Project Progress

Achieved

I worked on this project for roughly a year in 2023. Features that I build in this time include

  • a pipeline for AI based picture generation. And an App prototype, including explicit content after age verification.

  • a recommendation system to show you the content you like. The recommendation system considers hair color, ethnicity, clothing, breast size etc. And can also be influenced by setting user preferences to allow the algorithm to zoom in quicker on what you like. Additionally, I have

  • well integrated ads that you will barely notice (disabled) and

  • a platform for the community with Discord!

Post-Mortem

There hasn't been an update in a while and I shifted my focus to other projects a while ago. At this point, I have to be honest to you and myself: This project is dead. Time for a post-mortem.

What are the main reasons for failure?

The biggest nail in the coffin is, without a doubt, the lack of revenue. While I am a very sex-positive person and like consuming and producing pornography, the main incentive has been to build something that earns money. I have not managed to do that within the time frame I set myself and the opportunity cost is too high. With a PhD in CS there are plenty of job opportunities for me.

The second problem is what I would describe as regulatory threads. Germany has surprisingly restrictive laws when it comes to pornography and child protection. I like the solution I have come up with, but ultimately it is a competitive disadvantage. Additionally, generative AI is a probabilistic process and thus can produce unintended results. As an example, I can build a toolchain generating pictures with bare breasts but no genitals with high probability. But this means there is still a probability that pictures show up in the age-unrestricted area that should not.

What could I have done differently?

An obvious solution to avoid regulatory problems would have been to create cooperation abroad, but this step is too unethical for me.

For the lack of revenue, there are two things that I could/ should have done differently. Have more risk capital. The risk capital I was willing to spend right away was my cost of living for 12 months. The problem is that I did not set aside additional risk capital, even though I would have had capital available. I did spend additional money, but it was mostly on things that still have value if the project fails (e.g. an AI-ready PC and the domain thepornai.com) or was low cost (e.g. server upkeep). I was way too hesitant to spend money on other things. Especially design and advertisement, which I noticed is something that I don't like doing myself.

The second mistake was not focusing enough on creating a minimal viable product (MVP) and instead spending too much time chasing the wrong trail: I underestimated the progress of generative AI (multiple times) and concentrated on incorporating 3D renders into the generation process. However, with the announcement of Sora from OpenAI it has become clear that generative AI does not need explicit 3D models to produce consistent video, to just mention the most obvious.

What are my lessons learned?

Nonetheless, I think this project was a great success for me. It was my first attempt at something like a startup and taught me many small things necessary for a business. It also made me much more comfortable to continue as a freelancer/ self-employed and hopefully continue building something greater from here. Additionally, I have two main takeaways:

I do enjoy chasing technical challenging aspects more than focusing on the "easy" way to make money. A related line of thought is that I would rather compete on technical ability and skill than on raw time spent. This probably makes me a bad startup founder, or at least is something I need to consider and plan for when trying to start a new project.

I had wrong expectations on how easy it is to make money with (generated) porn. Even if what others do does not seem challenging from a technical point of view, there is so much more required to make a successful company. With what I have learned during my project, I can now have much more respect for the work of others and be less jealous of seeing their success. Thoughts of the form "How can he make so much money with this, everyone with a bit of technical expertise could do that!", will probably be things of the past for me.