I’m excited to announce the launch of the Gold Dragon Ventures website! The focus of the site is right on the home page - all of the startups that I’ve backed, whether as an angel investor, AngelList syndicate lead, or via Asymmetry Ventures.
I’ll be adding news and community features later on, but I wanted to give my Substack readers first look at all of these incredibly innovative companies.
In other news, the opportunity fund that I mentioned in January is officially on hold, because I’ve been busy helping a partner with the planning for a $30M fund called Victorious Ventures. The fund will back seed stage startups in emerging markets with a focus on Africa. More on that coming soon when that website is ready.
In keeping with past tradition, I have a short follow-up on the generative AI thoughts from my Jan post:
Elad Gil has a brilliant post that distinguishes various technologies being lumped into the generative AI bucket, and walks through the underlying economics.
Cognitive Revolution dropped a detailed competitive analysis of the foundational models that compete with OpenAI/ChatGPT, by Nathan Labenz, co-founder of Waymark (AI video creator).
Nathan is “talking his book”, but even with the rosy outlook he proves my earlier point - the price of the results for the type of advanced prompt engineering that really got people excited on social media was far too high for mass market. Even at his 5 cents per day per user estimate for OpenAI’s new API, I’d debate whether the results for typical searches are superior to the info from top sites on a Google results page.
Don’t get me wrong - ChatGPT is still a big deal. I expect OpenAI and Microsoft will find many ways to integrate GPT with products that drive increased productivity without raising prices. And OpenAI cutting its API prices by 90% for a slightly inferior version of its competitors’ service is a savvy land grab by way of innovator’s dilemma theory. It’s already spawning tons of new services targeted to underserved niche sectors. But I expect most of the value capture here will flow to incumbents like Microsoft who can bundle with other offerings, or to startups serving a vertical in a way that Google never could in the first place.