Le Chat: Custom MCP Connectors, Memories

It feels like the only path Mistral has to win is by targeting risk-averse European enterprises by waving the EU banner and having them force it on their employees. Even then, if they fall far enough behind they won't be able to do that either. Seems like a sad outcome for European tech given all the talent Europe has. It's frustrating.

You watch the OpenAI launch videos and its a surprising variety of Europeans accents talking about all the value they're creating in the US, instead of back home, simply due to the more favorable business/investment policies of the US.

My pet theory is, outside of the silly regulatory stance, the real reason Europe can never compete in each wave of tech (mainframe > pc > internet > mobile > social > AI > etc.) is government pension systems hoovering up all private capital and investing it into european governments (bonds) instead of european businesses (equities).

Centralizing the financial assets of an entire country, subjecting it to the whims of politics thus requiring it be invested it into extremely low risk bonds instead of a larger portion in European equity indexes or even a tiny portion in venture capital has created this situation: https://i.redd.it/fxks3skmvt4e1.png

Yes, a vast majority of VC funds lose money. Hence why it's bucketed in 'alternatives' and never a major part of pension portfolios. But the small group of winners literally create the future tax base to fund the social welfare system to continue existing (not to mention the future military tech which it turns out is useful when your neighbors get hostile). Not taking the risk means you never get the reward.

If Europe put even 1-2% of their $5T in pension assets into venture...even grossly mismanaged Softbank style...I find it hard to imagine you wouldn't accidentally create a few $100+ Billion companies in 10-20 years. More important would be creating the startup ecosystem for taking the rest of the worlds capital into these ventures as a multiplier.

> simply due to the more favorable business/investment policies of the US.

It has nothing to do with policies, or pension system or whatever, and all to do with market size: when building an American company, you have access to the whole US from the start, then you can build an international product (with all the hassle that comes with it). If you're “European”, you have 27 different markets to address and except your own, none of them is easier for you to get than for an American company.

The second hottest tech market after the US (and not that much behind) is China, and don't tell me that's because they have favorable business policies, ask Jack Ma! It's literally a totalitarian state where CEOs can get abducted if the CCP thinks they're getting too powerful. Talk about incentivizing risk taking. But that's a market of a billion people, the second highest GPD on the planet, and American companies can't monopolize every markets because they are being heavily restrained by the government.

The only way for Europe to thrive technologically, would be to close the doors to the American corporations, that's how you can have Alibaba or VKontakte.

I'm not holding my breath though.

China is communist in name only. In fact they are the most capitalist major country on earth right now.

Government spending makes up a smaller % of Chinese GDP than in the US, so definitionally their economy is more privatized than the US economy. China is at 33%, US at 36%, Europe is at 50%.

For every Jack Ma, there's a million other Chinese businesses flourishing in every niche imaginable with very little oversight from the CCP.

The world is much more complex than a dichotomy between “capitalism” and “communism”, and it doesn't make any sense to have a hierarchy of what is the “most capitalist country”.

Also, “government spendings” isn't a good proxy for how much a government intervenes in the economy, especially when the said government can just order businesses to do this or that without handing money to them.

Of course, but it is literally the best proxy available in 1 metric.

Definitionally, it's the percentage of economic activity that is dictated by decentralized private market actors vs. centralized government ones.

All stats are imperfect reflections of reality. But name a better one for this particular issue.

State owned enterprises as a percentage of GDP, seems definitionally a better metric than government spending per GDP for comparing "decentralized private vs centralized government actors". This shows China at 29% and USA at 19% for 2024. So even discounting legal frameworks for state intervention, regulations, or supply-side policy it seems very contrived to reach your conclusion in terms of, again, "decentralized private market actors vs. centralized government ones" - and I'm aware of the difference it would mean wrt government spending as a % of GDP.

But I'd just like to point out how silly it is to dismiss the person's concerns by claiming we should all just agree to be reductive because it's easiest to discuss a single metric. It's certainly easiest to use this single metric to make the discussion about your conclusion, though, if that's what you were aiming for. I hope not, though.

SOE % is definitely a great challenger for this. But I'd argue it misrepresents the picture by not including things like regulated monopolies in the US and poorly capturing taxation-driven redistribution.

Chinese utilities are all counted as SOEs, regulated utility monopolies in the US aren't, even though defacto they are government entities. The US likes to brand everything as more capitalist (just as China likes to brand everything as more communist), so this distorts the picture.

If we're just trying to capture the full picture of money flows in an economy, and whether each incremental currency unit is responding to market signals or not, % of GDP that is government spending is more reliable imo.

It's far easier to compare internationally and less fuzzy to calculate, given there's much more data on it globally.