Working on the innovative of AI is unfortunately expensive. For instance,Along with DeepMind, Google also has Google Brain, Research, and Cloud. And TensorFlow, TPUs, they usually own a few third of all research (in reality, they hold their very own AI conferences).

I also strongly suspect that compute horsepower can be vital (and possibly even sufficient) to achieve AGI. If historical trends are any indication, progress in AI is primarily driven by systems – compute, data, infrastructure. The core algorithms we use today have remained largely unchanged from the ~90s. Not only that, but any algorithmic advances published in a paper somewhere may be almost immediately re-implemented and incorporated. Conversely, algorithmic advances alone are inert without the size to also make them scary.

It seems to me that OpenAI today is burning money and that the funding model cannot reach the size to noticeably compete with Google (an 800B company). If you happen to cannot seriously compete but proceed to do research in open, you may in reality be making things worse and helping them out “free of charge”, because any advances are fairly easy for them to repeat and immediately incorporate, at scale.

A for-profit pivot might create a more sustainable revenue stream over time and would, with the present team, likely usher in a number of investment. Nevertheless, constructing out a product from scratch would steal focus from AI research, it could take a protracted time and it’s unclear if an organization could “catch up” to Google scale, and the investors might exert an excessive amount of pressure within the unsuitable directions.Essentially the most promising option I can consider, as I discussed earlier, could be for OpenAI to connect to Tesla as its money cow. I feel attachments to other large suspects (e.g. Apple? Amazon?) would fail as a result of an incompatible company DNA. Using a rocket analogy, Tesla already built the “first stage” of the rocket with the entire supply chain of Model 3 and its onboard computer and a persistent web connection. The “second stage” could be a full self driving solution based on large-scale neural network training, which OpenAI expertise could significantly help speed up. With a functioning full self-driving solution in ~2-3 years we could sell a number of cars/trucks. If we do that very well, the transportation industry is large enough that we could increase Tesla’s market cap to high O(~100K), and use that revenue to fund the AI work at the suitable scale.

I cannot see anything that has the potential to achieve sustainable Google-scale capital inside a decade.