Photos would need a lot of work to rival Photomator.
If they're essentially shutting down Photomator development, after doing similar with Aperture many years ago, they do seem very determined to drive people to Lightroom....
It increases the countries' soft power if people around the world watch content from there.
Eg. a self-reinforcing cycle that you get the best from the other immigrant categories arriving because they choose the country where everything "seems to be happening".
I think the era of buying templates is over, when you can get a tool that listens to you patiently, iterates again and again till you're satisfied for pennies, why would you pay hundred's for a template that is there for anyone else to buy as well.
The selling feature is that it's more polished (and has good accessibility etc), they're still intended to be customised, which you could use AI for. Why use Tailwind itself when you could generate one with AI? Because it's solidly tested and polished, similarly.
But the broader, more important point: an open source project previously could be funded by using attention to sell other services or add-ons. But that model might be gone if users no longer visit or know the creators.
Google was not profitable until they rolled out ads, either.
The scope of use of AI assistants in people's lives are significantly higher than google search, imo. People use it in far more scenarios already than just information retrieval. That's why some are betting there's a chance it's more valuable than present-day google search.
IIRC Google had no monetization at all until ads. Even then the cost of providing search with ads is orders of magnitude lower compared to running LLMs.
They made money by licensing their search technology to other sites, as well as selling physical search appliances for businesses. They were considered by some to be struggling to find a way to monetise well.
Computational cost is indeed higher than search (though remember, search has been heavily optimised for many years!), but search and web companies were one of the lowest cost, highest-margin businesses in human existence. Many higher-cost businesses have been supported by ads.
>Many higher-cost businesses have been supported by ads.
Not at the scale of a trillion dollars, though. You can't make that kind of money back with eyeballs. You either need government subsidies or insane vertical integration. And if your program threatens to neuter the GDP of a country, I don't know how long subsidies will last. At least not in a democracy. People are so mad about immigrants taking jobs, and this would be 10 times worse (and bipartisan, eventually).
Even then: we're quickly hitting a resource wall as well. Are we really going to go to war just so we can have some dude generate AI sheep memes? Something's got to give.
It's not only the computational cost though. Hardware requirements are much higher and GPUs need to be replaced every 2-3 years. Plus model training expenses which are considerable. I imagine it's easily 100x more expensive and the margins (if there's ever any profit) will be very low.
People did say the same thing about Youtube, which was unprofitable and extremely expensive to run in the early years. I remember thinking everyone would leave once ads were added.
At youtube's ad income rate (~$13/year), the current (but growing) ~800 million chatgpt users would add ~$10 billion. At facebook's rate (~$40-50/year) $32-40 billion. Potentially, an assistant would be more integrated into your life than either of those two.
The "audience retention" is the key question, not the profitability if they maintain their current audience. I've been surprised how many non-technical people I know don't want to try other models. "ChatGPT knows me".
The network effects aren't the same. All the viewers watch youtube because it has all the content, and all the creators post on youtube because it has all the viewers.
How can a model achieve this kind of stickiness? By "knowing you"? I don't think that's the same at all. Personally, one of the reasons I prefer Claude is that it doesn't pretend to know me. I can control the context better.
the problem with the YouTube analogy is that media platforms have significant network affects that NN providers don't. OpenAI can't command a premium because every year that goes by the cost to train an equivalent model to theirs decreases.
Youtube didn't either at the time. The front page was widely seen as garbage, and everyone I knew watched videos because they were embedded or linked from external sites. "If they introduced ads, people will just switch to other video hosts, wont they?". Many of the cooler creators used Vimeo. It was the good recommendation algorithm that came later, that I think allowed an actual network effect, and I don't remember people predicting that.
The field is too young to know what will keep users, but there are definitely things that plausibly could create a lock-in effect. I mentioned one ("ChatGPT knows me") which could grow over time as people have shared more of themselves with ChatGPT. There's also pilots of multi-person chats, and the social elements in Sora. Some people already feel compelled to stick to the "person" they're comfortable talking to. The chance of OpenAI finding something isn't zero.
That's a bit revisionist. Network effects were obvious when Google acquired Youtube. Google Video had the edge technically, but it didn't matter because Youtube had the users/content and Google saw that very clearly in their user growth before they made their offer.
I'm not sure about it having the edge, I thought Google video had a worse interface between them at the time. But that point feels eerily relevant anyway: a lot of normal people I see don't care if Claude/Gemini/etc are better models technically, they're comfortable with ChatGPT already.
A lot of YT's growth at the time was word of mouth and brand among the population, which is currently ChatGPT's position.
ChatGPT is losing their brand positioning to Google, Anthropic, and Chinese Open Source
Altman knows this and why he called code red. If OpenAI hasn't produce a fully new model in 1.5 years, how much longer can they hang on before people will turn to alternatives that are technically better? How long before they could feasibly put out a new model if they are having issues in pre-training?
They're losing their benchmark lead to those companies. But no chance that your average user is even aware of Anthropic, much less OSS models. The brand is mostly fine IMO, it's the product that needs to catch up.
Maybe ChatGPT is sticky enough that people won't switch. But since we're talking about something as old as Google Video, we could also talk about AltaVista, which was "good enough" until people discovered a better and more useful alternative.
A lot of "normal people" are learning fast about ChatGPT alternatives now. Gemini in particular is getting a lot of mainstream buzz. Things like this [1] with 14k likes are happening everyday on social. Marc Benioff's love for Gemini broke through into the mainstream also.
Youtube didn't have a significant competitor, once the quality started declining and the ads started creeping up, there were no alternatives to switch to (as a user) because the content creators were in on the profit.
The same isn't true about ChatGPT.
Anthropic and Google provides a similar product, and switching to a better/cheaper platform is fairly easy as it only depends on you and not on others (content creators or friends) doing the same.
It also didn't generate billions in income before even adding ads^1, nor grow anywhere near as quickly. ChatGPT is larger on most axes.
YouTube was ambitious for its time. "In 2007, YouTube consumed as much bandwidth as the entire Internet had in 2000" but they weren't believed to start breaking even until 2015.
Whether it "generated billions" is a wrong angle to look at this. What is relevant is the relation between the spend, or the committed spend and the income. I don´t believe that YT at any point committed to investing literally 1000x of its yearly revenue into a partner company, nor do I remember its CEO using made up words such as "annualised revenue" that keep being spat out by both OpenAI and Antropic CEOs, in the sense of "projecting the max monthly revenue as annual and fooling the investors".
I suspect some of the downvoters hate the idea of ads, which is understandable.
But a lot of HN users use gmail, which has the same model. And there are plenty of paid email providers which seem far less popular (I use one). Ads didn't end up being a problem for most people provided they were kept independent of the content itself.
1. Yes, we're also talking about ChatGPT's free plan here too.
Ads could fund more quota or bigger models for users who don't wish to pay (and/or just make it more sustainable)
Google will almost certainly be doing this with Gemini, and if ChatGPT can't offer as much it leaves an easy reason for people to switch.
2. It does have ads in the default interface, though they're quite unobtrusive. You might also have a blocker. But yes, I suspect their size allows them to provide it mildly "at a loss" to support their ads elsewhere.
Pure fact-based, niche questions like that aren't really the focus of most providers any more from what I've heard, since they can be solved more reliably by integrating search tools (and all providers now have search).
I wouldn't be surprised if the smallest models can answer fewer such (fact-only) questions over time offline as they distill/focus them more thoroughly on logic etc.
To be entirely honest, yes, American leadership is currently very childish while Chinese one is everything but childish. And the simple observable consequence is that China is winning whatever pissing contest is going on while America is busy shooting itself into own foot, applying bandage and then claiming it won cause it is not bleeding anymore.
Yes, though money is still a big part of it. An Australian company (Lynas) developed the capability but was struggling to get investment largely because they couldn't produce them as cheaply as China's scale/subsidies/etc could.
When Japan was temporarily cut off from rare earths they became an investor (willing to pay more to reduce single-vendor risks), but apparently it was hard to get the US at the time to care enough. At least that's the story that was floating around.
If they're essentially shutting down Photomator development, after doing similar with Aperture many years ago, they do seem very determined to drive people to Lightroom....
reply