Makes sense. You’re saying why would regulations matter when clearly they can just be ignored.
For venture ultimately it’s a soulless moneyman’s game. Really they have to pick winners, and anybody can look at the landscape and see there’s just not gonna be a Pierre Zuckerberg or a Klaus Kalanick. And if there ever is, he’ll need to raise lots of money anyway, which would come from venture.
> there’s just not gonna be a Pierre Zuckerberg or a Klaus Kalanick
Which is very much a positive. Those two aren’t a boon to humanity, they very much made everything worse at a global scale. We need fewer people emulating them, not more.
> The mechanism consists of a revenue-based levy applied to all commercial providers placing AI models on the market or putting them into service in Europe, reflecting their use of content publicly available online. This levy would apply equally to providers based abroad, creating a level playing field. The proceeds would flow into a central European fund dedicated to investing in new content creation, and supporting Europe's cultural sectors.
Presumably Mistral is putting forth the most pro-AI position possible for the region.
So it sounds like anyone doing what you described is at risk of a tax that will make their offerings uncompetitive.
Why not just force them to to build the common infrastructure and services, and in exchange they get to keep the money? e.g. Jeff Bezos has to build some subway stations in NYC or something.
That way you get somebody with a proven track record of building big projects who is also motivated by money, so the common infrastructure and services is handled competently.
> Why not just force them to to build the common infrastructure and services, and in exchange they get to keep the money?
Because it is undemocratic, ripe for corruption and abuse, will never work in practice (as the rich will inevitably find ways to game the system). What you are describing is basically just aristocracy, where the rich get to decide what is best for the rest of us.
Ah yes. Let's trust civic engineering to a man who ran a company that had front-line workers using piss bottles to keep up with quotas. This cannot possibly end badly.
I prompt it and check CI later. I couldn’t tell you how Codex feels. I’ve never had any conversation. You may want to try this sort of workflow if you’re affected personally in a negative way.
Litigation—the hope or fantasy to make a buck—soaks up a lot of the million-man animus I’d guess.
If that’s accurate, Luigi Mangione would be the exception that proves the rule. The “unwashed masses” generally want money more than they want to effect change in the world.
A lot of people spend mental energy fantasizing about getting rich off lawsuits. Like, a lot.
That kind of tracks as the source of the concern. My first thought was it’d be something IDMS-related as well. I don’t know enough about that system to pinpoint exactly what.
The article references “public panic buttons” and how
> There is active participation by the citizenry, where they connect their private security devices to the command centers run by the state
You don’t really believe anybody using a “public panic button” or hooking up their own alarm system to law enforcement deserves the worst fate imaginable. That’s a little extreme.
What are we even trying to accomplish here? It sounds like individuals in parts of Mexico are trying to protect themselves.
There has to be some compromise between ideals and reality. If you reflexively tell people “you can’t help the cops for the sake of democracy,” they’re gonna throw out the democracy part and keep the cops part.
Maybe a short stint in jail in the case of misconduct, but the worst fate imaginable? Chopped up in a suitcase?
It is admittedly pretty goofy to get exactly what you want—an army of people making rules for everything under the sun—and come on here and complain about what we’re doing.
Even TFA, which is about yet another rule, has a goofy quote from the Minister of something or other about breaking free from American tools. Linux seems pretty American to me [1]. Maybe they’ll fork. Would be cool.
I mean can’t you have a grammar on both ends and just set out-of-language tokens to zero. I thought one of the APIs had a way to staple a JSON schema to the output, for ex.
We’re making pretty strong statements here. It’s not like it’s impossible to make sure DROP TABLE doesn’t get output.
You still can’t predict whether the in-language responses will be correct or not.
As an analogy: If, for a compiler, you verify that its output is valid machine code, that doesn’t tell you whether the output machine code is faithful to the input source code. For example, you might want to have the assurance that if the input specifies a terminating program, then the output machine code represents a terminating program as well. For a compiler, you can guarantee that such properties are true by construction.
More generally, you can write your programs such that you can prove from their code that they satisfy properties you are interested in for all inputs.
With LLMs, however, you have no practical way to reason about relations between the properties of inputs and outputs.
reply