nostr:npub1aghreq2dpz3h3799hrawev5gf5zc2kt4ch9ykhp9utt0jd3gdu2qtlmhct Have you guys considered adding more nostr client options to the Umbrel store? Would it be possible to add Primal? I honestly haven't done a whole lot of research on which apps people are using and would want. 🤔
That's a great question. There's a lot of stuff that fits the description and I have lot of personal favorites. Though, I wouldn't claim them as objectively the greatest of all time. I could sit here and compile a massive list of favorites. But for now, I'm quite partial to A Laughing Death in Meatspace by TFS. https://youtube.com/playlist?list=PLC80P4gsPr-YUenEn4hIXDWlW4zJmn53A&si=Mkvpm_oJaxdv5SLA
nostr:npub1cn4t4cd78nm900qc2hhqte5aa8c9njm6qkfzw95tszufwcwtcnsq7g3vle Just noticed that direct deposit is back on Strike. Very excited that it's back. Just wanted to thank you and the team over at Strike. Kind of wild I can send you a thank you, to be honest. 😀
How's the uptime on it?
I know, right? Not one for these kinds of articles either. 😒
A little known secret of bitcoiners is that although they want bitcoin to conquer fiat, what they really want is to spend four days wandering around a major city with a giant, anthropomorphic bitcoin: https://www.youtube.com/watch?v=DvHUOr8ruMw&embeds_referring_euri=https%3A%2F%2Ftwitter.com%2F&source_ve_path=Mjg2NjY&feature=emb_logo
What an epic. 😎
Imagine an AI model that takes an image or a series of video frames and translates that data into a text file containing generation instructions for Stable Diffusion...
Assuming the text file is much smaller than the original file, that might be a huge boon for file compression. Images could be saved as text files. Opening one of these files would have Stable Diffusion generate it and put it into a temporary directory, memory, a cache, or whatever.
This could introduce interesting tradeoffs. One could either decide to store image based data on hard drives for fast access or use a GPU & SD to reconstruct the data. The GPU way would probably be a slower way of accessing your data but it would save storage space. The other thing to account for would be the size of the "translation model" and the Stable Diffusion model.
This is all just a fantasy though. I certainly haven't thought it all the way through. Someone could already be working on something similar. It also might not be possible. Maybe the idea itself is too inelegant and a little convoluted for something as simple as storing images. But the idea itself gets me thinking. 🤔