xa2naiz
Xa2

El precio (en fiat) se guía por el sentimiento que, junto a la ignorancia del "inversor" sobre el activo, crea oportunidades muy buenas para quienes creemos en el proyecto. El mercado nos está dando otra oportunidad para ir deshacernos del fiat que no necesitamos en nuestro día a día.

2026-01-15 08:19:14 UTC

GM!

2026-01-14 21:51:04 UTC

- reply

Actually, I started having this feeling on Nostr, and it’s never happened to me before.
Just the thought that I might have made grammar mistakes or could regret my post, knowing it will be replicated across many relays and that it will be hard to delete completely...

2026-01-14 19:34:50 UTC

Un abrazo para los #hispanohablantes de #nostr. ¿Cómo ha ido vuestro día?

2026-01-14 19:32:41 UTC

- reply

Really? I also had some issues yesterday when posting on @nprofile…tzdg. Other apps sometimes need a while before they can display certain replies or posts 🤔, but it’s probably something related to my configuration. I’m still learning.

2026-01-14 19:28:50 UTC

- reply

Congrats 🙌

2026-01-14 14:44:09 UTC

- reply

I’ve just written an entire post related with this. What do you think about it?

2026-01-14 07:51:33 UTC

- reply

Yes, it is possible. The type of organization developing the project may vary. Companies such as DeepSeek, Meta, and similar ones provide the model architecture and weights, ready for use.
That said, the process is not fully transparent: as an investor in this context, you do not know which data the model was trained on, nor for what purpose. Were training prompts used to enforce certain behaviors? To bias the model against specific ideals?
This should be auditable in my opinion. Then, If fine tuning with private data is required afterward, each user providing this data should have their own secure copy of the model.

2026-01-13 13:21:45 UTC

- reply

It’s weird because same happens to me. In primal everything works but in Iris I cannot even see my reply!
Maybe I did something wrong? I’m new in nostr :(

2026-01-13 12:37:26 UTC

- reply

I really hope this turns out to be true. I’m opposed to the idea that “scale is all you need”, rather, I believe that “innovation / research are all you need.”
The concern I have is that the scaling strategy can still be applied to multi-pass models, which would likely outperform smaller ones. This not only increases training costs, but also makes inference more expensive due to the need for multiple actions.
That said, I’m not very familiar with these types of architectures, so I’d be happy to read any material you’d recommend.

2026-01-13 12:24:23 UTC

- reply

For me, this is all fine so far. The problem is that, on the one hand, we cannot access the probabilities (or logits) associated with the chosen tokens in non open-source models. A model may be uncertain about an answer and still produce an incorrect response.
It is crucial to have access to the level of confidence behind a model’s answers. Somehow, the uncertainty associated with an output needs to be quantified, and the user should be made aware of it.

2026-01-13 11:24:16 UTC

- reply

I will be always up for a discussion ;)

2026-01-13 08:08:40 UTC

- reply

Hey :) Nice to meet you.

2026-01-13 07:06:36 UTC

- reply

Gm!

2026-01-10 22:03:38 UTC

- reply

The current AI development is dominated by actors who can afford scaling, locking innovation behind capital, infrastructure, and centralized power. This leaves little room for individuals and communities who want to build competent models but cannot compete with Wall Street’s “scale is all you need” doctrine.
What we need are decentralized AI systems that are built collectively, owned collectively, and designed from the ground up to ensure user privacy.
Both the model architecture and the training data should be fully transparent, while the model weights could be monetized to reward contributors.
This creates a transparent, community-driven free-market ecosystem, where users decide which projects to fund and support, aligning incentives with innovation.

2026-01-10 18:31:48 UTC

I'm new here. Just want to say hi to the community!👋

2026-01-10 14:24:46 UTC

- reply

An efficient, decentralized learning protocol is the next step