If AI Makes Us Faster, Why Aren't the Apps We Use Getting Better?

This week a Jonathan Blow video titled "If AI solved coding, why is my software still the same?" landed on the Hacker News front page and pulled more than 150 comments.

The question stuck with me because I don't have a good answer.

If AI really makes developers several times faster, I'd expect to notice it as a user by now. Fewer bugs. Faster apps. Features I've wanted for years finally showing up. But the apps on my phone feel about the same as they did two years ago. Some feel worse.

A few possible explanations I keep going back and forth on:

• Maybe we aren't actually faster. METR ran a randomized study in 2025 with 16 experienced open-source developers on 246 real tasks. With AI tools allowed, they took 19% longer, while believing they'd been about 20% faster. That's one study with early-2025 tools, but the gap between feeling fast and being fast is interesting.
• Maybe the speed goes into more code, not better code. If writing code is cheap, teams ship more features, not fewer bugs.
• Maybe coding was never the bottleneck. Deciding what to build, reviewing it and getting it approved take the same time they always did.
• Maybe it's too early. Better software might just take longer to show up than faster typing does.

Personally, I feel faster on small, boring tasks and not much faster on anything that needs real thinking. I'm honestly not sure that adds up to better software for the people using it.

So, what do you think?

• Can you name one app you use that clearly got better because its team uses AI?
• If you feel faster with AI, where does the saved time actually go?
• Is "coding was never the bottleneck" true where you work?

One-line answers welcome. I'm especially curious to hear from people who think the improvement is real and I'm just not seeing it.