models getting better will give them a 0% chance of failure. To get to a 0% chance of failure, you need to be creative. You need to innovate. You need to radically rethink your abstractions.

Every human has a “personal page” through social media, but less than 1% has a personal website. What is the same shift for AI?

Every human will use _____ agents, but less than 1% use coding agents. What is _____?

What is the “idiotproof” sandbox for agents?