And cost of inference tripled from $3B in 2024 to $10B in 2025, so cost of revenue linearly grows with number of users, i.e. it does not get cheaper.

https://www.wheresyoured.at/oai_docs/

The interesting question is: In what scenario do you see any of the players as being able to stop spending ungodly amounts for R&D and hardware without losing out to the competitors?

I mean we just have to look at old discussions about Uber for the exact same arguments. Uber, after all these years, still is at a negative 10 % lifetime ROI , and that company doesn't even have to meaningfully invest in hardware.

IMO this will probably develop like the railroad boom in the first half of the 19th century: All the AI-only first movers like OpenAI and Anthropic will go bust, just like most railroad companies who laid the tracks, because they can't escape the training treadmill. But the tech itself will stay, and even become a meaningful productivity booster over the next decades.

> there was a single brain region where we saw that higher cannabis use was actually associated with lower brain volume – the posterior cingulate, which is part of the limbic system and is implicated in processes like memory, learning, and emotion. That said, some research suggests smaller posterior cingulate volume is actually associated with better working memory, so it’s a little unclear what this means.

(I'm ignoring installation costs etc. because actually creating the satellites is ignored here, too)

Can you share the source? The last time I looked the association was both clear and pretty strong, e.g. "There was a linear dose–response association of TSSM and risk of depression. The risk of depression increased by 13% (OR = 1.13, 95%CI: 1.09 to 1.17, p < 0.001) for each hour increase in social media use in adolescents." DOI:10.3390/ijerph19095164

> “Social media” lumps together very different things

HN does this, but the research is usually pretty clear in spelling out they mean FB, Insta, TikTok and so on.

> If you want a lever that actually changes incentives, go after business model & design

I too would like changes in that direction (mostly because adults are also affected negatively by social media), but keep in mind even a non-optimized, strictly chronological feed produces these negative effects, see keyword (and associated studies for) "upward social comparison", i.e. people are always more inclined to post about things that went well or are fun, and thus even a pure chronologically sorted feed produces a warped perception of normal social reality.

People around the world distancing themselves from these actions is hardly nationalism.

https://www.thelancet.com/journals/lancet/article/PIIS0140-6...

So even if these lobby talking points would be true, and everything had to be 100% subsidized, that wouldn't be a problem.

https://www.reuters.com/world/europe/trump-vows-tariffs-eigh...

If you want more performance, you could try running llama.cpp directly or use the prebuilt lemonade nightlies.

And personally I don't care at all about the performance delta - we are talking about a difference of 6 to at most 12 months here, between closed source SOTA and open weight models.

The company I work for for example, a mid-sized tech business, currently investigates their local hosting options for LLMs. So Mistral certainly will be an option, among the Qwen familiy and Deepseek.

Mistral is positioning themselves for that market, not the one you have in mind. Comparing their models with Claude etc. would mean associating themselves with the data leeches, which they probably try to avoid.

Are you sure? They are advancing, sure, put look what they paid for to achieve this: 300k dead, 700k wounded, depletion of their souvereign wealth fund, 20%+ inflation, lower oil production and so on.

Things look decidedly different if you exclude the ad companies (Google, Meta, ...) and associated shovel sellers, see the WaPo article about the S&P 493 from a few days ago.