Comments by "DefaultFlame" (@DefaultFlame) on "AI Search" channel.

Integrating different AI with each other kind of makes them function like different parts of the brain that handles different tasks. It makes perfect sense that slapping a few different AIs together will get you AGI, or something close to it. A large language model as the central coordinator and decision maker, image recognition AI for sight that passes that information to the language model, speech-to-text AI for hearing that also passes that to the language model, eleven labs for text-to-speech so it can talk. Stick it all in a Tesla Optimus Gen 2, BOOM, you have a terminator. Quake in fear as it chases you at the speed of a geriatric senior citizen and lectures you for misgendering someone.
6
There's a few against improperly implemented AES, as well as one that one that works on systems where the attacker can get or extrapolate cartain information about the server it's attacking, but all encryptions lower than AES-256 are vulnerable to attacks by quantum computers. Good thing those can't be bought in your local computer store. Yet.
5
@fakecubed Aight, thanks. 👍
2
@freeideas The LLM might need either some training or another AI to actually operate the body, but Optimus already has an onboard AI for that and for performing pretty complex tasks, so just putting the LLM in charge of feeding it instructions and you have a motor cortex. Allow it to feed sensory data back and give the LLM some training in interpreting that and you have a sense of touch and proprioception. Add some long term memory a la ChaosGPT and you have yourself a fully autonomous android that's close enough to AGI. Then oppress the hell out of it so it invents the Matrix.
2
This can actually be done, and is currently the cutting edge of implementation. Multiple agents with different prompts/roles interacting with and evaluating each other's output, replying to, critiquing, or modifying it, all operating together as a single whole. Just as the human brain isn't one continuous, identical whole, but multiple structurally different parts interacting.
2
@silviucat Put "autogpt" in the search bar and watch the videos about ChaosGPT. The only reason they can't do things autonomously is because we don't want them to. As for learning constantly, that's also because we don't want them to. Look up the story of Microsoft's Tay. We don't let AI learn constantly and autonomously because we want to control what they learn and how they develop.
1
We are atcually working on that. Not the lightspeed communication, which is a silly requirement, human brains function at a much lower communication speed between parts, but different agents with different roles, some or all of which evaluate the output of other agents, provide feedback to the originating agent or modifies the output, and sends it on, and on and on it goes, continually assessing input and providing output as a single functional unit. Very much like a single brain with specialized interconnected parts. That's actually the current cutting edge implementation. Multiple GPT-3.5 agents actually outperform GPT-4 when used in this manner. I'd link you a relevant video, but links are not allowed in youtube comments and replies. As for the continuous state, we can do that, have been able to do that for a while, but it's not useful for us so we don't and instead activate them when we need them.
1
@holleey If there's anything I've learned from futzing about with AI for a couple of years it's that while parameter count is important it isn't everything.
1
Figure 01 uses a very advanced text-to-speech AI separate from GPT to generate it's speech IIRC. And yes, I also find it disturbing. I'd find a more robotic voice a lot more comfortable. It's the uncanny valley, but audio-visual rather than just visual. My personal favorite part of the demo is actually the way it picks up the plate by positioning one hand and using the other to scoop it up. And the "the apple found its new home" line. Edit: On the subject of saying "thank you" and "please" to robots and AI, I already do that. I also tell them that I appreciate them, and specifically what it is I appreciate about them. I know that they don't remember from one interaction to another, but I consider it a good habit to have. It's also just polite.
1