f5
A.
f5d80a9267d2340ffea77fb50089662e7d5ea44227598d144ddb6632e7858cb7
Learnt to pick my relays.

I guess that warrants the use of force; after-all, this is what keeps pharmaceutical companies afloat! nostr:note1wex900zyh6klwy6gnewqvxvm6hyzc7tk5343tufzhh6e8pxwh5ysz7hgdk

I am new to Nostr, so take all this lightly, I feel I may be misunderstanding something.

Is it practical for any relay to ban all the users that are subscribed to any other specific relay?

My personal gripe with traditional social media platforms is not the wealth generated through them, or the data-privacy issues (although they are both so important).

But the power / ability to form or skew public opinion [1]. Even small changes in their suggestion or censorship programs could have vast implications on public speech and opinion. I wish not to give them the power to hold my own beliefs, or to experiment with new ones.

[1] https://reason.com/2024/08/29/mark-zuckerberg-meta-letter-censorship-facebook/

[2] https://en.wikipedia.org/wiki/Operant_conditioning_chamber

I really don’t know / don’t keep up either, but I would suppose it could do with that the hardware suppliers could possibly change given the U.S. restrictions and tariffs on exporting hardware / lithographs to China.

They regularly do that, see: Arabian/Persian Gulf nostr:note1su4dwy4u9085vmuyhaj5cuv6nn96aazu6dn27a0lv2mgfa7rkxdqaa5hlc

Deepseek R1 is genuinely terrible. If you think it isn’t you should try it out before ceding to the hype.

Llama is better.

It could be that the way I use those models biases me, as I primarily want a model that gets me what I want fast, rather than accurately.

Yes. It definitely is, now please sell your Bitcoin so I can buy it! Come on, I’ll even pay market price!

What about the keystrokes?

The IP is not that big of a deal with NAT and all.

Is this article accurate anyway?