People underestimate the lead OAI has with their post-5.2 models. The author does not strike me as someone who closely follows the progress frontier labs make in US and around the world.
It's a joint ignorance of how these frontier models get baked and what consumers want.
Many pundits think it's just a matter of scraping the internet and having a few ML scientists run ablation experiments to tune hyperparameters. That hasn't been true for over a year. The current requirements are more org-scale, more payoff from scale, more moat. The main legitimate competitive threat is adversarial distillation.
Many pundits also think that consumers don't want to pay a premium for small differences on the margin. That is very wrong-headed. I pay $200/month to a frontier lab because, even though it's only a few % higher in benchmark scores, it is 5x more useful on the margin.
> They have internal scale and scope economies as the breadth of synthetic data expands.
These frontier labs will have a hundred or a thousand teams of people+AI working in parallel generating synthetic data to solve different niches. A few teams solve computer use. A few teams solve math. A few teams solve various games. So the org is basically a big machine that mints data, and model research is only a small part of it. Scale then is the moat.
The second leg of the moat thesis is that open weights competition will die off soon because the cost to keep up with the scale will be too excessive.
The third leg of the moat thesis is that customers are happy to pay big margins for differences that appear small if the benchmark is the measuring stick.
If the paradigm was still scrape internet -> train model, I'd agree that there is no moat.
I disagree that the model is a moat; distillation of models is going to happen, and even without it all the current players have models that are virtually indistinguishable for the use-case.
Model capbilities have converged over time, and I don't see this trend reversing. OpenAI owns only the model.
The provider who does have a moat is Google - they own the entire vertical, from the hardware, to the training data, they have it all.
OpenAI has to buy GPUs, Google makes them.
OpenAI has to rent data centers. Google owns them.
OpenAI has to scrape the web for all training data. Google's collection of user emails (not counting their Android data harvesting, ad data harvesting user-tracking, etc) alone gives them a ton of training data which will never be available to scrapers.
Google has billions of signed-in users, OpenAI has to market to and attract users (800m user count last I checked, but also last I checked that growth was asymptotic and flattening out).
Thats what a moat looks like. Better technology and/or results has never been, in my memory, a moat.
Look, I'm upvoting your posts in this thread because you make some good points, but I'm not really convinced that a) synthetic data will result in good models, nor that b) quality synthetic data can be generated by labs outside of those orgs that have a ton of user-info.
This is why I say that OpenAI has no moat - even if synthetic data (however it is generated) is 90% of training data, there are still only two possibilities:
1. Orgs like Google, Microsoft and Amazon have a ton of user-data with which to produce synthetic data (after all, it's not produced out of thin air).
and
2. You don't need a ton of real data to seed the synthetic generation.
In the first case, yes, that looks like a moat, but not for OpenAI, more like for Google, etc al.
In the second case, what's to stop an upstart from producing their own synthetic training data?
In either case, companies who provide only tokens (OpenAI, Anthropic, etc) don't have a moat. The moat is still the same as it was in the 90s - companies deeply embedded into users' workflows.
In my memory, like I said, I struggle to think of even a few successful moats that were technology. The moat is always something else.
I pay OpenAI but I would also be a happy Anthropic customer.
My view is that OpenAI, Anthropic and Google have a good moat. It's now an oligopolistic market with extreme barriers to entry due to needed scale. The moat will keep growing as the payoffs from scale keep growing. They have internal scale and scope economies as the breadth of synthetic data expands. The small differences between the labs now are the initial conditions that will magnify the differences later.
It wouldn't be surprising to also see consolidation of the industry in the next 2 years which makes it even more difficult to compete, as 2 or 3 winners gobble up everyone and solidify their leads.
When people worry about frontier lab's moat, they point to open weights models, which is really a commentary that these models have zero cost to replicate (like all software). But I think the era of open weights competition cannot be sustained, it's a temporary phenomenon tied to the middle-ground scale we're in where labs can still do that affordably. The absolute end of this will be the end-game of nation state backed competition.
I am smoking this thing called: putting same prompt in four different apps and seeing which ones give me answers and which ones hallucinate and patronize me, but considering your comment I can see how you would prefer ChatGPT
Having the same experience during development of my MCP App. ChatGPT is by far the worst, slow, hallucinating or just quitting. Claude is the best with amazing results and Mistral Le Chat surprisingly good.
Have you tried actually holding a conversation with it? I'm really puzzled in which world Gemini/Claude is better than ChatGPT for day-to-day tasks/conversations.
Claude can't even search products on Amazon, Jesus.
You know, I just tried to search on Amazon.de and it worked without ChatGPT. Is it a thing with the .nl-tld that you have to use ChatGPT for something simple like that? ;-)
Agreed, compare the frontier models from Google and OAI. It’s like night and day. Anyone who says “the tech has caught up” has not spent even one day using Gemini 3.1 to try and accomplish something complicated.
