My metric is: how much code that I create today do I depend on in 3 months (incl learnings and one off data pulls etc.) and its waaay less than what I generate. A lot of it is frankly noise.

If you have particularly specific knowledge in pretty much any domain, combining that with AI can lead to huge gains.

Yes

> It incentivises high-impact research

It incentivizes work that sounds impressive to laymen. Actual work tends to be technical and might not sound super exciting.

If 20 years ago, a politician had to get up and explain that we were spending millions of dollars training computers to recognize a strawberry, likely the entire field of machine learning would not exist today.

^ I say shouldn’t because I work in research engineering. Most of the needs of our users are pretty unique. We’ve had people come in and try and specify every piece of work, -and ended up building a crud app no one wanted or used.

I’m going to guess that this is Gell-Mann amnesia more than anything, and it’s going to get a lot of organizations into a lot of weird places.

The things that the software does might have value, but the marginal utility of your software is effectively 0.

Which is what I’m seeing at my job. All of these “afternoon vibe code” projects never actually get users because everyone just vibe-codes their own.

I’m being glib, but there’s a whole class of software (eg simple crud apps) that just don’t have any marginal value anymore. So it doesn’t matter if it’s 10X faster or 100X faster. 100 x $0 is still 0.

As an example, if you think there might be any sort of pushback, just never stop talking. Once a manager talked for 35 straight minutes to answer a question on an unpopular decision. By the end there were no follow-ups because everyone was too confused and checked out to care.

1. Everything is specified, written and tested by me, then cleaned up by AI. This is for the core of the application.

2. AI writes the functions, then sets up stub tests for me to write. Here I’ll often rewrite the functions as they often don’t do what I want, or do too much. I just find it gets rid of a lot of boilerplate to do things this way.

3. AI does everything. This is for experiments or parts of an application that I am perfectly willing to delete. About 70% of the time I do end up deleting these parts. I don’t allow it to touch 1 or 2.

Of course this requires that the architecture is setup in a way where this is possible. But I find it pretty nice.

I asked Claude (Opus High Effort) and pasted in all the logs. I went back and forth and it very confidently made over 20 separate changes in the repo, none of which fixed the issue. Eventually I stepped in and figured out it was a versioning issue.

I fear what would happen if I ran “10 agents for 10 days” on this simple issue.