Post

2. enterprise rewrites of already-working software, to get some enterprise-desirable property, that isn't immediately visible and therefore would've previously been unaffordable 3. software made by non-programmers (academics, hobbyists) that couldn't get programmer resources before

12/9/26, 3:50 pm2 reposts55 likes

4. rewrites and ports of legacy or even abandonware software that would've never happened otherwise 5. algoslop, like "claude make me a profitable saas app on the app store". exists but it's actually quite rare to *see* it because nobody's sharing this stuff

what I haven't seen as much of, is rapid buildouts to prove out new end to end research programs in software that unlock truly new paradigms in architecture, development, or deployment. that *aren't* just new Gastowns

another one i've wanted to build for a while is a compute environment based on hot reloaded wasm modules with fully qualified namespacing/content-addressing so you can have N branches of any given function or module within the binary and do things like parallel evaluation and shadowing

a few companies have tried something like this and they've all had to seriously cut down scope a human trying to build this by themselves would go insane but I don't think it would even be especially _hard_ for like $1mil tokens at this point

build seems to be an area ripe for total disruption in this space. imagine a Bazel-like where all components are distributed, mesh-based global caching, worktree oriented, agentic integration - that would've been an intimidating years long buildout. now it seems approachable for a small team

historically it's been hard for any meta-package-manager to keep up with all supported OSs and PLs, which are constantly changing but OSes and PLs won't be able to move anywhere near as fast as small tools/teams now can due to needing to keep a stable interface so that seems like a huge inversion

Jack Schumann

@jackschumann.com

14 Sept

Im on it 🫡 @meetcassette.com

@notrimskiy.bsky.social

Also anecdotal, but I have been seeing that some companies evolve internal systems for better support of automation with human in the loop processes. This does show up in earnings reports when especially promising-looking, so not exactly secret.

Benjamin Boorowski 👻

@typeoneerror.com