wqaatwt

6 ms·

91.
by wqaatwt 3mo ago
Technically yes but its not that straightforward. Take a country like Sweden for example everyone is reasonably well off (if not exactly thriving) since income inequality is quite low. At the same time wealth inequality is extremely high si
92.
by wqaatwt 3mo ago
Or hoarding real estate.. which is not particularly productive since it turns out higher prices don’t really stimulate supply sufficiently. Also its not exactly obvious how much inflated stock market valuations benefit the economy that much
93.
by wqaatwt 3mo ago
Well because they would risk losing a handful of sheep occasionally (which the government would likely pay for anyway). Also the prevalence of general cultural hatred of natural habitats and ecosystems prevalent amongst some sections of rur
94.
by wqaatwt 4mo ago
Gemini 3.1 is still up there, though? If Google started to compete on price they could be very successful.
95.
by wqaatwt 4mo ago
Mostly they kind of do since we do live in an utopian society of unlimited abundance. Extremely few people can afford to (or want to) spend a very large number of working hours without ever getting anything directly in return for it.
96.
by wqaatwt 4mo ago
On aggregate mainly because humans often tend to behave “irrationally” due to various reasons though
97.
by wqaatwt 4mo ago
> Open-source is also altruistic Contributing to it might not necessarily be. Most open source development is funded by large companies after all and from their perspective it can function as a cost saving measure. Allowing them to focu
98.
by wqaatwt 4mo ago
All the tech CEOs had no qualms about groveling before Trump and licking his boots, so yeah I assume they must be 100% onboard with stuff like this as well
99.
by wqaatwt 4mo ago
> what kind of scheme the administration is up to I’m sure they are wondering just as much. I assume exhorting Anthropic/OpenAI for personal bribes, favorable government contracts with no restrictions and public acts of submission.
100.
by wqaatwt 4mo ago
> can we make sure these models aren't dangerous They can’t, though. The models might or might not be too dangerous but the people running the US federal government are too incompetent and/or corrupt to do anything useful about
101.
by wqaatwt 4mo ago
It’s going to be a bit trickier to do that, even banning US providers from hosting them legally might be tricky to do.
102.
by wqaatwt 4mo ago
Or significantly increase their market share outside the US and give them some breathing space to catch up with the currently available closed models
103.
by wqaatwt 4mo ago
Is there any evidence that’s true? i.e. they are losing money on every additional token they sell?
104.
by wqaatwt 4mo ago
Maybe the implication here is that supposedly the real reasons are personal/political and the stated ones are just the justification?
105.
by wqaatwt 4mo ago
Well again.. you are mixing up inference costs and their other mostly fixed expenses (in addition to sales and marketing) Is there any indication that if they could sell X * N more tokens than now at the same (or even quite a bit lower) pri
106.
by wqaatwt 4mo ago
Since what you said applies to the company as well. If they don’t like customers breaking their possibly (or certainly at least in some jurisdictions) largely legally unenforceable ToC conditions they are entirely free to stop doing busines
107.
by wqaatwt 4mo ago
For some time capital costs seem to be dwarfing the operating costs (or at least energy costs) when it comes to GPUs. Of course that will likely change but it might take quite a while (unless the AI bubble violently pops)
108.
by wqaatwt 4mo ago
I don’t exactly see how the modern stock market is superior based on the (moral not economic) criteria being discussed in this thread? It’s even more speculative and detached from productive behaviors.
109.
by wqaatwt 4mo ago
We were talking specifically about inference and I don’t think there any indication that their gross margins on the API tokens (if not the personal subscriptions) are negative? Obviously they have R&D and other fixed expenses that make
110.
by wqaatwt 4mo ago
Based on what? There isn’t a lot evidence that’s the case.. Prices on OpenRouter for GLM and other large open models indicate that Anthropic/OpenAI must have pretty high gross margins even if their models are several times more expensi
111.
by wqaatwt 4mo ago
Why? Well it depends, most evidence is suggesting that Anthropic and OpenAI are making a lot of money on inference so the question is whether its more profitable for them to sell 100X tokens for Y, or 1X tokens for 100Y. In most industries
112.
by wqaatwt 4mo ago
The company is completely free to void the contract and stop selling their services to anyone they don’t want to?
113.
by wqaatwt 4mo ago
> how the terms of service can simply say you're not allowed to sue. That doesn’t necessarily mean much. You can put plenty of outrageous statements into any contract that automatically doesn’t make them binding.
114.
by wqaatwt 4mo ago
We really don’t know what are Anthropic’s margins on inference. Most available data indicates they are quite high on the API so it’s not that obvious that subscriptions are unprofitable.
115.
by wqaatwt 4mo ago
Stealing credit cards seems not directed related to using legitimately acquired LLM outputs for whatever legitimate purpose you want to use them.
116.
by wqaatwt 4mo ago
An endless list of tangential analogies isn’t really a valid argument.. DMCA has as little to do with this as streaming copyrighted content
117.
by wqaatwt 4mo ago
Using a bunch of nonsensical/irrelevant analogies to somehow make a point seems worse than these “word games”? What does streaming copyrighted content have to do with LLM outputs (which are public domain)?
118.
by wqaatwt 4mo ago
No access to capital can and does constrain supply to a very high degree.
119.
by wqaatwt 4mo ago
A basic calculator is 100% better than humans at all the things its designed to do.
120.
by wqaatwt 4mo ago
> Once locals get to Opus levels I find it hard to imagine it would ever be cost efficient vs hosted/cloud i.e. you should always be able to run faster and/or better models remotely at a comparable price since its just way more

More ›