got its own post

looks like

I’ve been pondering whether the intermittency is also (by now) intentional - this way he only “sets direction” and the group self enforces. it’s a handy way to avoid really hard-ticking all the cult boxe

nfi if he’s really that clever tho

The AI start-up is continuing with plans to list before the US Thanksgiving holiday this month, according to multiple investors in the group

someone seems a bit rushed

“It’s just one milestone along the way,” said one early Anthropic backer. “They should treat it like every other fundraise: a means to an end.”

I certainly hope they mean for it to end

Anthropic released its financial data for the third quarter earlier this week, revealing revenue growth that had slowed but still jumped almost 50 per cent to around $17bn.

I like how this doesn’t point out that the same data showed their compute costs have doubled

candlestick charts X matrix mult

just imagine what cursed nonsense they might go for

it was quite funny how, while counts were going up, he just entirely shut the fuck up for about 3~4 hours. not a tweet in sight

must’ve been tragically hard overcome that rage. poor pissbaby, so sad. anyway,

yeah they’re far past public-only sources, there’s more than a few non-public sources going into this shit too. and still it’s not enough to work and sort out the issues! almost as tho the whole thing is a bit of a silly idea

I’m glad there’s finally some financial reckoning coming, I just hope I’m wrong about how utterly fucking bad it will be

hopefully also a lot of accountability and secrets opened up

More people using it, more training data fed into it

yes but that’s not actually the hard and costly part. the hard and costly part is the actual training action, which runs in the hundred of millions if not more

which, y’know, when lower prices……

lol, I saw that going aflame among the promptfondlers nearest me, too

model auto-tune. they’re getting hype for model auto-tune. I expect we’ll little outside breathless tests of autotuned models for a few days

this combined with the anthropic amended s-1 (which I very briefly dug into here, will dig more later), the numbers are all just about as bad as I’ve been thinking they’d be

so there’s a new nber econometrics paper out, covering our favourite industry

it has some bangers in it

Paid adoption is substantially more common among younger and higher-income consumers: those earning more than $200,000 per year are 1.7 times as likely to pay as those earning less than $50,000, and those younger than 55 are 1.8 times as likely as those aged 55 or older

starting strong!

This force creates a dynamic incentive to set lower prices today: lower prices attract more users, improving future model quality and raising future demand

I don’t entirely follow how setting lower prices today improves future model quality but sure I’ll go along with the ride

Despite this limited horizontal differentiation, we find some degree of vertical differentiation across providers, consistent with observed market shares. At the end of our sample, the average consumer was willing to pay a $10 monthly premium for OpenAI products relative to Anthropic products. Part of this difference reflects model quality: our estimates indicate that consumers are willing to pay more for higher-quality models. For instance, moving from GPT-4’s capability level to GPT-5.5 Pro’s raises average willingness to pay for OpenAI products by approximately $25 per month

We also use the model to examine how consumer preferences for AI services shape competition. We begin by considering substitutability across services. By the end of our sample, all pairs of AI services are substitutes on average

“do you want tea or coffee?” “yes”

OpenAI has the highest retention of all AI products we study, with 77% of people still paying after one month, 65% after three months, and 55% after six months. Anthropic and xAI are lower, but similar to each other: around 65%, 50%, and 40% after one, three, and six months,

in the words of basement jaxx, down down downdown down downdown downdown

chat is it good when your world changing ai god not getting picked for anyone’s team during lunch football

hah, nice. yeah the template layout immediately made me think that, but your word has far more weight :)

I really dislike how he believes that LLMs and fascism are “separate axes”. Where does he think the LLMs come from?

valsorda worked at google for a very long time, well past when it became clear that they were no longer the good guys. his motivation for this was (only very slightly paraphrased) “well, it’s better than whatever else I might get in my home country” (italy)

he’s also had a few choice other takes over the last few years, enough such that I stopped following his work. I would not be too surprised if he turns out to be quite a shit in some years to come

you’re the kind of programmer your colleagues hate having to deal with

looks like iocaine with a papers corpus (although I guess kyle could’ve made his own thing too)

The only thing that would really prevent me from migrating to greener pastures is if they stop collaborating with the Zionist Swiss University that support Israels military.

so they’re actively doing it, but you’re also still actively on their service. and you’ll only consider not moving if they stop that collab? you don’t consider moving now, while they’re still doing it? wat

support mail, “hi I was considering using your service but on learning about your CEO’s shitty views have decided to take my money elsewhere”

“hey chat what’s an EDR and is it good that we don’t have one?”

saw that earlier as well, and I had a real good laugh about the myriad ways that will go wrong. both the obvious (derping proprietary extensions, etc), and the less so (crashing a control plane because the chatbot doesn’t understand 256MiB ram)

the sheer utter range of the crisco product range as it is…. phenomenal display of how removed from knowledge the execs are