Interesting link, thank you.

If the oil prices stay up (and the orange T seems to be doing his best to make it so) then I agree that EVs should drop in value much slower than petrol cars.

All of what you said but there is more: most people will do some reading before buying an EV. What will they probably see? A few scare stories about charging on the road (if I don't go to a supercharger or a fast charger I've never used or not recently used it's always a bit scary whether the fast charger will work - and of course mostly they do).

But mainly they'll read how much better other cars and mainly Chinese cars are.

Who wants to buy 2nd or 3rd best?

I spoke with a colleague here in the UK who recently bought a car (not an EV) because he was worried what the resale value of current generation EVs will be when the 400 mile WLTP range, 5-10min charge next generation (mostly Chinese) hits the market. And he has a point.

To be fair even I bought my EV (Model 3) second hand to shave of the steepest part of the depreciation curve.

Given that citizens of the US have elected Trump as the US president twice it is pretty clear that EU countries cannot count on the US being a 100% reliable ally in the future.

That has all sorts of consequences and will require the EU to develop all sorts of capabilities.

The question of course is: will this mean willingness to reduce benefits / increase working hours to pay for all this to develop genuine competence through significantly more effort or will it be just performative?

Your analysis is mostly correct except UK uses no coal

https://grid.iamkate.com/

But indeed, better transmission, more storage and more renewables (or nuclear) would mean that gas isn't needed more of the time and wholesale cheaper.

Looks like we're close to the peak of the bubble. The question is which: the AI bubble or the SF property bubble or both?

Yeah sure, you can trust is with backdoors to encrypted communications but only for the good guys because they know how to take good care of such important mechanisms.

All of them really. What's typically open source is
1) the code used for training, but never the dataset for initial LLM and never the RLHF (reinforcement learning with human feedback) data used to make a text vomiting LLM into a useful question answering maching.
2) the resulting weights - these are totally uninterpretable.

So it's never fully replicable; even if you had the infra and were willing to burn electricity you don't have a way of going to 2) yourself.

AFAIK that's not just the Chinese but also open-source / weights Llama and Mistral.

This sounds dismissive, but I wouldn't read it as such.

"China doing something first, however, has never been a reliable indicator that the thing will prove durable, economic, or widely replicable. China is large enough to try almost everything."

This has always been true, you can for example read the history of improving iron / steel production during the Industrial Revolution. Either you had existing outfit and capital to try things. Or you raised capital and set-up an outfit to try things. If you had something viable, you made money, if not, the world (and hopefully you) moved on.

And in the end VCs work more-or-less on the same principle (which is why at some point someone was trying to do Uber for xyz).

This is a PR "thought leadership" BS article by Benjamin Riley, Cognitive Resonance, who "provides direct consulting support to organizations to improve understanding of how generative AI works."

This doesn't mean they're wrong but it's probably nothing terribly original (there is a reason why it's not on openreview.net as a submission into one of the relevant AI conferences).

What I don't get is this: let's say they believe that to get to "genuinely useful AI" (generating value in business, automating science) they need 1024x the compute they have now; which I think is ballpark with their spending.

But Moore's law says that number of transistors doubles every 2 years, so if we believe we can keep this going then in 4 years time they'll only need 256x the data centres and in 8 years it's only going to be 64.

So all the chips they put in will be next to worthless in 8 years and they'll only need a fraction of the other infrastructure. So all this massive spend is to be first because? Because whoever is first will be able to use their "genuinely useful AI" to out-innovate everyone to singularity?

The whole thing is a one-way bet on the empirical LLM scaling laws (https://arxiv.org/abs/2001.08361) scaling to something useful, which I don't buy.

The whole situation is so bad, I can't help but suspect the former Bungie employees had some incentive to make the acquisition worthless.

The ignorance of this fact seems intentional.

Funny how there's this big push for people to join some ephemeral, federated network where nothing but text is discoverable, content moderation is limited to complaints to disinterested admins, and privately-hosted servers can be added / removed with a minimal amount of effort.

It's almost as if someone realized the available tools for identifying CSAM have become more sophisticated and a new platform that's harder to monitor is now a necessity.

The EU's Mastadon instance will average 2k daily visitors who will spend less than a minute on the site. Some of those visitors will be bots. In about a year, those numbers will go down and discussion of the site will be limited to some wonks in Brussels arguing over whether it was a success.

Mastodon / Diaspora / Matrix - each of these operates with a very different set of incentives. In theory, yes, it's nice to think of social networks in some ideal form, as an expression of a public good. Jack Dorsey seems to have that vision, you may share it and good for you if you do.

But there's a reason half of Apple's revenue comes from iPhone sales, 95% of Facebook's revenue comes from ad sales, 75% of Alphabet's revenue comes from ad sales, and 50% of Amazon's revenue comes from fees they charge companies who sell through their portal.

https://www.visualcapitalist.c...

FAANG are a highly targeted system for fostering, promoting and satisfying desires. While it's possible to think about their features in a civic-minded, pro-social manner, they constitute a network and these features represent a small part of what they do. Value extraction and disruption are the chief reason they exist, which is anything but civic. They have highly sophisticated tools to keep people on platform and that is what explains the high engagement levels and participation rates.

Decentralized networking apps have no corresponding capabilities. Another metaphor would be a Roman Soldier versus a Tank. One lacks the ability to imperil the other, no matter how well-trained and fearsome it might be. And everyone can see the problem with a tank, it's mostly an expression of oppression that takes up space and occasionally blows things up.

But no one's going to replace their tanks with Roman Soldiers. They don't serve the same purpose.

Do people spend money on what is useful or on what they crave? 2 very different things.

Futile gesture, these numbers are meaningless.

Mastodon / Matrix are conceptually defective. Chronological timelines are not what drive people to social media. Network theory, especially Triadic Closure, drives people to them.

https://wikiless.org/wiki/Tria...

People have lots of reasons for hating social media, but it's nature as a emotionally manipulative and coercive medium is what attracts large audiences.

Decentralized social networks apps, by their nature, lack similar capabilities. While you might regularly visit a Mastodon instance, the likelihood of experiencing a craving to be there all day is slim.

The difference between these platforms is like heroin and aspirin. One compels attention, one relives pain. You don't always need pain relief, but you always need to feed the dragon.