Yeah, I actually agree with him and not in a loud or defensive way. It’s more like… if you’ve really used these tools in real work, you eventually see what he’s talking about.
AI is useful. No doubt about it. It speeds things up, helps you move faster, and can take care of a lot of repetitive work. But once you go beyond experimenting and start building real products people depend on, you notice something else pretty quickly.
It’s fast… but it’s not careful.
You still have to review everything. You still have to think through decisions. You still have to clean things up. And if you don’t, the codebase slowly becomes something you don’t fully trust anymore.
That line about it being like a junior dev you have to watch closely that’s honestly spot on. It gets things done, but it doesn’t understand consequences the way you do.
And that’s where a lot of the hype feels disconnected from reality.
People keep focusing on how fast code can be written now. But after a certain point, speed isn’t the real bottleneck. It’s judgment. It’s clarity. It’s knowing what to build, what to simplify, and what to leave alone.
If anything, AI just exposed that.
That’s actually why I wrote The Vibe Coder, which is now live on Amazon. Not to argue against AI, but to talk about what happens when coding becomes easyand thinking becomes the real work.
Because once generation is no longer the hard part, you’re left with better questions:
Are you making good decisions?
Do you understand the trade-offs?
Are you taking ownership of what gets shipped?
Or are you just moving fast because the tools let you?
I think that’s the part a lot of people are skipping right now.
AI didn’t replace developers. It just made it obvious that typing was never the hardest part.
The vibe coder- https://t.co/dwRDOLcvTx
@Seq_Protocol Since funds are already using Sequoia to manage downside risk, what kind of institutional strategies are currently being executed on the protocol and how do they differ from what retail traders will be able to do?
@Seq_Protocol Since funds are already using Sequoia to manage downside risk, what kind of institutional strategies are currently being executed on the protocol and how do they differ from what retail traders will be able to do?