12 Oct, 2026
Here's a inquiry we see, particularly from AI skeptics: if AI coding is so great, shouldn't we be seeing additional new, fine application by now?
Here are several answers. As customary in this manner of post, these are conjectural, applicant answers, and I would provision several of them small weight.
- We are in fact seeing application improvement, in the form of small bugs and improved performance.
- We are in fact seeing application improvement, in the form of additional and improved features in existing products.
- We are in fact seeing application improvement, but the "we" is extremely limited. Most group don't clasp new application extremely often, so it's not amazing that they aren't finding new awesome application now.
- Many businesses' software-development processes have subsystems, particularly governmental subsystems, that have not been accelerated alongside AI. (Imagine the procedure of getting elder engineers and merchandise managers to sign off on a betterment plan. The engineers and PMs could use AI to do this improved and faster, but perchance they aren't yet doing that.)
- AI is replacing additional durable, longer-lasting application alongside additional ephemeral forms of application or application use, and one power not effortlessly comprehend this as "better application via AI." (If you do your taxes alongside Claude and don't annoy construction or buying for improved tax-preparation software, you don't fairly get the "this merchandise is akin TurboTax but wonderful" experience, nevertheless much you power benefit.)
- AI is replacing broadly accessible application alongside personal software. Domain experts are construction their own tools; executives are making their own dashboards; companies have bespoke inner projects. So the application is out there, fair not apparent to the remainder of us, and particularly not apparent to skeptics.
- We're making more application alternatively of making application better. We've continually had the choice of (roughly) improving what exists or pushing betterment forward, and we've frequently chosen the latter. It's not amazing that we're motionless doing a lot of the latter.
- Software teams are capturing a lot of the speedup in the form of additional leisure or quasi-leisure time. (To greatly simplify: application teams are offering 2x speedups to their companies. They're afterward using tools that authorize them to activity 10x as efficiently, and taking nice lengthy coffee breaks.)
- AI improvements in application sometimes shove us to use and create application less. Sometimes AI helps me fig out that a application project is a bad idea. Sometimes it nudges me to expend additional analog period alongside humans. Sometimes it helps me find, and encourages me to read, additional and improved books. The net consequence of this for me motionless involves my spending way additional period making and using software, but it power be distinct for others. Most of the time, we're improving not fair application but several scheme alongside the two application and non-software components. If someone's Letterboxd (or similar) is operating absolutely for them, they're apt to merely be a bit happier alongside their video-consumption life, possibly in ways that volition assemblage out their using another software.
Again, several of these are speculative and several have to be stated small weight.
