On: Kimi K3.1 Pro and K3.1 Pro Highspeed

[delis-thumbs-7e]

Why would they not release those models, since it would help them become profitable and beat the competition becoming the one company rule them all? Oh yeah, they are so dangerous. Well Anthropic revenue comes from only few big corporate customers, so just make them Golden Partners etc. and give them access to super-models. I doubt Google and Meta wants to crack web security.

Besides, if we start imagining super models US companies have in secret, why would we expect Chinese companies have them? Why would not Moonshot have their own super intelligence hidden in the cellar?

[satvikpendem → delis-thumbs-7e]

Vertical integration. They want to control all sectors, that's why OpenAI and Anthropic are slowly rolling out solutions for law, medicine, design and so on.

[disgruntledphd2 → satvikpendem]

This is just cope on the part of their funders.

That being said, Claude is extremely pleasant to use and work with, but it's not 10x as good as the pricing suggests.

[satvikpendem → disgruntledphd2]

How is it cope when that's literally what they're doing?

[disgruntledphd2 → satvikpendem]

It's delusional. Like so many failing tech companies claim to aim for this and it basically never works.

Ads might save them, but they'd need to shift their company structures significantly to make that work.

[satvikpendem → disgruntledphd2]

Why are you comparing these companies to previous tech companies (at least some of which have, by the way, been able to vertically integrate, like Salesforce and SAP) when AI is completely different from what came before? The real delusion must be thinking it's still a fad like Paul Krugman thought of the internet.

[disgruntledphd2 → satvikpendem]

Because they are tech companies?

LLMs are super useful tools, but their valuations are based on building the machine god, and it doesn't look like that's going to happen.

Instead, it's looking like a commodity market which is great for users but very very bad for investors.

Like Claude is great, but is it 10x better than GLM 5.3 or 20x better than DeepSeek?