First published in Jul 2024 on LI

The GenAI field right now gives a very similar vibes to last decades’ hype on Self Driving and “Uber for X” startups.

Close to $500B has been invested across startups/GPUs with an end resulting revenue to show of just $10B, majority of which imo is Experimental Revenue (ERR) than ARR.

Some of our VC brothers know about it and even talk about it, yet Greater Fool Theory is still in full play. Keep marking up our investment for next round VC (a certain LLM provider is now 25x valued in just one year ) until we find the Greatest Fool to dump it all on, classic stuff of last decade when SoftBank took such dumps.

I think it’s good to bet on future. But what’s better is to bet on knowing what future is possible. Too much money floating in a space can build a hype for sometime but not a genuine trend that has strong legs.

I think LLMs are quite cool stuff with some def use cases. They have changed most of the natural language processing work.

At a basic level Tranformers are just n gram model on steroids.

However not every cool thing is going to revolutionize economy. Last decade Uber was the coolest thing apparently, and gig economy was heralded as the next frontier of growth and productivity. Yet, I don’t see any tangible impact on economy productivty from these.

The issue I see is in too much focus and obsession with technology than what actually runs the economy. Silicon Valley starts plucking their hair in amazement if they see they can get instant response to their query on an LLM app than having to spend few minutes searching over Google.

Voila! It saved few minutes for me, napkin maths and sketchy extrapolation takes over and boom, we can now save billions of human hours - here comes the BCG slide deck on trillions in economy productivity.

The problem is, you saving few minutes/hrs of time on not having to google or reading documentation to figure out the code, research knowledge base, doesn’t help you go and use that time to now dig a well to find water or build a nuclear fusion technology to generate more energy for society.

Most often you end up spending that time in your consumption of entertainment on social media apps or even outside in nature. Your time saved, does not generally contribute to any productive activity for economy.

For Silicon Valley, it’s the perfect Attention Flywheel. We give you LLMs so you can save time to go back and spend it on our other entertainment sink holes.

The biggest economy productivity we have ever witnessed in society was actually due to fossil fuels.

Instead of having 75% of population growing food 200 years ago, we now have 15% grow food for a far bigger population while allowing others to focus on developing science, technology and other areas for society. That’s real economic productivity.

This link starts to break as ppl move up the Maslow’s Hierarchy. When your food and basic survival is already taken care of, ppl’s desire to be ambitious and “productive” starts to flounder.

← Few things, wish you learned early The End Game on EVs →