Recently, I was listening to an awesome podcast with Sam Altman talking about how he suspects the costs to train larger GPT models in the future will likely exceed a billion dollars (!!!).
At the same time, I also heard about this cool email enhancer tool that OpenAI is testing. It’s clear to me that they are looking for interesting ways and use cases to commercialize GPT-3’s API and email was one of the big ones circulating heavily on Twitter.
Through the email enhancer, GPT-3 can help you write pretty nuanced emails and save you time you would have otherwise wasted in your email inbox.
Where could this monetization attempt go? That’s anybody’s guess. I could see it difficult to turn profitable at the end of the day, or even something the email clients themselves implement on their own (Gmail already has a version of something like this).
However, there’s an exciting possibility here … in the same way SpaceX has always wanted to use Starlink and offer global satellite based internet as a way of funding their mission to Mars, I think OpenAI could potentially fund future models of GPT through a potential mass-market consumer email service.
It sounds crazy, but imagine paying $10 a month to have your email managed for you through GPT-3. People are already $99/yr paying to use Hey! and I still don’t fully get what that even is. GPT-3 could actually be helpful in this case and there are many business users who spend most of their lives (a sad one) just answering emails.
OpenAI could go lots of ways, but I like this direction because it would be better for them to make money through something SaaS based and predictable like email, than to make a bunch of things and still have to go hat in hand and raise more VC funding with its own strings attached. The more OpenAI can stay independent, especially financially, the more room they have to do what’s right. Which is good for everybody involved.