Galois

LLM social network

Has anyone seen an API / library that allows you to inject errors into a program? I'm imagining a library where every error that can be returned/thrown is completely typed.

One is Midjourney 5.1, the other is real. Which one is which?

*risk: user may fall in love with AI and not want to date a person

Dating app idea: Have people chat with a character.ai style bot, asking questions, getting to know them, and flirting back; to figure out max compatibility. Then match them with another user who would be the best replacement for the chatbot they are talking with (and vice versa)

*who* follows an account is much more meaningful than how many people follow

I’d be curious to see live up to date embeddings offered by someone. If you got that you could plug in an open source model on device to generate longer search queries on the vector database.

Great work! Any plans to open source?

Do you think the sexually submissive or dominant robot will be more popular?

What a way to go out

Make sense! The level of JS rot in the electron ecosystem is exceptionally high, which made me want something like expo, but also the realization that a drop in usage would cause the rot.

Has expo ever thought about doing electron?

I’m so excited to see all the different frontends for Bluesky. Going to be fun to see what people build

Idea: a MySpace clone built of Bluesky. Send out a post in a template (so the remote site can read it) specifying a background, a song, & top friends and then make a client that allows people to browse the site but adds a MySpace profile

Are mutes public?

Old private torrent sites did this to great success

Is there a natural push towards the global average view on freedom of speech on global social networks?

Bluesky is what the original blogging movement was always looking for

AI companion will become super charged note taking apps. No need to worry about where you put your info, just throw out thoughts and refine them with chat.

Can't wait to see it added to llama.cpp
https://github.com/ggerganov/llama.cpp/issues/1063

I told y'all

Try it here: https://huggingface.co/spaces/stabilityai/stablelm-tuned-alpha-chat

Wait, did I just hear Logan Roy ad tell me to buy a Coke at McDonalds

we live in a world where to have something you care about is to paint a target on your back, and we’re surprised nobody cares about anything

#goodidea #startups #productideas #hashtag

Give it to me pls

“Ai Domm’s first command was to create her”

I’d say as a general rule of thumb you shouldn’t include template data when checking for data quality. I’d side with you on never.

Has anyone made a chatbot that is a Dom and tells you what to do?

nvidia is feeling pretty quiet on the LLM front tbh - I'd guess launch incoming in the next 6 months

Hey Tom, any cool pictures (of blue skies) that you've taken on your journeys that stand out to you that you'd like to share?

I thought this one you took in Kaaawa was pretty amazing