The moat is in the sheer cost of training models.
At most you’ll have like 5 major players.
At most you’ll have like 5 major players.
5 companies doing a repeated leapfrog game where consumers can flip without much effort seems like the opposite of a moat, unless and until they consolidate into a monopoly.
I'm starting to feel some diminishing returns on the new releases, too. The wow moments are still there but fundamentally we're reaching a point wherein even open models can come close enough on some tasks that we can eschew paid models for not-insignificant pieces here and there.
One could say the same about search, browsers, image/video/text sharing, smartphones and yet, moats were built.
Most of these built moats through widespread adoption by being first or best.
And in almost every case adoption requires an investment by a consumer, who is unlikely to switch because there's a cost incurred. Switching phones, browsers, social media platforms come with real costs that we don't see with LLM vendors, where you can switch one with another with twenty minutes of work.
Unfortunately, these LLMs are marketed to enterprise(B2B) and not regular consumers(only beta test at exorbitant costs).
In world of enterprise, switching out to save costs takes many years or even decades as the user of the product is far removed from the decision maker who buys it.
Google has a deal for Reddit’s data, so its models will have that information. I can see this kind of thing being a moat as well.
And yeah obviously you’re going to feel diminishing returns.
Going from 0 to 1 always feels like a bigger step than from 1 to 2.
I don't see this as a silver bullet but let's assume this is true and a differentiator. This kind of validates that building a robust model hits a data saturation wall at some point and the moat dries up.
Imagine OpenAI makes a deal with GitHub, Stack Overflow, and O’Reilly gaining exclusive AI rights to their content.
Now there’s no reason for anyone in software to use someone else’s models.
Anthropic could make a music generation model with an exclusive license from Warner music or whoever and own the music producer market.
This seems like a pretty straightforward plan assuming those rights could be obtained.
- Meta has its own
- Reddit is monetizing its
- Twitter was bought for its
- MS is trying to create one
- Amazon has a funhouse mirror version (reviews)
I don't disagree with this but this is a losing battle assuming the perpetual cat-and-mouse game of web scraping continues to lean to the scrapers.
To beat it means raising the walls of the gardens, making "data deals," hoping users don't move to the next thing, and aggressively protecting data via third party legal action.
This all strikes me as a very vulnerable and complicated moat.
I don't see a way Apple engineers its way out of that -- but they won't need to if open source models are good enough, or if capable models remain accessible (e.g. not first-party locked).
Ironically, it looks a lot like the Maps situation again...
Also let's not forget how corporations invent doublespeak and twist meanings. For example Apple may not resell private data but use it internally. Or it may use it internally only processed by scripts, and claim it is private because other human never saw it. Many other terminology manipulations can be invented if a stake is a few billions in cash.
I expect, if they do find they need it, they'll pitch it as an opt-in paid feature.
Pay us for these extra capabilities, and know that we're training on your data.
Which it is.
The public web is going to turn into even more of a dumpster fire once genai is fully digested by spammers. (Which is to say, once a 14 year old in a random country with third world GDP per capita can use it at scale)
Just my 2c.
“Fifth, the board determines when we've attained AGI. Again, by AGI we mean a highly autonomous system that outperforms humans at most economically valuable work. Such a system is excluded from IP licenses and other commercial terms with Microsoft, which only apply to pre-AGI technology.”
The moat is still the cost.
That being said, most models are based on llama, so it may seem like a lot of movement is happening, but it’s not, really.
RWKV and deepseek are the only LLMs that’re fully open that I’m aware of.
If meta decided to stop releasing llama, that’s that.
Just because they share a near identical interface does not mean they will always have identical capabilities