“‘The best way to predict the future is to invent it.’ Alan Kay Today, AI doesn’t just invent , it manufactures the future. Hardware is finally catching up to software.

Max Tegmark said AI could be “life’s last invention.” That is unless we steer it to be our first true collaborator. Factories provide the sandbox. Max Tegmark

Henry Ford built modern mass production, but that scale created rigidity, which was necessary with the tools available. Now, AI-driven factories can adapt in real time with flexibility and precision. Ford’s dream has been upgraded.

I have found that any conversation that I have had about abundance all comes down to false scarcity and this is currently a distribution problem in a very abundant world. I like this by one of the best, William Gibson: “The future is already here — it’s just not evenly

Many are saying “hardware is the bottleneck.” Maybe. But bottlenecks are where real builders live. That’s why I pay attention to Jim Keller and @zelooftron at Atomic Semi . they’re inventing and building , not shouting about rivals.

I hear often about hardware as a limit to AIs future. It is certainly is a challenge, but it is also also an invitation to invent. Jim Keller and @zelooftron at Atomic Semi are an examples of an important way of breaking through that barrier..

I love that David Bowie said” Tomorrow belongs to those who can hear it coming.” What are we listening to? I listen to the hum of ai factories. Certainly not as cool as Bowie, but relevant. Listening to sounds of (((((((577))))))) also.

“If Anyone Builds It Everyone Dies” by Eliezer Yudkowsky ⏹️ & Nate Soares ⏹️ deserves the attention. Scary and important . But if building alone means everyone dies, then building together is the only path. Not yet ASI, but safer. Check out Damas Limoge, ieeexplore.ieee.org/abstract/docum…

I just watched a video where Peter Martin 🎹 was quizzed on whether several piano solos were AI or human. He got them right, but it was hard. I stick by a longtime view that even if indistinguishable, art is human. This from 2017 youtu.be/KCDVEn5Wzmg?si…

“AI agents” are not the endgame. In 1982, my father built a company eliminating analog devices with PCs for factories. That revolution was hardware first. The next leap needs to be infrastructure, not avatars.

This is my daughter years ago with Jane Goodall. I respected her willingness to change viewpoints when she was wrong. Many scientists defend their own ideas no matter what evidence presents itself these days. Not very scientific .

I really appreciate your openness to work on energy and chips for data centers to make AI successful David Sacks .

GPUs have proven that hardware can transform software. The next shift that will be coming is AI transforming hardware itself.

It seems like low hanging fruit is being missed about energy efficiency. It can come from novel power devices, that can only be made using AI. cubefabs.com Sam Altman Elon Musk Jeff Bezos Dario Amodei Demis Hassabis

Like most people I am thrilled that machines are getting smarter, but the the problem is that we have forgotten what we wanted to be smart for.

I would love to see a day of as many people using AlphaFold instead of image generators . We might cure diseases instead of just drawing them. Demis Hassabis @DeepMind @BakerLab

This is why we have come up with CubeFabs.com We need to de-risk the semiconductor supply chain.

Everyone chases “big” as the path to scale. Cubefabs.com changes the paradigm. each fab can start small, fast, and autonomous , but together they act as the world’s largest megafab. The network itself is the scale.

This too, is why we have come up with Cubefabs.com Using novel materials, we can make chips with better power efficiencies for data centers, EVs, and more.

RL and transformers don’t just train consumer models, they can train matter. The future factory is a living neural network, not a static assembly line.