AI Hype vs Crypto Hype

TLDR

The AI hype wave and the crypto hype wave from a few years ago feel very similar, and I wanted to reflect on why I bought into the AI hype wave so much more than the crypto hype wave, despite possible long-term similarities in value.

Hype men be hypin

I’m not sure if technology has always been this way and I just haven’t been paying attention, but in the past ~5-6 years it feels like we’ve been through two identical hype cycles: crypto and AI.

Crypto (aka “Web3” 🙄) hype went something like this: “crypto provides solutions to everything; if you don’t adopt crypto now, you’ll be a technological dinosaur and completely irrelevant to the tech market in 5-10 years.”

AI hype is currently going something like this: “AI provides solutions to everything; if you don’t adopt AI now, you’ll be a technological dinosaur and completely irrelevant to the tech market in 5-10 years.”

I feel like we’re just over the cusp of AI hype, where the critical mass has started to realize that AI will not be the panacea it claims to be. Just a couple months ago I don’t think we were there yet, and even I was more willing to consider that AI could completely transform and revolutionize our lives. Now I’m not so sure, and I think more and more people are seeing it this way too. My suspicion is that it’s because AI tools have “stabilized” to some degree in 2025 and people have finally dug in and deeply explored them, only to find that the benefits do not live up to the promises.

I never bought into crypto hype, so I find myself wondering why I bought into AI hype? I think the answer is in the application of the technology. AI claims to (and, to some degree, appears to) be applicable to any knowledge-based job or industry. It does a great job at appearing smart and competent at tasks which take up a lot of human labor hours. By some measurements such as lines of code written, AI clearly outperforms any human developer; although lines of code is a poor metric in itself, the clear victory of the machines on this front felt like an indicator of great things to come.

In contrast, crypto was always pigeon-holed, and no amount of hyping could cover up the simple truth that the fundamental tech is just a fancy ledger. Blockchains have a lot of interesting features, and smart contracts do feel useful for some specific issues, but it was extremely easy for non-experts such as myself to see that crypto was never going to take over the world. Even in the realm of finance where blockchains got their start, it never felt like crypto stood a chance. Decentralized services sounds great to a certain class of people, but typical users do not gravitate towards them, and they have significant downsides which are hard to argue away. To top it off, there are significant technical hurdles such as energy consumption (yeah, I know, “proof of stake” or whatever). Try as they might, crypto hypemen simply couldn’t convince me that it was going to change the world, because it was reasonably easy to see the holes in the logic.

On the other hand, AI hypemen have done a really good job at spinning a narrative of AI as a perpetual motion machine, continuously powered forward by its own innovation. The AI hype narrative states that AI will improve at an exponential rate because AI will continuously unlock new and exotic capabilities through emergent behavior. It is much harder for me to argue against this because I can imagine a hypothetical reality where a sufficiently powerful computational system is capable of improving exponentially by way of automatic self-improvement. Perhaps I read and watch too much sci-fi, but this feels like a plausible scenario to me, at least in some universe somewhere. And critically, it isn’t quite possible for a non-expert to peer behind the curtain of AI and accurately predict how likely it is for that scenario to unfold. Unlike crypto where the deficiencies are quite obvious (and indeed fundamentally a part of the technology), AI’s deficiencies are harder to argue against because the hypemen claim any deficiencies will soon be erradicated by its own automatic, exponential improvement.

However, there is more and more research by more and more people showing that AI is likely plateauing right now. And if that’s the case, then oof, it really wasn’t worth much. AI coding assistants (the flavor of AI I’m most familiar with) are certainly powerful and useful in some cases, but they are not what I would call a transformative technology. My daily work schedule and process has not changed much since I’ve adopted AI into my coding experience, and I wouldn’t say my productivity has changed much in any absolute terms. The one tangible benefit I can point to: I probably type less these days, which might be good for long-term nerve health.

I don’t write this to argue that AI isn’t important or won’t be seen as a transformative technology, but rather as a reflection on what has felt like a wild few years riding two back-to-back hype waves. I remember the relief when the crypto hype wave finally died down, and I’m starting to think we’re headed that way with AI. Here’s to hoping.

Here are some recent pieces I saw trending on HN which have motivated me to write today

• Future AI bills of $100k/yr per dev (kilocode.ai)
• 84% of developers use AI, yet most don’t trust it! (shiftmag.dev)
• Why LLMs Can’t Really Build Software (zed.dev)

Wow, that’s nuts, the last time I wrote about AI was in 2023 🤯