Mika

Real. Also: AI bad. If you think that AI is actually useful you’ll be downvoted to hell and below.

You do realize that reactive jet shahed interceptors are also using starlink and also hit the speed limit wall?

What we call hallucinations is operating on a lot of assumptions while you don’t have enough data, instead of saying “I actually don’t know what am I doing”. People do it too.

It’s not like that’s not a problem, sure, but the further this tech goes, the better it is at not hallucinating hard.

Confidently incorrect. It has reasoning layers. It creates new text based on the context of the previously written conversation. That conversation could be in the context of things that weren’t in it’s training dataset, and you still will receive some sort of an answer.

You don’t need a datacenter to run LLMs. A decently strong PC can run useful stuff already. The real fight is not anti-AI, the real fight is locally run AI (means of production you own) vs datacenter AI (means of production you rent, that can be restricted any moment).

AI is means of production today.

Positive line (for them) is that it helps russia.

Kinda the opposite. Openrouter exist and you can host any model, but even within the model there are bunch of params that affect how good is the output you get.

And while most hosts list these upfront, & you just gotta be smart enough to request with specific requirements, I’ve heard there are still good reasons to just stick to a list of trusted hosts sometimes, I don’t really remember what was the issue, quantinizastion level that they aren’t honest upfront maybe.

Kimi k3 has 2.8T, so it’s not even the chonkiest open weights model now

You don’t even need to attack, they are open weight. Distill as much as you want.

Just don’t read the mails and you won’t get fished

Futureposting