Comment by himata4113
10 hours ago
Engineers at google have publically stated that the models are too big and are far from their potencial. Glad they're being proven right with every release.
They continue to focus on smaller models while openai and anthropic are increasing compute requirements for their SOTA models.
Given the cost increase associated with this model, and previous model releases, I think the size is trending upwards, not down.
The speed says otherwise. I think they're increasing costs since they want to start seeing ROI.
Those are (mostly) new, faster TPU
3 replies →
> Engineers at google have publically stated that the models are too big and are far from their potencial
Can you link to a source?
I wish I could, it was one of those youtube podcast type interviews with one of the engineers, there was a lot more shared, but that line stuck with me the most.
Source please cause i dont believe that for once second
Don’t let that fool yourself. Google will have SOTA models as big as or even bigger than their competitors.
They are just refining their current models while they finish training the next generation.
They will all come out at about the same time. Anthropic, OpenAi, Google, xAI
Anthropic has been sitting on Mythos for a while now. I guess they don't feel pressured to fuck it ship it until anyone else gets a 10T to work.
According to people who have access to Mythos, it is slightly worse than GPT-5.5-xhigh. At least for security tasks.
Hold on, I think this claim needs some hard data. Here you go gentlemen:
https://www.aisi.gov.uk/blog/our-evaluation-of-openais-gpt-5...
4 replies →
Anthropic can sell Mythos to Fortune 500 companies and bypass the average user. I'm not sure how much is hype but I see things like this https://blog.cloudflare.com/cyber-frontier-models/
It's doubtful they have the compute to make mythos publicly available even after the SpaceX datacenter deal. And why sell it publicly if people are still willing to pay for Opus 4.7?
I suspect that Mythos doesn't have a business model that works
Google’s pro models are almost certainly bigger than Openai’s lol
Why would that be? I am curious why do you think that.
E.g. because they are behind on research and so must compensate with size to achieve similar level of intelligence. At least this is what I heard.
For intelligence/size only OpenAI and Anthropic are the frontier. Google has more compute so it can compensate for that with size of the models...
1 reply →
Because TPUs are more efficient, and its cheaper for them to field them in higher quantity since they own the chip.
I mean, yes and no.
Nobody really knows the answer to which one is more optimal
* Large model trained on a large amount of data across multiple domains, that doesn't need any extra content to answer questions.
* Smaller model that is smart enough to go fetch extra relevant content, and then operate on essentially "reformatting" the context into an answer.