I agree with your conclusion. I can’t see where the income is coming from to fund all the datacenters being built.

The bubble is based on creating AGI, but people already don’t want to pay for sub AGI intelligence that is already beyond most people’s needs.

If run locally you get none of those problems.

S&P 500 PE Ratio hit 120 in 2008, but thats because earnings collapsed.

My last company they didn’t delete messages. That would be to obvious.

“I am sorry we didn’t get around to answering all the questions live. We will respond to the remaining by email”

No more questions were answered.

Apparently the spacex ipo is 2x oversubscribed

Your plan is too bloodthirsty.

We just need the assets.

Now you are being anti-semantic

It’s in the title so you are not mistaken. The problem is that various statistics have been reported and we don’t know what is correct.

Very true. You can build a pool on top of a mountain and pump/discharge water but it is super expensive for limited capacity.

This project is only 500 MW here

https://www.pv-magazine.de/2024/09/20/flexbase-plant-500-megawatt-redox-flow-speicher-in-der-schweiz/

And other places say 800 MW

Both of which are comparable to large hydro.

Modern pumped hydro has a ramp up in the 10s of seconds range.

Anyway. Same ballpark in terms of power.

I see your Luton and raise you a Swindon.

They are not standard sized.

FYI. Hydro power has similar capacity and start up times

The downside is that does batteries are not be energy dense. Perfect for grid storage but useless for car batteries (where the bulk of battery R&D money goes).

I hope the AI lies and makes me look hot.

Ok. How is this different from cameras on phones, or cameras hidden in shirt button?

Also true for spy cameras. And they don’t have the disadvantage of needing to be attached to a person to work.

And private smart phones.

How is running a local model empowering anyone else?

Which is why deepseek releasing it’s model open source was such a big deal. The monopoly isn’t feasible.