@@ArtOfTheProblem I didn't watch it for sure, but I can't remember if I skipped the notification and forgot to come back. Some days you're in a different mood 😅
The video actually delves into all the relevant terms we hear about in Reinforcement learning but have to go through an hour it teo hour long lecture to understand. I will keep coming back to this video again and again while learning the topic ❤😊
Very informative, I loved the ending note of using successful architectures in completely different domains. Especially interesting with the recent release of OpenAI’s o1
@@ArtOfTheProblem very excited to see it! As someone who loves to learn about AI but has no formal education on it (at least yet) these videos are my lifeblood and only make me more energetic to learn more
I posted this video a month ago, but due to incorrect settings, it wasn't shared. I'm reposting it now for another chance to fly! I'd greatly appreciate if you could like, comment, or share it this week. FULL AI series: ru-vid.com/group/PLbg3ZX2pWlgKV8K6bFJr5dhM7oOClExUJ
"Robots learned to act through reinforcement learning by trial and error, mimicking how animals learn. Over time, algorithms evolved, allowing robots to make better decisions and adapt to complex environments."
Really cool video, thanks. I don't understand most of it, but the implications are clear, robots with general intelligence. That's huge. I haven't been jazzed about the language models, they are being used to take away creative work from people. But intelligent robots could take away the shitty work, leaving people to do creative things. I suspect that our inherent human fuckery will prevent it from being a utopia. We aren't built for endless pleasure so we will produce strife if the situation doesn't present us with any. I'm thinking of Anna Lembke with all that, how our brain tries to maintain an even mix of pain and pleasure as part of homeostasis. Perhaps that emotional need can be expressed completely in the digital space so that the war instinct stops making fuckery in real life. The momentum of the nation-state global structure we've had for a while will be hard to overcome, it directly incentivizes conflict.
Amazing its as if all these videos are set up like a documentary of whats happening so in future yeara its like the most accurately recorded development in history ❤❤❤❤❤❤❤
Thanks, this was very helpful and informative . If your middle age is like me, then you grew up around lots of early leftover vacuum transitor antique tech that was very sensitive eltrodynamical systems if shorted or flawed you've played ham radios or bunny ears and distortion in amps, or even like a losly screwed in light bulb on a shaky table making it's plasticity stobe like gates or switches opening and closing . Something I think younger generations are lacking but may have more understanding on this level about the algorithm, code, or step by step functions after the facts screen time. I retired in precision machining where early g code cnc was very widespread before the internet reply took off but it wasn't much to be learned on its history.
Thanks for alerting me to this video. I have now subscribed so I won't miss any future videos. There was so much food for thought presented, and it filled many blanks in my knowledge.
@@ArtOfTheProblem It came up as a suggestion on my RU-vid home page. What caught my attention was the phrase "learned to act", and you explained that concisely. Best of luck and keep up the great work.
The easiest example I can think of It’s slightly dystopian but the cart and pole kind of reminds me of twitch streamers trying whatever to get engagement So it makes me wonder if you can RL engagement and may be ratings
@@ArtOfTheProblem But without the human in between just directly optimizing for engagement generating content I’m quite frightened of that and intrigued too much power
Loving the video but the music is giving me a headache. It's like trying to think near a railroad crossing signal that just won't stop ringing. I don't mind music in videos but I'm a firm believer that it should only be there when I'm not trying to listen to anyone talking.
Super interesting stuff! But maybe don't encourage people to apply their expertise devising ways for wealthy capital owners to more effectively extract wealth from the working class? Machine learning has a tremendous list of potential beneficial applications, so using it to solidify existing wealth and power hierarchies seems counter-productive to me.
@@uvectoru1327 this is a great question , I’m actually working on a series on economics right now - it will look at all sides of that question , such as efficient market hypothesis , access to markets , and value itself (power and risk)
@@ArtOfTheProblem I'll definitely check it out! I'd be really interested if there are any applications being worked upon in the leftist economic sphere. (e.g. cybernetic algorithms applied to resource distribution like Project Cybersyn from the 70's)