menu Home

GAME OVER! New AGI AGENT Breakthrough Changes Everything! (Q-STAR)

TheAIGRID | April 2, 2024
GAME OVER! New AGI AGENT Breakthrough Changes Everything! (Q-STAR)

Written by TheAIGRID


This post currently has 27 comments.

  1. @jimt7045

    April 2, 2024 at 8:26 pm

    There is simply no point in slowing down research into AI. And analogy would be where someone is going to climb a cliff. If the cliff has to be climbed, and in this case it will be cllimbed, then with sensible safety precautions in place, there is no point in hesitating before you start the climb. Get on with it!

  2. @user-zs8lp3lg3j

    April 2, 2024 at 8:26 pm

    Loss of Autonomy—the fear of being immobilized, paralyzed, restricted, enveloped, overwhelmed, entrapped, imprisoned, smothered, or otherwise controlled by circumstances beyond our control

  3. @Dave-Goodwin

    April 2, 2024 at 8:26 pm

    Unfortunately during that woman's TEDTalk she didn't go into what constitutes safety. In our current environment it's not too hard to see how that can mean supporting someone's narrative.

  4. @TheMillionDollarDropout

    April 2, 2024 at 8:26 pm

    I’ve been watching for months. For the first few weeks, I would go to your channel, but then kind of abandon it because it got to me how you had these ADHD tiktok brain subtitles on that went rlly fast and appeared with every word and I saw other people complaining so I had kind of like a confirmation bias, whatever lol but then I just kept listening and I was like man he’s bringing that information and I love that shit and I love how fast you want to talk about all of it so thank you AI GRID guy lol

  5. @Hypersniper05

    April 2, 2024 at 8:26 pm

    Ok so you keep the model in training mode but use it for inference and when it does a good job -cross checked by another LLM you train it on that interaction on real time. Repeat this over and over thousands of times and eventually youll get AGI and technically you are keeping that new skill in memory. I have been able to do this but only on consumer GPUs, imagine having more resources

  6. @gregmattson2238

    April 2, 2024 at 8:26 pm

    I don't get it. Gemini has already said they basically have this. Gemini's context window can go up to 10 million words or perhaps longer.

    Don't get me wrong this would be awesome, especially if they did something like hook up an agent to a damn docker image for actually iterating over code. As it stands, that's the major bottleneck to using this stuff here. yes you can iterate over code samples but only in the limited playspace that gpt4 offers.

    I'd rather setup a docker container and let it test over that container.

  7. @TheAiGrid

    April 2, 2024 at 8:26 pm

    Will OpenAI Be forced to release their findings on Q* Early? This new company is claiming they've done something similar and will release their product in a few months. This years going to very interesting for sure.

Leave a Reply

play_arrow skip_previous skip_next volume_down