The Bigger Picture, November 12, 2023
In last week’s ‘The Bigger Picture’, “Going Small to Go Big”, I made the point that we’ll likely see big things in this AI Tech Wave at both end of the AI infrastructure spectrum, the Big Foundation LLM AI models being built by over a dozen public and private companies, and the Small devices that are always at hand every day for billions of us. This week I’d like to do a follow up focus on the other Bigger Picture on smaller AI efforts around platforms, applications and services, that sometimes surprise the most. OpenAI’s ChatGPT, less than a year old still, started out like that as a back-room experiment in ‘reinforcement learning loops’, vs the company’s grander, multi-year, GPT LLM AI models. They may seem far less ambitious in the beginning than the big dollar AI efforts sometimes in the same company. They often start small, sometimes as features. Then become massive platforms seemingly overnight, for applications and services. Boiling the perfect cup of tea while trying to boil the ocean. Let me explain.
All around us are examples of companies that today have made substantial dents in the lives of hundreds of millions, if not billions of people, that started doing simple things that became world-changing platforms. One example that comes to mind is from a pithy quote by a prominent VC a few years ago:
“We wanted flying cars, instead we got 140 characters.”
The reference of course was to Twitter, now X, that ended up changing the world in how we communicate in short bursts of texts, that most importantly, could be CONSUMED in short bursts of attention by many receivers, shifting the focus from the sender. Another company that has built a world-changing business around this same idea is of course TikTok. They did the same with ‘short videos’ that now occupy over two billion souls on that service every day, for sometimes an hour a day or more. Videos from 30 seconds to 10 minutes, are now a business opportunity pursued by most of the other US big tech companies like Google YouTube, Meta, and others. And they’re already generating billions in revenues.
So small, prosaic features turn out to allow billions to fly around with ideas at Scale, far easier than former ways of human communication. And we’re still waiting for Flying Cars, despite the best efforts of some of our most successful founders to fund efforts to re-create the ‘Kittyhawk’ moment with personal flying vehicles.
Most of our best VCs today want to:
“Invest in smart people solving hard problems ‘that really have the potential to change the world.’”
Laudable goal indeed. But often much smaller efforts get to that place sooner with boiling smaller pots of water rather than boiling the ocean. Think Facebook with a humble college yearbook online, Snapchat with ‘ephemeral posts’ that disappear after being read, videos by YouTube that could be shared by and to everyone with a simple URL, text messages that could be sent from and to anyone via most proprietary phone networks around the world (WhatsApp), low-quality smartphone photos that could be livened up by digital filters (Instagram), and so many more. Billions of value created by serving billions with prosaic efforts to have the seemingly smaller problems set to boil.
The reason I bring this all up is that we’re at the earliest stage of the ‘AI Tech Wave’, and all around us are entrepreneurs and investors pouring hundreds of billions in efforts to ‘boil the ocean’ in so many areas around ‘AI’. And that’s a good thing.
But often these investments are incentivized just by a competitive peer doing something, and everyone following with billions in the pot. Or as I highlighted in a recent post, driven by what’s worked in previous big tech cycles like applications dominating a platform, or an app store that took the lion’s share of the economics around a smartphone.
Those larger opportunities are definitely there this time around too. But often the bigger opportunities are where the elephants are not trundling around at a fast pace. But they often take longer periods of time to make real. While, smaller, more humble efforts lead to larger, far more impactful creators of value in so many forms. It’s a point I’ve made in other posts here, here, here, and here. A point close to my heart.
But it’s a point remembering again after OpenAI’s highly anticipated inaugural Developer Day (aka ‘DevDay) this past week, that was a starting gun to Foundation LLM AIs going from a centralized model to personalized, ‘smart agents’ driven models, and efforts to kick-start an AI app store. As I mentioned here and here, these are events that will again get the elephants moving in new directions chasing old dreams in new ways.
But also keep an eye out for founders be they large and small boiling smaller pots of water and make a great cup of tea. They may come up with something that dents the world in unexpectedly good ways ahead. While we wait for oceans to be boiled. Stay tuned.
(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here)
“Invest in smart people solving hard problems ‘that really have the potential to change the world.’”
Really? AFAIK, VCs are in it make monet with high risk, high returns. Given the sorts of companies many have invested in, their ideas of "changing the world" are not aligned with the pressing problems we have.
Having said that, AI could help solve some pressing problems. Google's Deep Mind has perhaps been best at that with its protein folding AI that has now populated the protein database with their work and made a difference in understanding their functions and how to manipulate them. There are other computationally heavy tasks that AIs have streamlined, such as orbital perturbations and powered escape orbits. There are many NP hard problems that might be provided with very good AI solutions, that would be of benefit. Of particular interest would be AIs that could design genetic manipulation of organisms to change metabolism. Improving carbon fixation would be a huge boon for food production. It should also be obvious that similar models would streamline new pharmacological treatments as Big Pharma continues to struggle towardsi ts inevitable dead end.
Flying cars weren't a particulalry good idea (and that includes taxi drones)., but Twitter/X was certainly not a good substitute for good ideas, and the monetizing model has resulted in its enshittification being accelerated by its apparently clueless current owner.