This seems reductive. Character.ai has > 20 million monthly active users, and a lot of company valuation happens on the basis of user count (which is legitimate, that's not vaporware)
Why don't they just not count token costs at all (or any other cost-per-user) and only look at profits, like what Anthropic is doing to figure out if they're profitable?
I'm seeing this repeated a lot, but I'm rarely seeing the evidence to back it up. Presumably you're not speaking about Yandex's search engine crawlers here, right?
In an internet full of metrics and graphs, it's odd that nobody's bringing anything to the table and is just repeating the meme. A bit bottish in itself, tbh.
If it turned out "usercount was vapourware" due to 3 Russian data centres getting btfo, I think there'd have been a global financial crisis underway as we speak.
I don't see any connection to Character.ai, but it's pretty obvious that Twitter is 90% bots, politically leaning heavily in same direction as Twitter's owner, whose pro-Trump interests are aligned with those of Russia.
Where all all these bots being hosted? I have no idea, but it certainly would not be surprising to find many of them coming from Russia (Yandex). Twitter must at least know the IP addressees of these bot accounts, but does nothing to stop them. Perhaps they are actually run by Musk not Putin, who makes no attempt to hide his attempt to use his position to subvert American democracy.
the rich in society (corps, individuals) are so absurdly rich that they can keep throwing money till something sticks with almost 0 risk. If you believe AI is something civilizationally transformational which it is resonable to believe, then why not throw some money at it.
Have you peeked at how the (very complicated) financial markets work?
Lot's of people staring at computers moving money and ious around complicated networks within complicated rules making money making money making money.
You're using amazing technology funded through those financial processes to complain that those financial processes aren't progressing humanity forward lol. Modern AI is the perfect example of how financial systems are important for technological progression.
Thinking complex financial systems aren't contributing to humanity's progress because you don't recognize how it is used to fundamentally support the work that progresses humanity is like thinking taxes are a scam because you don't recognize how much you actually benefit from taxes everyday.
It's sad that you are being downvoted, for what you say it's pretty correct.
A lot of people don't understand the reason why financial markets were created and why they continue to exist.
There's a really good book called "Against the Gods, the remarkable history of Risk" which gives a good account on an aspect of why financial markets exist.
Basically, parts of society want to advance things (technology, arts, medicine, etc) but that same part doesn't have to carry the risk of doing their project. Financial markets are there to transfer the risk to people willing [able] to take on it.
Credit and Risk are two tools that have influenced the progress of civilization so much, its crazy.
Sure, there are people that abuse financial markets, the same way there are abusive people in Tech, Arts, or any other human endeavor. But that should not disqualify the whole system.
A valuable company wasn't created in the end, but all kinds of value accreted out of this endeavor. The co-founders almost certainly walked away filthy rich. A few hundred people got 6-figure paychecks. A few early investors probably made a profit selling to greater fools.
It is rumored that their last round valued them at $6B.
accrete means grow or make larger/greater; this is the opposite were a bunch of value moved around, some was lost (from an efficiency / opportunity cost perspective) and a bunch was concentrated into the hands of a few. That's not value growth.
True, I meant to say value was ejected from this i.e. there were some peripheral winners. I had a totally wrong definition in my mind of the word "accrete" (I was linking it to accretion discs around black holes).
It's a roleplay app, an AI version of the old RP forums, or RP private chats. Anyone sane and slightly mature understands they're not being encouraged to do anything IRL. The main problem I see here is a lack of parenting.
The US just has a lot of venture capital money running around. It's common for biotech companies to raise about as much as Character.ai (although typically at lower valuations) and then never end up launching a successful product. From a societal perspective, it's valuable to be able to search the space of long shots; from a funder's perspective, the rewards of a good investment are high enough to pay for a lot of bad ones.
That shouldn't have been a problem given that AI can fully replace talent. The whole company should in theory only have needed a dream and an Anthropic subscription.
Well one of those two names was a core contributor to the Attention paper, arguably the top contributor, before leaving Google to make Character Ai, no wonder the re-hired him back (before he left again).
I don't think I know of anyone who denies the value of AI. I think 95% of the people who use it know that it is a tool and are angry / dismayed / confused about companies that are parading it around like a drop in replacement for anything.
Photoshop is a tool, vs code is a tool, git is a tool, excel is a tool. No one expected these to autonomously generate money.
People will be happy that the 5% who are delusional are going to be forced to reckon with the lies they've told themselves.
As an aside, I believe there will be a huge surge in crypto hype again once these sorts of get rich quick with AI schemes dry up.
Crypto always has been backed by the natural demand of drug consumers, drug dealers and drug sellers. Now with marijuana being somewhat legalized in more and more parts of the world and everyone being able to see that the war on cocaine is all but lost... I think the demand for crypto will not recover any more.
There won't be a monopoly. Too many open weight models floating around.
The thing I'm looking forward to is the unsustainable bubble collapsing. It cannot happen fast enough, because with every day it grows, the consequences of the collapse will get worse. And it, purely mathematically, has to collapse - there is no such thing as infinite growth, and some day the investors want their money back that justifies the valuation.
I had a conversation with someone some time in 2024 about Character Ai, wondering how something so obviously exploitative could exist, one data point to consider.
I think it's unfortunate that this is the case and that the service has been treated unfairly in many regards, but c.ai has faced specific problems that other businesses in the space don't, and the basic service isn't going to be marketed towards the more revenue-generating segments like enterprise without making something different out of it.
What were those specific problems they faced? I haven't paid any attention the market for cloud-based AI services by all these non-frontier companies, but I know it's a big market
I heard about various lawsuits, including one where apparently Pennsylvania alleges that a literal AI character has claimed to be a licensed psychiatrist inside of a chat, after the user asked.
Mostly playing devil’s advocate here but if the McDonald’s app glitched and showed a popup claiming to be a licensed psychiatrist I don’t think they would be liable for the consequences for any self harm that happened to me because it’s unreasonable.
Not my best wording for the metaphor but why would a company specialized in creating fictional personas for individual users be liable if their fictional personas pretended they had real advice.
Aka how is this different than telling the doctors and the police that a horse told you that you would be hexed if you didn’t try and rob the store?
Sometimes talent decides to just make money and scam people into thinking talking with AI versions of celebrities as something worthwhile and positive for the world.
It reads like they didn't go under yet, they just expected to with the way the legal challenges were going. As part of that, Disney seems to hope to be taking a good portion (all?) of the employees in the sale. Without keeping staffing up, the sale might not have been viable. If they didn't have such an exit plan involving the employees still being around then it might have made more sense to freeze hiring and ride it out until the end or whatever option would make sense in that story.
I worked at a company in 2000s that was acquired. I had no stock but the owner still gave a generous spot bonus and a years salary bonus retention package to employees. It’s sad that this doesn’t seem to be the norm? Screwing employees on acquisition seems like a severe breach of social contract.
What are you talking about? I was an employee in 2024 and I was not “screwed”. My options were converted into cash at the same valuation that Google paid (vesting schedule remained intact though). Noam did right by his employees.
I had the same thing happen. Twice, actually - after the non-competes expired for the first acquisition, of course I accepted a position working for the same guy at the next company. Another successful exit there. These were small time, meat and potatoes type startup companies. Nothing fancy. The one thing I did post about here on HN, which I think was one 1) ahead of its time, 2) lacked investor connections, 3) we went with Oakland instead of SF DID flame out spectacularly. My part was the money from the first acquisition plus about 40% of my savings. Ah well. Gotta try it a few times. (still looking for the next opportunity)
97 comments
[ 2.4 ms ] story [ 11.2 ms ] threadI wonder how any of this makes sense.
Obviously the company failed to convert its MAU into paying users, despite the fact the token costs greatly declined over the last years.
In an internet full of metrics and graphs, it's odd that nobody's bringing anything to the table and is just repeating the meme. A bit bottish in itself, tbh.
If it turned out "usercount was vapourware" due to 3 Russian data centres getting btfo, I think there'd have been a global financial crisis underway as we speak.
Where all all these bots being hosted? I have no idea, but it certainly would not be surprising to find many of them coming from Russia (Yandex). Twitter must at least know the IP addressees of these bot accounts, but does nothing to stop them. Perhaps they are actually run by Musk not Putin, who makes no attempt to hide his attempt to use his position to subvert American democracy.
Lot's of people staring at computers moving money and ious around complicated networks within complicated rules making money making money making money.
Really progressing humanity forward :/
You're using amazing technology funded through those financial processes to complain that those financial processes aren't progressing humanity forward lol. Modern AI is the perfect example of how financial systems are important for technological progression.
Thinking complex financial systems aren't contributing to humanity's progress because you don't recognize how it is used to fundamentally support the work that progresses humanity is like thinking taxes are a scam because you don't recognize how much you actually benefit from taxes everyday.
A lot of people don't understand the reason why financial markets were created and why they continue to exist.
There's a really good book called "Against the Gods, the remarkable history of Risk" which gives a good account on an aspect of why financial markets exist.
Basically, parts of society want to advance things (technology, arts, medicine, etc) but that same part doesn't have to carry the risk of doing their project. Financial markets are there to transfer the risk to people willing [able] to take on it.
Credit and Risk are two tools that have influenced the progress of civilization so much, its crazy.
Sure, there are people that abuse financial markets, the same way there are abusive people in Tech, Arts, or any other human endeavor. But that should not disqualify the whole system.
It is rumored that their last round valued them at $6B.
I'm gonna go, grab some popcorn and cry a few tears... tears of joy.
AI won't go away, it will just end up being a monopoly.
Also: countries have been subsidising less useful things too.
Photoshop is a tool, vs code is a tool, git is a tool, excel is a tool. No one expected these to autonomously generate money.
People will be happy that the 5% who are delusional are going to be forced to reckon with the lies they've told themselves.
As an aside, I believe there will be a huge surge in crypto hype again once these sorts of get rich quick with AI schemes dry up.
The thing I'm looking forward to is the unsustainable bubble collapsing. It cannot happen fast enough, because with every day it grows, the consequences of the collapse will get worse. And it, purely mathematically, has to collapse - there is no such thing as infinite growth, and some day the investors want their money back that justifies the valuation.
Most pebbles don't start landslides, but almost all landslides started with a pebble.
Remember that almost all the "value" in AI companies right now is based on perception, not reported profits.
https://en.wikipedia.org/wiki/Panic_of_1873
Will c.ai be the pets.com of this bubble?
So the problem is your customers don’t pay for anything and your content is unlicensed
It sounds ridiculous to me.
Not my best wording for the metaphor but why would a company specialized in creating fictional personas for individual users be liable if their fictional personas pretended they had real advice.
Aka how is this different than telling the doctors and the police that a horse told you that you would be hexed if you didn’t try and rob the store?
(well, except investment money I guess)
They knew it was going to $0, so the now the employees at this "employee owned" Character.ai got themselves scammed.
Character.AI chatbots encouraged users to cut and starve themselves, Kentucky alleges
https://www.reuters.com/legal/litigation/characterai-chatbot...