AI's Ability to Forecast the Future
I'm still blown away by the power of massive data analytics that these AI models can take on. The idea that the most accurate 10-day forecasts are now coming out of AI models is mind-boggling. Of course, that's what AI does—it takes a seemingly massive amount of uncorrelated data and makes interpolations and extrapolations.
A recent paper published in Nature revealed that a new AI-powered system for forecasting weather and predicting future climate achieves results comparable with the best existing models while using much less computer power. The implications? Enormous "computational savings" and the potential to "enhance the large-scale physical simulations that are essential for understanding and predicting the Earth system."
But here's where it gets really interesting: AI being able to predict earthquakes, stock markets, and a whole slew of different things. The ripple effects are immense.
Salim shared a fascinating story about the real-world implications of accurate forecasting:
"You may have heard me speak about the car washes in Buenos Aires," he said. "A 50% drop in revenues over a 20-year period just because we were better able to model when it's going to rain. When you know it's going to rain, you don't wash your car. That's just a huge impact on an industry. And the key there is you can be the smartest car wash owner in the world and you'll just never see that coming."
This is the orthogonal effect of technology on domain after domain. Better or worse, it's going to change everything.
As an entrepreneur, you need to be asking yourself: If I could know anything or predict anything, how would it impact my business? Whether you're in fashion, car washing, or gardening, the challenge—and opportunity—is in asking, "What would I want to know? And if I could know that, how would it give me different economics, a different business model?"
Defaulting to Disbelief – The Dissolution of Truth
Check out this video here or a photo realistic women generated with Flux with Lora + Gen-3 Alpha image-to-video. It is indistinguishable from reality. For reference:
- Flux: Refers to a specific software or framework used for machine learning and AI tasks, especially those related to generative models.
- Lora: This is a specific type of machine learning technique that allows for fine-tuning pre-trained models on smaller datasets. It means that the model is used for image-to-video generation.
- Gen-3 Alpha image-to-video: The core technology used to convert images into videos is a third-generation AI model specifically designed for this task.
Salim's reaction? "We've now crossed that Rubicon, and now we need to navigate a future that is going to be really, really crazy."
On the positive side, there is a benefit where we can now use this technology to bring back our loved ones—those who have passed. Important Note: If your mom, dad or grandparents still with you, take the time to record them on your iPhone – record their voice, sit 'em down for two or three hours, give 'em a glass of wine and get them to talk whimsically about their history. You can train up an AI model that's going to be with your family line for the next century.
A Bombshell on AI Superintelligence
What happens if someone actually wins the race to achieve AI superintelligence?
Vitalik Buterin, the crypto genius who gave us Ethereum, just dropped a truth bomb that's got the tech world buzzing:
"If you imagine every AI growing exponentially, existing ratios of power get preserved. But if it's growing super exponentially, if you're a little bit ahead, the ratio of the lead starts increasing. The worst case is a step function. Whoever first discovers some magic leap - nanotechnology, increased compute, algorithmic improvements - would immediately turn on that improvement, quickly expand, find all other possible improvements before anyone else, and take over everything in an unknown and unpredictable environment."
This "step function" could be anything from cracking the nanotechnology code to a compute power explosion. Whatever it is, it's the golden ticket. The AI that grabs it first doesn't just win—it takes over everything. We're talking world domination, but with silicon and algorithms instead of armies.
Why does this matter?
Well, picture this: OpenAI, Google, China, and Anthropic are all in this “AI arms race.” If one of them stumbles upon this magic leap, it's not just about bragging rights. It's about reshaping our entire reality. The implications are staggering—we're talking about an intelligence that could solve climate change, cure diseases, or... less optimistically, decide humans are an inconvenience.
And this isn't some far-off sci-fi scenario. It's happening now, in labs and data centers around the world. Every breakthrough, every new model, could be the one that tips the scales.
We're not just spectators in this race—we're the track it's being run on. As entrepreneurs, innovators, or just plain citizens, we need to be asking the tough questions. How do we ensure this leap forward benefits humanity? Can we create safeguards? Or are we unleashing forces beyond our control?
Eric Schmidt on “What Failed at Google”
I've got to imagine that in the executive conversations at Google, they're like, "WTF is going on here? How did we possibly get into this situation?" Well, here's what Eric Schmidt thinks made them second tier to OpenAI:
"Google decided that work-life balance and going home early and working from home was more important than winning. And the startups, the reason startups work is because the people work like hell. And I'm sorry to be so blunt, but the fact of the matter is, if you all leave the university and go found a company, you're not going to let people work from home and only come in one day a week."
Wow. Pretty damn spot on of Eric to say that.
The companies I've invested in, the companies I've built—when you're there in the thick of it, physically all together 7 days a week, you get something special. I just visited the drone delivery company Zipline and did a podcast with their CEO. When I was physically there, the energy of hundreds of people literally crammed into this office with dashboards up on the walls... you don't get that virtually. You don't get that kind of energy, that kind of drive. It's a new organism that comes out.
But is remote work the only thing holding Google back?
Here’s Salim: “It comes down to organizational design … in Google today, to try and get anything done, there's so much bureaucracy, it's taking forever. That's the thing that's slowing it down."
I discussed this with Eric when he joined me at the Abundance Summit last year. Google previously had a massive lead in AI but was being careful and didn’t put things out before they were fully tested. And when Sam Altman came out with ChatGPT, it was like, “Okay, gloves are off, put on the racing sneakers, let's go."
And Google has responded incredibly well. For example, Gemini has some amazing capabilities.
But as Salim notes, there’s a lesson here for all entrepreneurs and CEOs:
“Eric is commenting on the organizational structure of Google versus OpenAI being comprised of small teams that are just doing whatever the hell they want … It's the curse of being a big company. The control frameworks you have to put in place prevent you from being nimble. And today the name of the game is being nimble."
I agree with Salim: Agility always wins.
