Tootfinder

Imagine ChatGPT but instead of predicting text it just linked you to the to 3 documents most-influential on the probabilities that would have been used to predict that text.
Could even generate some info about which parts of each would have been combined how.
There would still be issues with how training data is sourced and filtered, but these could be solved by crawling normally respecting robots.txt and by paying filterers a fair wage with a more relaxed work schedule and mental health support.
The energy issues are mainly about wild future investment and wasteful query spam, not optimized present-day per-query usage.
Is this "just search?"
Yes, but it would have some advantages for a lot of use cases, mainly in synthesizing results across multiple documents and in leveraging a language model more fully to find relevant stuff.
When we talk about the harms of current corporate LLMs, the opportunity cost of NOT building things like this is part of that.
The equivalent for art would have been so amazing too! "Here are some artists that can do what you want, with examples pulled from their portfolios."
It would be a really cool coding assistant that I'd actually encourage my students to use (with some guidelines).
#AI #GenAI #LLMs

I'm quite impressed by Fluxpose. It's making me consider selling half of my base stations (not all cause you never know if I will ever need them again) and get a kit once I get my hands on the Steam Frame.
I've rooted for BS tracking for so long because it was the best option out there, but the downsides it brings can be quite annoying sometimes. Plus, the Steam Frame is fully portable and BS tracking is not.
Also I recently learned that Fluxpose was made by an Spanish…

YouTube is like heroin because it pulls you in, hooks you, and slowly drains you. It is designed to be addictive and destructive, and that is exactly why it feels so hard to step away from it.
And yes, I know the excuse, maybe your favorite dealer has not set up shop on PeerTube. But that does not change the truth of the comparison.
PeerTube is like weed, it grows more naturally, it is gentler on you, and it still gives you a high without tearing apart your health and your lif…

The ARRL has a long online text band plan which they have condensed for HF into The Considerate Operator's Frequency Guide which is not really in a convenient form to print and carry with you into the field when operating portable

Let's be honest. I've been a strong supporter of #OpenPGP (or #PGP in general) for a long time. And I still can't think of any real alternative that exists right now. And I kept believing it's not "that hard" — but it doesn't seem like it's getting any easier. The big problem with standards like that are tools.
#WebOfTrust is hard, and impractical for a lot of people. It doesn't really help how many tools implement trust. I mean, I sometimes receive encrypted mail via #EvolutionMail — and Evolution makes it really hard for me to reply encrypted without permanently trusting the sender!
The whole SKS keyserver mess doesn't help PGP at all. Nowadays finding someone's key is often hard. If you're lucky, WKD will work. If you're not, you're up for searching a bunch of keyservers, GitHub, or perhaps random websites. And it definitely doesn't help that some of these may hold expired keys, with people uploading their new key only to a subset of them or forgetting to do it.
On top of that, we have interoperability issues. Definitely doesn't speak well when GnuPG can't import keys from popular keyservers over lack of UIDs. And that's just the tip of the iceberg.
Now with diverging OpenPGP standards around the corner, we're a step ahead from true interoperability problems. Just imagine convincing someone to use OpenPGP, only to tell them afterwards that they've used non-portable tool / settings, and their key doesn't work for you.
That's really not how you advocate for #encryption.