start now. start imperfect. learn along the way, figure it all out step by step.
everyone that says it won't work is right. it won't work for them.
but for you, you'll make it work. start now
If OpenAI is successful in its goal of AGI, the irony is that it will make itself easily duplicable.
The real winners will be those who control the infrastructure AGI needs.
OpenAI knows this. custom chips (Jalapeño), Stargate, data-center hardware.
Elon knows this too.
One thing these new models are making more clear, the AI bubble thesis misses the point.
Intelligence creates recursive demand for itself. That is worth sitting and thinking about.
Even if we get cheaper inference, this enables faster scaling of persistent agents, whose discoveries create more actions, decisions and further inference, driving structural demand for cooling, connectivity, power and grid infrastructure. I expect this demand to increase in orders of magnitude from current levels.
My positioning reflects this view. $INV for exposure to Accelsius data-centre liquid cooling, $GCTS for the connectivity-semiconductor layer, and $IPWR as a power-switching play across data centres and grid infrastructure.
Make Apple beautiful again.
I love @Apple, and I'm unashamedly hooked into the Apple ecosystem. But they have become so big, optimized and scared to take risks, and this shows in their products.
If John Ternus brought me into Apple's hardware team, this would be a glimpse into my vision.
Products that feel emotional, and deeply personal.
$AAPL
Apple HomePod.
instead of just a smart speaker, why not a product that makes your entire home smart, while actually looking like a part of your home decor.
Make Apple beautiful again.
I love @Apple, and I'm unashamedly hooked into the Apple ecosystem. But they have become so big, optimized and scared to take risks, and this shows in their products.
If John Ternus brought me into Apple's hardware team, this would be a glimpse into my vision.
Products that feel emotional, and deeply personal.
$AAPL