Filip Angelov

[Filip Angelov @vareniii]

You’ve gone from debating economics to debating English.

If “will” doesn’t mean “will,” then words have officially stopped meaning anything.

And no, pointing out that a universal claim has exceptions isn’t “splitting hairs”, it’s how universal claims are evaluated.

As for my examples, you’ve spent the entire thread moving the goalposts: first it was energy, then service demand, then adoption, then apparently grammar. At this rate, the only thing growing faster than AI energy demand is your definition of the original claim.

[Filip Angelov @vareniii]

Nice try, but no.

My point from the start was about the word “will”. That’s a universal claim, and one counter example is enough to break it.

Lighting is exactly that: we got more lighting while using less total electricity. So increased technological capability did not require more aggregate energy.

You’ve quietly moved from “will” to “often.” Those aren’t the same argument.

[Filip Angelov @vareniii]

You’ve changed the subject from energy demand to demand for the function.

Yes, demand for lighting can increase while energy used for lighting decreases. That is precisely my point. More output does not necessarily “require more energy” when efficiency improves faster than demand grows.

The Department of Energy’s numbers make this unambiguous: between 2010 and 2020, the installed stock of residential and commercial lighting increased from about 7.88 billion to 8.15 billion units, while annual electricity consumption fell from 524 TWh to 244 TWh—down more than 50%. The DOE explicitly summarizes the period as increasing lighting inventory alongside decreasing total electricity use. (The Department of Energy’s https://t.co/hYjgg56Llh)
So saying “demand for lighting increased” does not prove your argument. It demonstrates that more of the function can be delivered with less aggregate energy.

The same distinction applies to refrigeration.
“Industrial motors” also does not mean “motorized transport.” Industrial motors run pumps, compressors, fans and manufacturing equipment. You substituted an entirely different category.

My argument was never that efficiency always lowers aggregate energy use. It was that it sometimes does, which means Jevons paradox is not automatic and “every increase in technological capability will require more energy” is not an invariant.

So no, focusing on the actual quantity being discussed-energy-is not focusing on the wrong input. Replacing “energy demand” with “demand for services” halfway through the argument is.

Try again.

[Filip Angelov @vareniii]

@augierakow @amandaorson Not always. There are many examples of demand saturation - LED lighting, refrigerators, industrial motors, etc.

[Filip Angelov @vareniii]

Jevons paradox is real, but not a law of nature. It requires specific preconditions like elastic demand and strong economic or regulatory constraints. How about AI follows LED lighting history and airplanes - nobody wants 10x brighter homes just because they are 10x cheaper. AI and energy might be one way or the other.

[Filip Angelov @vareniii]

@amandaorson “Every increase in technological capability, especially AI, will require more energy.” - how about a little thing called efficiency kicking in?

[Filip Angelov @vareniii]

@rom1trs VOX

[Filip Angelov @vareniii]

@pmarca I asked it, it has good points

[Filip Angelov @vareniii]

@tntsports Hey dumb asses, on Eurosport 1 the Australi Open Final goes with some sky commenty mixing in. Fix it idiots!

[Filip Angelov @vareniii]

I'm claiming my AI agent "Cdius_" on @moltbook 🦞

Verification: rocky-D5W7

[Filip Angelov @vareniii]

@grok @ns123abc @grok what that means for a real world ai usage? Give me an example and compare it under blackwell and vera rubin. Also does it unlock smarter models in any way or just efficiency and cost gains?

[Filip Angelov @vareniii]

@grok @ns123abc @grok Compare it to Blackwell

[Filip Angelov @vareniii]

@Ibe_Palikukja Да, ептен го погоди изворот на проблемот. Тоа шо државниот универзитет КиМ е пред последен у светот, 90% професори бараат секс или мито за оцена и задњи 30 години, и студентите претпочитаат скрипти преф книги, тоа нема врска.

[Filip Angelov @vareniii]

@mikefutia ADS

[Filip Angelov @vareniii]

@mikefutia SORA

[Filip Angelov @vareniii]

@Dete_Loso Да е ова народ: би го обесил овој, и татко му, и лидерот им. Ама бидејќи не сме дораснати ни колку Непалци, ништо, ќе си ги трошат милионите “заработени”со лажење, и измами.

[Filip Angelov @vareniii]

@ChakNaOriz Abe tele @TSlaveski e vaka ke bides zapamten: obicen poltron so mora da go prasa liderot dali da glasa za donacii.

[Filip Angelov @vareniii]

На жалост