The Future of VC and Tech
Here are my purely qualitative thoughts (no data backing them) and I would be so interested to hear people’s opinion on how right or wrong I am and why:
-
Chatgpt has shown that the main driver for growth in the tech sector will be AI in the next 10 years. What Chatgpt has also shown is that narrow AI is yesteryear. The public will only settle for general AI- one software/ platform to rule them all. The trouble is there will only be one, maximum two winners in this entire sector as the first company to inch ever closer to AGI will rule everything. After that any other AGI platforms will be nought but imitations for which the public will not care less.
-
The lifespan of these AI companies in the race for AGI will be measured in short years given the rate of progress and so this sector will be a total blood bath. We already know most startups fail but the AI sector is going to be setting records in failed startups. The public only has the attention span for 1 or 2 general AI platforms that they may use in the daily life. Everything else will offer no unique value proposition. We are seeing early examples of what happens when an inferior AI is set into competition with a competent one (Bard vs Chatgpt- nobody is using Google’s Bard).
-
If the premise is correct that the future will be a race towards AGI then the era of overnight app millionaires, software millionaires etc is surely over. Why would people rely on using heuristics and algorithm based archaic software platforms when one AI platform is intelligent enough to do everything for you? Even if one argues the value of apps/ software will be to provide a UI for the AI to work from, one cannot deny the value proposition is in the AI itself and not the app offering a UI for ordering say taxis.
-
Therefore the focus for investment will shift towards robotics and devices. Such devices will be integrated with AI and provide fascinating use cases in industry, housework, deliveries etc. When Musk flaunts his Teslabot many of roll our eyes but actually he seems to have anticipated what industries a new AI world will drive. The downside is that these robotics/ devices companies are far slower to develop and less easy to scale as businesses compared to software that has driven the tech boom.
-
This all leads me to wonder whether the massive gains made by VC in the tech space are coming to an end. Those high VC salaries and great carries are about to dwindle. Perhaps a VC firm out there will get lucky on betting on the company that figures out AGI or the next skynet but the amount of losses firms will incur in trying to find that diamond will render the search futile. As I say I would hazard a guess by saying this sector is going to be a bloodbath.
-
so…is the era of an abundance of VC tech multi millionaires about to come to an end? Are tech VC returns going to be as large as they have been in the last 20 years? Is software tech VC no longer the sector to be focusing on?
Did you try and fit as many bad and uninformed takes into one post as you possibly could?
I love you Liam. I even gave you a banana.
Hey what's up, I know an ok amount about gen AI, first thought is on this:
Definitely not necessarily true. Personally I think it shakes out a little more nuanced than that. Different models have different strengths. It's fairly common to chain together different models, or plug the gaps with things developed internally, etc. You might have one that wins for personal consumption e.g., ChatGPT occupies a Google-like position in the advertising value chain eventually or maybe maintains a subscription/consumption model to become profitable, but definitely not the case. Here's an interesting article I like, it's not perfect, I probably agree with 2/3 of it, but things like user reinforcement, proprietary user data are important for fine tuning off the shelf models, but it'll always help to have different models.
Then going from that it's not necessarily true that the traditional paradigms of value accrual will hold. What makes a winner? Is it using the best model, or do models in fact become commoditized if you are right? Is it user data, is it best feedback loops, is it best tech stacks? Is it internal development vs. model hosting open source vs. using APIs of closed source models? Which gives you most defensible margins, lowest CAC, best retention?
Robotics is probably about 10-15 years behind what you think of as the maturity in AGI right now. A lot of the robotics advancement still occurs in academia unlike gen AI which is now about 80-90% done in the enterprise. You'll still see a lot of activity in this space as both academic and commercial movers are making progress and people with deep coffers start to take note (e.g., Hyundai acquiring Boston Dynamics).
I think you should read about the differences in model types. It's not just language models, there are plenty of other types of models. I think it might more likely shake out like the database market where you have 10-15 big ones that do different things and some are the Oracles of the world but there's a big play for lots of them with continued innovation.
Let me know what you think. This was a quick skim and response (<2-3 mins) so happy to dig into it further if you want a good discussion.
Est ut natus aut est repellat sit. Adipisci voluptas dolor voluptas eum aut. Et enim consectetur sapiente consequuntur.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...