Post

Simon Willison
@simonwillison.net

I wrote about the most ambitious form of AI-assisted software development I've seen yet - Strong DM's "Software Factory" approach, where two of the guiding principles are "Code must not be written by humans" and "Code must not be reviewed by humans" simonwillison.net/2026/Feb/7/s...

simonwillison.netHow StrongDM’s AI team build serious software without even looking at the codeLast week I hinted at a demo I had seen from a team implementing what Dan Shapiro called the Dark Factory level of AI adoption, where no human even looks …
Carlos Scheidegger
@cscheid.net

Because at that point you’re not *necessarily* going faster than other cheaper human-involved software engineering practices. The 1000/day setup might live in speed/quality/cost Pareto front with strongdm, but almost certainly doesn’t at the equivalent single max plan cost

Carlos Scheidegger
@cscheid.net

1000/day is 365k/year, or about 1 fte at ~180k/yr salary with the simplest adjustment for company costs to hire someone. If you convert token costs to subscriptions, there’s around a 10x subsidy. So you’re comparing 1000/day in token spend to 30 max plans. Your 1max plan idea is 1/30 of a 180k dev

1 repost10 likes

Carlos Scheidegger
@cscheid.net

And that’s at best. If strongdm requires regular token spend, then you get no subsidies and your effective speed is 1/300 of theirs. It’s like ec2 in this way. It only makes financial sense if you have spiky usage patterns and lots of cash. If you’re on a budget, buy the cheapest server and colo