ChickenOverlord
No bio...
User ID: 218
ChickenOverlord
4d ago
the most offensive examples that i'm aware of are all in the south and all in favor republicans
The ones in the south are all because the Supreme Court has mandated the creation of "majority minority" districts (minority = black in this case) and blacks tend to cluster in urban areas. Though recent court decisions might be changing that (and are actually what prompted Texas's redistricting in the first place).
ChickenOverlord
4d ago
And in practice, there was no free land; drop them in Northern Australia and aborigine advocates would hate them today (even if there were insignificant numbers of aborigines in the area).
A real life example of this is Rhodesia. The highlands where most of the whites settled were sparsely populated because the native Africans were using wooden tools and the soil in the highlands was largely unworkable for them. After Mugabe took power he seized tons of white farmland on land where the natives had never lived.
ChickenOverlord
4d ago
ChickenOverlord
4d ago
Given that pretty much every single model trains against the benchmarks, I'm going to go with the end users on HN reporting running out of tokens way faster (for the same sorts of tasks they were doing on 4.6) over another synthetic benchmark.
ChickenOverlord
5d ago
Famously, Cisco only just barely got back to its Dotcom bubble era peak (and maybe it still hasn't yet if you account for inflation?).
ChickenOverlord
5d ago
I'm still a fan of Butlerian Jihad General
I'm saying revenue can increase, can stay flat, or can decrease. So the $14B is anything but a sure thing. And given that Anthropic (and all companies) love making themselves look as good as possible, I would assume that they chose the best possible chunk of time to extrapolate into that $14B. Also it's especially strange that they show their run-rate revenue for previous years instead of just their actual revenue, making the likelihood that they're massaging the numbers even more likely.
But they're not a publicly traded company so there's really no way to know for now.
Edit: apparently actual total GAAP revenue from 2023 to the end of 2025 was only $5 billion: https://www.reuters.com/commentary/breakingviews/anthropic-gives-lesson-ai-revenue-hallucination-2026-03-10/
The $5 billion figure refers to GAAP revenue generated from 2023 through to December 2025, a person familiar with the matter told Breakingviews; the $19 billion is an extrapolation. Anthropic defines “run-rate revenue” in two parts. Use the last 28 days of sales from customers charged on a consumption basis and multiply it by 13. Then, multiply the monthly subscription take by 12, and add the two together.
ChickenOverlord
5d ago
Assuming zero growth, and zero decline in revenue, compared to whatever fraction of the year.
ChickenOverlord
6d ago
Note that your link says "run-rate revenue" which is a very different thing from actual revenue. Relevant XKCD: https://xkcd.com/605/
ChickenOverlord
6d ago
Opus is increasing end-user costs, not just changing the tokenizer, and that's the part I said is indicative of a bubble because it looks like they're finally needing to squeeze a profit out of customers instead of subsidizing their usage with VC money. That's the crux of my argument, not just "tokenizer chaged -> bubble." If you're not going to even try to summarize my argument in good faith I see no reason to engage with you.
I also talked about changing sentiment in the tech community, which you completely ignored.
ChickenOverlord
6d ago
OK but do you agree that "Anthropic has slightly altered their tokenizer in a 0.1 update for Opus" is not really "controversial issues that fall along set tribal lines"? Which tribe has a strong position on Anthropic's tokenization design choices?
What an excellent and totally accurate summarization of my post and the argument I was trying to make in it.
I think mods should intervene… somehow, because these posts are getting too frequent, too obviously agenda-laden, and aren't even remotely about the culture war (though AI discussion as such is necessary). It's becoming one guy's AI Bad blog.
I made 3 top level posts about it over the course of about two weeks, I hardly think that's excessive. And I'm still commenting on plenty of non-AI topics, I'm hardly a one-trick pony like some other posters. But if you think it's an issue, then feel free to report me.
He must be a Luca Goers or Al Bhed Psyches fan.
Edit: on second thought, he's probably not a Psyches fan, the Al Bhed are too Jew-coded.
More options
Context Copy link