Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Who's afraid of Chinese models?

stratechery.com

Post reply on HN
Print view View original

721–730 of 965 posts

Re: Who's afraid of Chinese models?

#721

Post by myrmidon »

flakeoil wrote: post #665
myrmidon wrote: post #637

Earlier quoted context omitted.

Outsourced manufacturing is much less of a factor than people typically assume: https://ourworldindata.org/grapher/imported-or-exported-co-e... US numbers are insanely high because cheap hydrocarbons are locally available (=> bad incentives) and everyone is wealthy (that correlation is very strong; just compare Luxembourg, which is much wealthier and more polluting than surrounding nations) and also population densit…

> that correlation is very strong; just compare Luxembourg, which is much wealthier and more polluting than surrounding nations Well, your theory does not hold at all if you look at Switzerland which pollutes 1/4 of the US per capita. CO2 pollution is not and does not have to be correlated with standard of living.

The reality is obviously more complex, and many countries have already managed to somewhat decouple economic growth from CO2 emissions (very good!), but as a crude approximation: More money => everyone gets more stuff => more CO2 emissions (producing and running that stuff).

If you want meaningful CO2/capita comparisons, you also have to be very careful with countries that get ("free") hydro power opportunities for electricity because that distorts the picture massively (same for e.g. Norway).

Big producers of hydrocarbons, on the other hand (like the US) have to work hard to resist the allure of cheap & convenient fossils.

Switzerland is so far ahead in this comparison because they get a lot of CO2-free hydroelectricity (>50%), have to import most hydrocarbons (=> incentive against) and also save massively on transportation because density is much higher.

Re: Who's afraid of Chinese models?

#722

Post by sanderjd »

stymaar wrote: post #692

> The defining characteristic of a commodity is that it is fungible: a gallon of oil is a gallon of oil; a ton of copper is a ton of copper; a bushel of wheat is a bushel of wheat. The concept of “commodity” as defined above is a model, a simplified abstract representation of reality, but that does not match the reality perfectly (the map != the territory). The author claims that a token isn't literally an ideal comm…

Yeah I bumped on this as well. Contra the author's claim, the analogy to energy commodities seems very direct to me. Natural gas is not useful in and of itself, what is useful is the energy or aggregates created from it, and those have very different levels of efficiency. Exactly like Sol more efficiently converting tokens into intelligence than Kimi, a combined cycle gas plant converts gas into electricity more efficiently than a simple cycle gas plant. But this does not imply that gas is not a commodity. And both the more efficient and less efficient kinds of plants have large markets; they just target different trade offs.

Edit to add: I think what he's saying is more like "tokens aren't the interesting commodity, 'intelligence' is", which makes more sense. To carry on my gas and electricity analogy, I would say the same thing about gas being the less interesting commodity than electricity, because electricity can be used for a broader set of useful things. But both things are commodities, despite one being an input and the other being an output in this case, and the conversion efficiency is one very important consideration, but not the only one.

Re: Who's afraid of Chinese models?

#723

Post by onetimeusename »

ballon_monkey wrote: post #233

The 2 things people need to remember: 1) China can (and does) use the models to influence the west. They train in false information about Taiwan and Hong Kong. Or pretend like history is in favor of China. 2) Ignoring the models containing false information, they are incredible. But you should be scared of running inference via the model creators directly. If you think your data is safe compared to running it via mod…

There's an irony in it. I have many reasons to believe China has its own geopolitical narratives and goals just like the US does. If someone dislikes when the US does it, why ignore when China does the same thing and praise them for their efforts? These models exist and sure, they accomplished something interesting. Use them with caution but I don't appreciate the political narratives about good and bad guys and AI models and politics.

Re: Who's afraid of Chinese models?

#724

Post by gigatexal »

samiv wrote: post #717

I'm afraid to trust technology coming from companies operating under the jurisdiction of a rogue aggressive nation that is continuously attacking other nations both (so called) ally and foe using economic and military actions.

Not saying the heavy hand of the Xi admin and Chinese communism is equivalent to the outright lawless, corrupt, grifting and vindictive and hateful administration that is Trump 2.0 ... but it's pretty much * makes the 6 7 motion that the kids do * this for me.

China will just do a better job -- if they do this at all -- of storing, collating, indexing, and using data from their state sponsored and championed AI labs to use that against the US. [1]

The US under this admin is doing the same, attacking universities, allies, it's own citizens.

The two governments are operating more or less the same. Ergo the ai models from each country's ai companies shant be trusted either.

[1] https://www.abc.net.au/news/2019-05-16/grindr-why-is-the-us-... The late Lindsay Graham comes to mind. Even larger bombshells probably.

Re: Who's afraid of Chinese models?

#725

Post by throw103882 »

inigyou wrote: post #619
throw103882 wrote: post #597

Earlier quoted context omitted.

Let’s not pretend that there’s no conflict of interest on the other side of the debate. There are dozens of countries and billions of people with obvious biases against Israel, and they have significant sway within the UN and its commissions. It’s trivial to show that the UN is biased.

Pretend it's 1939. How do you differentiate a bias against Germany from genuine knowledge that the Holocaust is happening?

By the standards applied to Israel, all sides were committing “genocide”s during WWII, and at any other major war past and future.

War isn’t nice, and innocent people dying doesn’t make it a genocide. The term is being diluted to just mean “war with many casualties”.

Equating Israel’s war with Gaza to the Holocaust is an attempt to simultaneously belittle the historic suffering of the Jews while exaggerating the evils of Israel. The two situations have nothing in common.

Re: Who's afraid of Chinese models?

#727

Post by cpburns2009 »

ballon_monkey wrote: post #233

The 2 things people need to remember: 1) China can (and does) use the models to influence the west. They train in false information about Taiwan and Hong Kong. Or pretend like history is in favor of China. 2) Ignoring the models containing false information, they are incredible. But you should be scared of running inference via the model creators directly. If you think your data is safe compared to running it via mod…

Personally I'm more worried about a US provider/model censoring for BS reasons than Chinese ones. Biology 101 is too dangerous for Anthropic. If I want an AI to scrutinize the CCP for Tiananmen Square, Taiwan or Tibet because I'm bored, I'll use a US or heretic model.

Your data isn't safe from exfiltration no matter who the host is. You should host the models locally if your data is truly sensitive.

Re: Who's afraid of Chinese models?

#728

Post by therealpygon »

throwaw12 wrote: post #486

Lets do "who's afraid of US models" version: * Me, as an individual, because I might not be able to pay price hikes, because my revenue (salary) is much lower than what they want and I can't support my expenses via huge bank loans. * Again, me as a new entrant to the industry, LLMs are basically pay-to-play games, again related to price hikes, new entrants might not be able to afford paying those prices 24/7 - which…

I think your second point touches on a bigger concern for small businesses.

US labs have consistently demonstrated their intention to paywall higher intelligence. Eventually the paywall for “hyper-intelligence” will set a bar so high that average small businesses simply won’t be able to afford the bill of what is used by the top corporations to keep themselves at the top. That’s already starting, when it comes to the volumes of tokens top corporations are burning.

This is a feature of the system corporations want to establish and OAI/Anthropic are happy to oblige. 10% of a trillion dollar company is the same as 10% of 1,000,000 million dollar small businesses. Whose hitch would they rather ride, and which size customer easier to obtain to meet their revenue goal?

Not to mention, it would certainly be possible for the EU and other world powers to equally disincentivize use of US labs as a data security risk, since our top models are impossible to run in private lab environments without specialized agreements and there no access to the model weights for auditing. As best as I can tell, AI regulation is a dangerous game that is a hair away from isolationism.

In my opinion, Google is one of the few hopes in this area. There is still a paywall, but I feel like they are the closest thing to a Chinese lab we have (for frontier) in terms of their targets (real business use cases) and they actually have both the infra and already have a pipeline for small businesses into their products; they already have the wide non-AI customer base to leverage unlike Anthropic and OpenAI whose only product requires convincing people to use their (more expensive) AI.

Re: Who's afraid of Chinese models?

#729

Post by fidotron »

jml78 wrote: post #548
kingleopold wrote: post #512

Earlier quoted context omitted.

ethics and other details are for humans. AI companies just proving it, even highest IQ teams are against ethics because they want more, even several millions is not enough for them.

AI has forced a dramatic shift in my view on intellectual property. The laws as they exist only protect corporations now. They do not protect individuals no matter how much people want to think they do. AI has proven this. I think to level the playing field all copyright, trademarks, and patents laws should be eliminated. If I want to make a marvel movie, I should be allowed to and profit from it. AI let the cat out…

> AI has forced a dramatic shift in my view on intellectual property. The laws as they exist only protect corporations now.

Even this isn't quite right. Publishers unambiguously have been screwed.

What has happened is if you can give the political/investment classes enough upside opportunity in the entities doing the IP theft then you're golden, and they increasingly have no problem with even pretending to hide it.

Re: Who's afraid of Chinese models?

#730

Post by flybarrel »

samiv wrote: post #717

I'm afraid to trust technology coming from companies operating under the jurisdiction of a rogue aggressive nation that is continuously attacking other nations both (so called) ally and foe using economic and military actions.

Sir, given what's happening in the world these days, I cannot really tell which nation you meant by your statement :D
Post reply on HN