I also strongly doubt that computing power is necessary (and probably sufficient) to achieve AGI. If past trends are any indication, advances in AI are primarily driven by systems (compute, data, infrastructure). The core algorithms we use today haven’t changed much since the 90s. Not only that, but an algorithmic advance published in a paper somewhere can be reimplemented and incorporated almost instantly. Conversely, algorithmic advances alone are inert without the scale to frighten them.
It seems to me that OpenAI today is wasting cash and its funding model won’t allow it to reach the scale to seriously compete with Google (an 80 billion company). If you can’t seriously compete and still do research in the public domain, you may actually be making the situation worse and helping them “for free.” Because advances are very easy to copy and quickly incorporate at scale.
A for-profit pivot could create a more sustainable revenue stream over time, and with your current team, could result in a significant investment. But building a product from scratch takes away the focus of AI research, takes a long time, it’s unclear whether companies can “catch up” to his Google’s scale, and investors are on the wrong track. You may apply too much pressure. The most likely option I can think of is for OpenAI to belong to Tesla as a cash cow, as I mentioned earlier. I think attachments to other big suspects (Apple, Amazon, etc.) will fail because their companies’ DNA is not compatible. Using a rocket analogy, Tesla has already built the rocket’s “first stage,” which includes the entire Model 3 supply chain, its onboard computer, and persistent internet connectivity. “Stage 2” is a full self-driving solution based on large-scale neural network training, and OpenAI’s expertise could greatly help accelerate it. If fully self-driving solutions work, many cars and trucks will be on sale within two to three years. If he does this really well, the transportation industry is big enough that he could drive Tesla’s market cap to a high level of O(~100,000) and use the proceeds to fund his appropriately sized AI development. I can.
I don’t see anything else that has the potential to reach sustainable Google-sized capital within 10 years.