#The future of AI

1 messages · Page 1 of 1 (latest)

candid monolith
#

A lot of stuff has been going on with AI with perhaps it being able to do 1st grade math, ai video improving a lot, and the use of ai generated text and video in robotics.

Different advancements and their implications for Neuro Sama:

  1. Ai video
    Summary of recent developments in ai video generation:
    https://www.youtube.com/watch?v=vc04Jp8sjW0
    First we saw the rise of AI images, and it improved steadily over time with more people learning how to use it with more refined control and with better tools. Now we see the rise of AI video, and it is looking pretty good. This raises a lot of questions.

What is a vtuber anymore if anyone can animate anything? Should Neuro take advantage of this technology in some way? Gain more range of motion in her 2d model than would otherwise be possible? Have her rants be turned into short films? Have an animated background that she can interact with?

  1. Synthetic data
    https://www.youtube.com/watch?v=o9e0wAOK7JE
    The idea of synthetic data is you generate your data set with ai. So AI video is not just making content, but it can be used to make AI smarter.
    In a sense AI generated video can give AI "imagination" to imagine things that never happened, and then perhaps answer questions about these events. It will be able to make predictions about what might happen if certain actions are taken. This could probably help with games, but would also result in more intelligence (probably).

  2. The power of language.
    We laugh at the things that Neuro says, but language can also be used as a control mechanism. More and more we see language models being used to improve robotic intelligence. Language is used to plan and execute tasks. I wonder how much Neuro is already doing this.
    https://palm-e.github.io/

  3. Embodiment
    Being in the world is a way to learn about the world. So give AI "bodies". Right now Neuro controls a 2d model. In VR there is potential to interact with a variety of environments and scenarios.

What are your thoughts?

YouTube
Matt Wolfe
An EXPLOSION of AI Video is Happening Right Now

It's crazy what advancements we're seeing in the world of AI Video! Thanks to TubeBuddy for sponsoring this one! Learn more about TubeBuddy here: https://www.tubebuddy.com/tools/suggested-shorts/?utm_source=youtube&utm_medium=video&utm_campaign=mw_yt_suggestedshorts

Discover More From Me:
🛠️ Explore hundreds of AI Tools: https://futuretools.io/...

▶ Play video
YouTube
AI Explained
Not Slowing Down: GAIA-1 to GPT Vision Tips, Nvidia B100 to Bard vs...

From GAIA-1 and UniSim showing that new worlds can be imagined with synthetic data, to Nvidia B100 and X100 suggesting no slow down in compute, this video will argue that AI is not slowing down. I’ll cover what that means in Robotics, Audio, Vision and even fraud, from Disney to LLaVA, Elevenlabs to exclusives from The Information (Nov 6 OpenAI ...

▶ Play video
PaLM-E: An Embodied Multimodal Language Model

Project page for PaLM-E: An Embodied Multimodal Language Model.

vocal trout
#

That's all very cool and interesting, but keep in mind that as far as we know staytuteled is the only one working directly on Neuro. He doesn't want anyone to know too much so it's difficult to say how things work currently. So yeah I doubt posts like this will help him at all neuroShrug But it's cool to talk about and imagine what could be if Neuro was a huge state-of-the-art project developed by the best AI scientists in the world neuroHypers

past rivet
#

with how fast AI R&D is going, expect for things to be radically different next year

#

OpenAI is apparently sitting on something big