The way I read it TFA is claiming that the Devs who built the Natascha AI suite did solid work.

Contrary to prior reporting there really was an LLM based product at Builder in addition to their earlier business of a large pool of contractors cranking out apps on demand.

Is it? Do you think it's a coincidence that this scheme (whatever it truly was) somehow "worked" from 2019 and happened to "bust" right when things like bolt & all hit the market? Excluding the management fraud with declaring more contracts, this thing produced solutions for 5 years. Now similar solutions can be 0-shot with pretty much the same consistency and quality. Sure, the quality is shoddy at best and there are huge problems (security, scalability, maintainability, etc.) but the happy path usually works.

Remember, we're just 2.5 years post chatgpt and now we have 0-shot abilities similar to what a company employing 700 devs used to provide. I think that's huge. This very much fits the overhyped in the short term (for the problems listed above) and underhyped long term (if this is where we're at 2.5 years later, where will we be in 5-10?).