How do you know if it’s open source? Well if it’s called something like “huggingface” or “redpajama” there’s a very good chance it’s made by people who have no marketing department. So good odds it’s free.
ChatGPT is pretty crap branding too, for the record. They just somehow managed to mainstream it. All the LLMs after it try to have cooler names (Bard, Copilot, etc.) but the kludgy first name is still better known.
I personally disagree, Bard feels very uninspired, and copilot i associate too mich with flying, and also sounds more competent than it is.
ChatGPT is probably not the best name, but at least it’s unique.
GPT in French is literally pronounced the same as “I farted”.
Cat I farted
My life is a little better knowing this fact. 😄
Having a huge backing (i.e. OpenAI) helps…
“Sure, it is not perfect. But, sometimes it is incredibly helpful. No matter what you do with it, unfortunately, it is not an open-source solution.”
This article needed a better ai to write it .
Not mentioned in the list, but a project worth keeping an eye on:
“llamafile: bringing LLMs to the people, and to your own computer - Introducing the latest Mozilla Innovation Project llamafile, an open source initiative that collapses all the complexity of a full-stack LLM chatbot down to a single file that runs on six operating systems.”
I’m honestly more interested by the OS-agnostic executable than by the LLM here. How?
Here’s the answer, but I have absolutely no idea what it means…
https://github.com/jart/cosmopolitan
Cosmopolitan Libc makes C a build-once run-anywhere language, like Java, except it doesn’t need an interpreter or virtual machine. Instead, it reconfigures stock GCC and Clang to output a POSIX-approved polyglot format that runs natively on Linux + Mac + Windows + FreeBSD + OpenBSD + NetBSD + BIOS with the best possible performance and the tiniest footprint imaginable.
Six?
Edit: they counted Unix 3 times.
Seems to me like they counted Unix 5 times.
I can tell its have huggingChat in list without even clicking the article.
Hands down the easiest
I run Kobold locally, it is awesome
All this talk about llamas and red pajamas…
Here is an alternative Piped link(s):
https://piped.video/shorts/A6LOVMymJhs?si=fcPt5pr8R2bCxhYi
Piped is a privacy-respecting open-source alternative frontend to YouTube.
I’m open-source; check me out at GitHub.
Good! I’m looking to ditch most search engines (with the possible exception of Searx) since they have become so inundated with so much junk links. Louis Rossmann mentioned in one of his videos that he pays $20/month for GPT-4 since it fetches better results. But I’ll look into this before I do so. Thanks for sharing this.
Only show interest in open source software,
let the anti consumer surveillance tech companies drown.For SearX,
checkout these random SearX redirectors:- GimmeASearX - Self Hosted
https://github.com/demostanis/gimmeasearx - SearX NeoCities - Publicly Hosted https://searx.neocities.org/
Both increase your privacy + prevent rate limits.
Love the sentiment, and I agree, but anti consumer surveillance tech is here to stay, sadly. Can’t tell you how many people in my life have Alexa, FireTV and random shit like that.
Can’t tell you how many people in my life have Alexa, FireTV and random shit like that.
This doesn’t mean that you have to surrender to it.
Little baby steps, at least we can make the difference for ourselves… But Google is in my walls
Why don’t you take google out of your walls?
It’s threatening me with breaking my house’s structural integrity
Oh yeah, load-bearing Googles are a trap
I was thinking the same until @Rikj000 responded to my comment. The defeat of the surveillance state (both private and government) relies on us.
Thanks for the tip. Dodged another proprietary bullet.
- GimmeASearX - Self Hosted
Kagi is the only search engine I use which has really good results and no junk links. …and you have to pay for it, of course. It’s a meta search engine but they use their own indexes for news results and Teclis, which indexes small commercial sites with fewer than 5 trackers. One of the cool features it added recently was an icon for identifying paywalled articles.
I’d like to recommend Mojeek, my default search engine, but it still has a way to go. If you’re just looking for an “answer engine” rather than a general search engine…I guess an LLM probably isn’t a bad place to start?
Thanks for mentioning us nonetheless! You can help out with that journey, if you want, by chucking some searches (either new ones or old ones you remember being not so great) into the Evaluation Page and voting.
It’s good to see you guys on Lemmy :)
I tested it a bit a few days ago, but I’ll see if I can give it a more rigorous go today. The ones I’ve found Mojeek to be weak in are bug strings that programs I’m working with spit out. Although I think I’ve had more luck in the past few months.
I’ve used GPT4All, and it’s one of the easier ones to get up and running I found. Everything just works out of the box.
