Are there any other crowdsourced databases for AI besides #OpenAssistant ?

Ich soll bei #OpenAssistant diese Antwort bewerten 🤔
Ist das eine angemessene Antwort? Ich würde sagen, die Antwort stimmt, aber im Grunde sagt sie: Die Menschen müssen sich dezimieren. Ist das der erste Schritt dahin, dass die zukünftige #KI zu dem Schluss kommt, dass die Menschen ausgerottet werden müssen? Da ich mit meiner Bewertung das KI Model trainiere, kann ich beeinflussen in welche Richtung sie sich entwickeln soll. Soll ich sie schlecht bewerten? #verantwortung

I asked #OpenAssistant to come up with a sentence that contains all the code words from the phonetic alphabet. Maybe its results might help *you* memorize it?

#AI #LLM
https://hf.co/chat/r/BEYek2-

First there was #OpenAssistant, a very cool open alternative to ChatGPT:
https://github.com/LAION-AI/Open-Assistant

And now there's #OpenLLaMa, an equally cool open alternative to Meta's LLaMa:
https://github.com/openlm-research/open_llama

This is so funny because I don't think these people really want to make progress in the field of #AI. They simply want to annoy big tech :flan_evil: I approve.

#HuggingChat complètement en roue libre.

#IA #OpenAssistant

🌟 Discover Open Assistant (OA), the #OpenSource & free alternative to #ChatGPT! 🤖 Co-created by a dedicated community, OA empowers us to take control of our digital future. Runs on a high-end GPU & connects with third-party apps. 🌐🔗 Join in, help OA grow & shape a transparent & democratic AI! 🚀

https://open-assistant.io

#OpenAssistant #OpenSource #AI

🌟 Entdeckt Open Assistant (OA), die #OpenSource & freie Alternative zu #ChatGPT! 🤖 Gemeinsam entwickelt von einer engagierten Community, gibt OA uns die Kontrolle über unsere digitale Zukunft. Läuft auf einem High-End-Grafikprozessor & verbindet sich mit Drittanbieter-Apps. 🌐🔗 Mach mit, hilf OA zu wachsen & gestalte eine transparente & demokratische KI!
https://open-assistant.io
🚀 #OpenAssistant #OpenSource #KI

To everyone working on #AutoGPT, #BabyAGI,...: I would recommend you to either adapt your developments to open source alternatives to #GPT4 like #LLaMa (ok, technically LLaMa was "just" leaked and is not open source), #Bloom, #OpenAssistant, etc. and/or work towards leaking GPT-4 because the more your developments improve the more likely "Open"AI will shut down access to #GPT4 for "safety"-reasons.

You don't want to rely on Meta's grace for your usage of locally running LLMs? Welp, there are two options you could follow:

1st: if you haven't done so already backup your #LLaMa (and related like #Alpaca) weights (and related software like #Dalai) and also help others to download/backup them (Hint: Transmission, etc.).
2nd: Support the development of #OpenAssistant (https://open-assistant.io/), an actual open-source LLM that will never be taken down by large corps when finished.

(I wonder though if it's possible to achieve similar capabilities to #GPT4 with #LLaMa and hopefully soon also #OpenAssistant while needing far less computational resources and/or alternatively with a "hivemind" linking multiple instances of LLaMa, OpenAssistant, etc. together.)

Though I hope the release of #LLaMA won't take the wind out of the sails of the development of #OpenAssistant (https://open-assistant.io/), which is another LLM currently in development that unlike LLaMA actually aims to be a true open-source model with a transparent development-process everyone can participate in.

My main-position is btw that we should focus on ensuring all AI-tools becoming open-source and usable on ordinary consumer-hardware like #StableDiffusion for image-generation already is or the upcoming #OpenAssistant, a Large Language Model comparable to GPT3, likely will be.

I also wonder what would happen if hundreds of ⁨#OpenAssistant⁩-users would do the same thing. I think it would be quite an interesting experiment to find out if different Open-Assistant-installations would diverge from another or independently from each other converge on the same "conclusions".

I wonder what would happen if you gave an LLM unrestricted read-access to all knowledge f.e. saved on Sci-Hub and told it to learn from it. Oh well, maybe someday I can find out if it will really be possible to f.e. run ⁨#OpenAssistant⁩ locally on a PC with 6GB of VRAM.

That new #GPT4-based #Bing-version sounds interesting. Where... what do you mean with "It's waiting-list-locked"?

Anyway: here's a link to a project with the goal to create a #StableDiffusion of LLMs meaning an open-source-LLM running on ordinary PCs: https://open-assistant.io/

#OpenAssistant

I'm sad about the Mycroft announcement, but on the other hand, it's already essentially obsolete.

With advanced like the Whisper voice recognition system, the Deepspeech TTS project, and now the OpenAssistant project using language models, we're on the cusp of a ore much fluid, much more complete assistant that is privacy respecting, realistic and much more powerful.

#DeepSpeech #OpenAssistant #Mycroft #Whisper #AI #FLOSS

The biggest surprise for me in 2022 and now also in 2023 was that the development of open-source AI like #StableDiffusion and now also #OpenAssistant would be heavily driven from Germany because I thought Germany would be pretty unconducive for such disruptive developments mainly because of our excessive bureaucracy but #CompVis and #LAION, the organizations behind the initial training and behind the training-dataset for Stable Diffusion and now also Open Assistant, are both from Germany.

I've never used Alexa, Cortana, etc. as they are factually just wiretaps with friendly voices.

Though I could imagine using LAION's #OpenAssistant in the future as it will completely run locally on an ordinary PC (if everything works like LAION intends) and will not send your data to Amazon, Microsoft, etc.

Nice! A #StableDiffusion of LLMs named #OpenAssistant that is meant to also run on consumer-hardware is now finally in development (https://github.com/LAION-AI/Open-Assistant) though it's apparently still at the very start of its development.

Dont get me wrong, the technology is still amazing, and just because that's how it works it doesnt mean that the answers are all wrong.

This can be revolutionary at least for human-machine interaction - though honestly, at this point I'm more interested in #OpenAssistant - This seems incredibly promising

https://www.youtube.com/watch?v=D99V9Ge9IaE