> I believe this is a very flawed way of thinking that misunderstands why companies actually pay for software services. What you have described is essentially to hire an AI to provide the service. Our current LLM-based systems are not capable of this and probably won't be with the currently used technology. Software services is not the code. The code is frankly the least important part.

I think you are mistaken. You still need to pay someone to host the service, but that someone can be a non profit or coop who operates the open source software for you. For example, Karakeep is open source (https://github.com/karakeep-app/karakeep/) but I pay monthly for it so I don't have to maintain an instance. At any time, I can dump my data, startup the open source software elsewhere, and I'm off to the races. Another example is OpenTofu and OpenBao (part of the Linux Foundation) replacing Hashicorp's Terraform and Vault commercial offerings, if your org has the appetite to operate those open source infrastructure offerings.

If the code is open source, I/we can pay anyone willing to run it to run it, either as an individual or a corporation. I can incorporate a non profit or B Corp to run it. The code belongs to anyone, not an entity to be passed around for future potential cashflows by squeezing the customers who require the software it has ownership of. Certainly, maintenance of the subject software and its operation will still require resources and people, but those consuming the software have the control instead of a for profit entity.

Organize software starting with the most profitable and start building with the software factory. For profit software industry margin is software consumer and user opportunity. Someone will still need to operate the open source outputs of this process, that is the economic opportunity. It is no longer "Who will build this?" The question is now "Who will operate this open code for me?"

(i have both worked as an early employee at a SaaS startup and currently have a budget for buying software and SaaS products in the context of risk management and cybersecurity, and also am encouraged by my org to build solutions using AI whenever potentially more advantageous than purchasing a product or SaaS solution, so I am very familiar with this topic)

Edit: Similar vibes to https://huddleup.substack.com/p/the-savannah-bananas-just-co... | https://x.com/JoePompliano/status/2085007330748383441

> The Savannah Bananas spent hundreds of thousands of dollars building their own secondary ticketing platform.

> No fees, every user is verified, tickets can only be sold for face value

> In just 5 months, it's already saved fans $13.69M.