Unleashing AI’s Full Potential
When giants dominate the innovation scene, the startup ecosystem gets disrupted
Image from Katerina Kambrani, symbolizing how a startup outside the enterprise and the data inside an enterprise need to be linked up.
I recently had the opportunity to participate in the Catalog & Cocktails podcast. I must say that the way the hosts managed the flow of conversation and the great questions that led to good nuggets being distilled, was refreshing. A big thank you to Tim Gasper and Juan Sequeda for hosting.
Some days later Giri Venkatesan posted a summary of the discussion on LinkedIn, which highlighted some of the key points I was making - thank you! I feel very passionate about the true potential of AI (I have for decades!) and can’t stop thinking about how to help the world become better with the help of AI. So I’d thought I’d double click on this again in a new blog post. Maybe we can co-solve this challenge together?
OpenAI enabled the general population to see the power of AI directly applied. AI’s potential has been clear to AI Scientists and Practitioners for quite some time (decades to be fair), but perhaps unfairly hidden away in labs or Data Science teams in large Enterprises that can afford the training and talent costs. Similar to what Edison did for electricity, yes? Everybody thinks he invented the light bulb, but there were actually variants before that, mostly accessible in labs and managed by experts, as electricity was dangerous and unsafe to handle for novice people. Then came Edison, who envisioned a safe way to put it in the hands of everyone - in homes, in offices, in workspaces, in stores. He came up with an affordable system - a cheaper lightbulb, with better quality results, and most importantly a focus on the system over incremental technology improvements - i.e. the switch on the wall, the integration with existing gas distribution etc. Suddenly people everywhere could use it. That is when the real magic happened and eyes were opened to the possibilities. I think Edison is a perfect analogy for what OpenAI started out as.
The unfortunate side note on this is that OpenAI, Google, Microsoft et al. somewhat hijacked a lot of the innovation space that normally serves the startup ecosystem. When companies like these build a LLM over public data with enormous resources and sufficient backing ($B) to cover all the expensive AI training required, then there is not much space left for starving startups who were on their way to serve greatness, built on the same technology and innovation ideas. Consolidation started to happen before the race was even run. Building something over public data is just not, in the long run, very competitive - but more of a first to market game - especially as giants can “move mountains” over just a few years by shifting budgets.
With the recent “app-store” move by OpenAI, we can now probably also sit back and watch the whole app-explosion wave all over again. This may look similar to what iOS enabled (and later perhaps Android to some extent) for mobile apps back in the day (or Windows and Linux, if you go even further back). If you're an entrepreneur who wants to ride this wave and is not afraid to fight hard for space around “your app”, by all means go build an app! And don’t forget to allocate a huge marketing budget and optimize for OpenAI app search engines... I tell ya, it's going to be the application race all over again, that we already know, if we lived through the mobile era.
Unfortunately, (I say “unfortunately” as I know there will be some happy Millionaires in the app space and I am sure they would be good opportunities for a VC to work with) this doesn't excite me as much as the other bucket of what GenAI and Traditional DL/NLP/CV can accomplish for enterprises.
The untapped potential of AI (or Machine Learning as I insist calling it) remains abundant. . The transformative impact AI could have on the vast amounts of data related to processes, workflows, and industry-specific blueprints that enterprises possess is enormous.
My message to the world and what I have been trying to accomplish here at DNX over some years now, is to establish trusted and close collaboration between enterprises and startups. As that is the only path I see towards unlocking AI’s full potential for industries. There have to be new ways to foster synergistic partnerships between startups (for their nimble and innovative nature) and enterprises (i.e. the data goldmines). This collaboration could significantly amplify AI's transformative benefits.
As Giri so well put words to my core intent: “Such a collaboration could propel us into the Innovation 3.0 cycle [around AI], positioning it alongside other revolutionary shifts like the web, mobile technology, and now, the transformative AI revolution."
How do you see startups get more access to unique data? How can we foster more willingness and confidence in enterprises to co-innovate with startups?
Such a fantastic conversation together, thank you Eva for joining us! You have such great insights on what’s happening and where it’s all going!