I don't know, but it's a good question? I have a bunch of test questions that Claude is unable to solve using the web, any version, Fable 5.1 to Haiku 5.5, and pretty much every OpenAI model gets it easily.
My only explanation is that Anthropic's search backend sucks.
What data? To my knowledge the only data point that exists is a 2025 document, which is completely irrelevant considering how massively they hiked the API token price.
To my knowledge, the subscriptions are the only part that may still be profiting from vc subventions. And if not, then their API margins are gigantic
- The suspiciously high amount of money they have to keep raising at a rapid pace
- The fact that they have to play around with silly metrics and leak bullshit to pretend they are healthy
- Their desperation for an IPO.
I feel kind of silly having to repeat myself, it was all there in my previous reply of you bothered to read. I point at that whenever people spread this bullshit of Anthropic being "profitable" especially when it comes to subscriptions. Everything points at it being ruinously unprofitable actually.
Profitability only matters if we consider that they are grifters willing to cash in as quickly as possible before this jenga tower falls apart. It matters less if the interpretation is that the fucking weirdos at Anthropic are death cultists that want to conjure their death god into existence.
but none of those are data points which actually meaningfully tell what we were discussing? They're not even data points, theyre merely observations from the outside, not at all objective.
Anthropic can be profitable and massively overvalued, AND deeply into debt from running too fast and loose with their money in previous years.
That situation would have them behave exactly the same: hurrying to get the IPO done while the stocks are still skyhigh to get over the debt burden and then buckle up for the profitability to weather them through coming correction...
> Why can't you cancel the credit card used for the subscription?
ianal - but I never understood that mindset. You're surely aware that a subscription is a contract, and terminating the payment method does not legally terminate it unless the company doesn't bother proceeding further?
Which may be the case for Adobe, but you're definitely in a legal gray zone in which I'm sure (again ianal) they can sue for damages if they want.
That game always felt more like supreme commander then TA to me... Personally I preferred the foss games spawned out of the spring engine like eg zero-k and the previously mentioned beyond all reason over it
It's true that the game director Jon Mavor, who wrote the graphics engine for TA, was also the lead programmer on SC. And he's not the only point of confluence between the two games.
Clearly. Ever since the government has decided people get to train AI ignoring everyone's copyright I imagine the respect for copyright has gone down the drain, as using other people's copyrights to steal their jobs was suddenly declared legal as soon as it became possible. That is so incredibly unfair there's ... there's no more pretence.
If big companies can make money from people's copyright, it doesn't apply. If big companies lose money from some person's copyright, BIG sentences, off to jail!
By the standards used in Nappster cases, combined with the 500.000 works Anthropic was caught using that becomes (2 months jail + $150.000 PER INFRACTION, because "wilful infringement"), Sam Altman, Dario Amodei, Mark ZuckerBerg and Larry Page deserve 500.000 * 2 months = roughly 100,000 YEARS of jail (each of course) PLUS 75 billion dollars each PLUS of course starting to do these cases internationally (they certainly didn't limit themselves to downloading US books only)
But of course, such things only apply to 9 year olds stealing big company copyright.
This is yet another step towards the inevitable total clash: when claude AI is good enough that it can build Halo, The Simpsons: Hit And Run, ... (slight variations, e.g. with different IP) without needing more than a promo vid and a wikipedia page on the game. Same for movies. Songs. That, too, is coming.
Or even do the Disney playbook from scratch. After all, essentially every story Disney touches is public domain: Mickey Mouse itself, of course, Snow White (
Once we are there, that will force congress to finally choose: destroy copyright or declare it's scope to include AI training and AI outputs. And not just congress but EU, EU country parliaments, even Russia and China's CCP will be dragged, kicking and screaming, into making this decision.
> Sounds like LLMs shouldn't respond to insults like humans do either then.
is impossible, because thats literally what a llm is. its a stochastic model which produces output according to the predictive calculation of how people wrote. it can never be separated from our own behavioral quirks, even if it may or may not be more intelligent then a person off of the street.
i dont know how many years ive been a subscriber -- the oldest email ive got from them is switching my email 8 years ago, and i know for certain i used it on another email before that... but emails on that previous domain were never long lived, so theyre gone.
anyway, the bigger issue ive with this is the doubling of the price right from the get-go.
with private equity on the steering wheel, i suspect this will keep going up every year from now on, so ... while i too have been a loyal customer to date, i suspect ill be driven out within the next 1-2 years, because if they double the price again next year, its gonna be way beyond the value i get out of it given how decent the alternative have become since.
That's an attempt at mitigation, but most definitely not solved.
It still causes both false positives and false negatives through that. The only true "solution" is to make sure generation always happens on the same platform as your ci... And various mitigation strategies around that (eg fall back to structure tests on other platforms vs actual visual diffs in ci
Also not unique to playwright. Been available basically everywhere since the start
> Outsourcing is not new, yet they haven't done it to date
I didn't say "outsourcing", outsourcing is where you have a different company do the thing (who may or may not be in the same country).
What I am describing is "offshoring" (specifically where it's the same company, though this is not strictly necessary for the word "offshoring").
> yet they haven't done it to date
As of June 30, 2026, we employed approximately 223,000 people on a full-time basis, 121,000 in the U.S. and 102,000 internationally. Of the total employees, 89,000 were in operations, including product support and consulting services, datacenter operations, and manufacturing and distribution; 77,000 were in product research and development; 43,000 were in sales and marketing; and 14,000 were in general and administration. Certain employees are subject to collective bargaining agreements.
reply