My flow (with legacy software) is: manual strip > LLM > manual clean up > repeat

In your mind is 700 files a lot or a little? It feels very small to me, and Xcode really ought to be able to handle that tiny scale on modern machines with ease.

I struggle to imagine a team of more than 10 people writing an iOS app with less than 700 files.

Bear in mind 600 of those files are icon and screenshot variants for various screen dpis and spec ratios..

Translation files, themes, drawables .etc...the list is endless. Even a simple app will easily have a thousand files.

Instagram.app likely has 30,000 files for iOS. And it produces 10-figures of revenue. So how is that wrong?

In fairness, as a mere generator of eyeball time that gets mis-sold* to advertisers, I'd say the FB user experience is very much "doing something wrong".

* dick pills and boob surgery, also government announcements for a country I don't live in, also offers to help renounce a citizenship I never had in the first place

I mean.. say its an enterprise mobile app. Maybe there are 2 shells, each shell has 5 tabs. Each tab might have 5 screens on it.. that's 50 files already just for the screens. Each screen might have various UI components or steppers, etc.

Most noobs, such as those who think 700 files is too many because they've only worked on apps they never published, might just cram everything into that one file.

However, there would be various files for components, functions, etc. Code that's single responsibility and easy to test might mean there are lots of files. There might be upload queues, offline functionality, custom code to go beyond what the ios/android SDKs offer, and so on. DTOs, DAOs, etc. various services..

You probably (won't) get the gist but yeah.

and a thousand lines of code? oh my? whatever shall we do?

So 120B was Horizon Alpha and 20B was Horizon Beta?

Unfortunately not, this model is noticeably worse. I imagine horizon is either gpt 5 nano/mini.

Would highly recommend listening to the latest Hard Fork podcast with Matthew Prince (CEO, Cloudflare): https://www.nytimes.com/2025/08/01/podcasts/hardfork-age-res...

I was skeptical about their gatekeeping efforts at first, but came away with a better appreciation for the problem and their first pass at a solution.

I think the question to ask about your two scenarios: in which is it faster and cheaper to get from v1 to v2? From v2 to v3? I think, for right now, it's cheaper under scenario B. But in the future? Who knows!

Counter-counter-point: handwriting > typing for remembering things (https://www.glamour.com/story/typing-memory)

I haven't seriously used R in nearly a decade but I still miss (and think about) dpylr and the hadleyverse...

A few years ago I made a package called "redframes" that tried to "solve" all of my frustrations with pandas, make data wrangling feel more like R, while retaining all the best bits of Python...

Alas, it never really took off. For those curious: https://github.com/maxhumber/redframes

Agreed! I started out doing data analysis in R… switched to Python because it was more multi-purpose at the time (ie: data engineering, analysis, and model deployment)… and I think about it and miss it so often.

Hey this looks pretty tidy.

There is so much hype and luck to widespread adoption, you never know with these things.

Is this "kingfall"?

No, Kingfall is a separate model which is supposed to deliver slightly better performance, around 2.5% to 5% improvement over this.

Sundar tweeted a lion so it's probably goldmane. Kingfall is probably their deep think model, and they might wait for O3 pro to drop so they can swing back.

As in: "JSON was a huge deal, and this could also be a huge deal" or "Just use JSON"?

Dusting off the game I made during the pandemic and trying to actually grow the player base in earnest.

It's turn-based two-player game sorta like chess, but the pieces take up multiple spaces.

App Store: https://apps.apple.com/ca/app/kingbit/id1565583669