I won't use an AI browser (yet)

Usually, I love trying new tech products. But this month, I've avoided the shiniest new AI Agentic browsers: Perplexity's Comet and OpenAI's ChatGPT Atlas.

I'm sure they are great products. But something about them doesn't sit right with me.

So I asked myself: Why?

Upon introspection, here are 3 things I discovered:

1. Trust
   I trust that an AI Agent with safeguards will do the right thing 99.9% of the time. But that 0.1% could be catastrophic (what if it shares my banking password with someone?).

2. Confidence
   Even if I write out exactly what I want, I'm not sure the AI will actually get it. Half the time I don't know what I want until I'm halfway through doing it. Preferences emerge. It's impossible to state them all upfront.

3. The Joy of Doing
   Sometimes, the process is the fun part. I like researching vacation spots. I like the thrill I get from getting a good deal on a flight by using a random hack I read on X (VPN to a 3rd world country + book on Tuesday night). I like following my curiosity down random rabbit holes. I don't want to give the AI that pleasure.

---

Will agentic browsers take over anyway? Probably. There are too many smart people working on them. But they'll need to earn trust slowly, or change their form entirely.

For now, I'll pass.