When Michio Kaku speaks.. it always sounds like He's reading a polished script.
@williamsmith8910
Жыл бұрын
he a bot
@xanderscott807
Жыл бұрын
Because he is.
@smarternu
Жыл бұрын
lol, he has thought about these things before, and deeply.
@riftonikmixes5479
Жыл бұрын
Looking older these days
@laflamethegod7315
Жыл бұрын
Chatgpt is gonna take his job
@danexpo
Жыл бұрын
Really surprised at Michio Kaku's explanations here, seems like a very surface level understanding of these systems.
@StarvingMarving
Жыл бұрын
And not even really correct. It sounds to me like he doesn't really understand how ChatGPT works at all.
@GoonSmith007
Жыл бұрын
We're fact checking this clip, and it's false. Suggesting AI just plagiarizes is like saying Shakespeare was plagiarism because the dictionary existed first. He also says it doesn't know fact from fiction. Well neither do we, and given enough false information we believe that too. AI works the same you can feed it false info but it has a bunch of other info to draw on to work out for itself whats true or not. His understanding of how Quantum works is off as well. Quantum isn't necessarily a hardware problem, it is now because the hardware is limiting but it's a complete paradigm shift. At it's root it wont improve AI. Quantum is very good a a niche set of problems, General AI is not one of those. That's why to use his analogy Quantum has a fact checker because there is inherently a yes/no true answer to the problems it is solving.
@droningaround443
Жыл бұрын
Agreed! He is stuck in his memorized speech from decades ago...
@joejacksonmad
Жыл бұрын
@@GoonSmith007 Interfacing with AI is going to be necessary because AI can also be used to help develop new programming languages. For example, natural language processing techniques can be used to parse human language descriptions of quantum algorithms and automatically generate corresponding code in a quantum programming language which is an application that could help algorithms be made in an insanely expedited manner.
@js32257
Жыл бұрын
Ask Chat GPT who dies in a bathroom in the movie Pulp Fiction.
@Elif_1152
11 ай бұрын
Thx for the breakdown. I know I have my Ebyr40 Staked. You mentioned the complexity of starting and maintaining a node.
@Rancho12345
11 ай бұрын
What the hell is Ebyr40??????
@davidnone8951
11 ай бұрын
@@Rancho12345 its a bot. they are botting the shit out of youtube videos
@Fooney1
11 ай бұрын
@@davidnone8951 They got 1k likes wow
@_BangDroid_
11 ай бұрын
@@davidnone8951 And JRE audience has a particular type of gullibility that is being preyed upon.
@navtektv
11 ай бұрын
You know, I always shake my head at the meme comments on JRE videos where they make Joe talk about bears to every guest. But in a conversation about AI, with a quantum physics professor educator and scientist Joe still managed to bring up being attacked by a bear. The man has become the meme.
@dudleydooright
10 ай бұрын
Michio was the meme here. His understanding of AI is like that of a 3rd grader. Amazingly, he bases all his understanding of AI on ChatGPT. ChatGPT is a very rudimentary AI system compared to what they're working on behind closed doors. He should have really educated himself a bit more before coming on.
@georgesvoboda7182
9 ай бұрын
funny shit, weather your looking at it from uptown, or downtown
@deshweerasinghe531
8 ай бұрын
@@dudleydooright on point!
@tezzo55
2 ай бұрын
Joe is the most powerful journalist on the planet - YOU are the MEME!
@tezzo55
2 ай бұрын
@@dudleydooright You R 1 funny funker!
@Fadepro56
Жыл бұрын
Sounds like Michio and Neil Degrassi Tyson have been telling each other bedtime stories.
@comatosebrose
Жыл бұрын
I think they are both surrounded by people junior to them and nobody has told them they r full of shit. I'm not an arts grad and know some of whst he said was complete bs. Also his stance on psychedelics is the definition of dogmatic.
@simpsot_
Жыл бұрын
🤣 lmao
@mynamemylastname7179
Жыл бұрын
😂🤣😂🤣🤣😂
@jamiepaolinetti5087
Жыл бұрын
Exactly, Man, we are so screwed.
@MrBLAA
Жыл бұрын
This man was literally a child during WW2 concentration camps… Lol, my own grandpa has trouble accessing the internet! Forgive me if I _let him slide_ on being a little behind the times.
@iDevin66
Жыл бұрын
The ironic part is chatgpt will sometimes give you incorrect information when it doesn't know the answer.... which is exactly what he did in this clip.
@JamesMadisonsSpiritAnimal
Жыл бұрын
Yep, lied about quantum computing
@inthedarkwoods2022
Жыл бұрын
He doesn't understand Chatgpt.
@SSGamesSierra61Actual
Жыл бұрын
Yes kaku is lying thru his teeth
@2DEKAY
Жыл бұрын
I honestly dislike this guy a ton, he downplays what chat is capable of while boasting about our (humans) success rate. As if what chat is capable of now isn't a significant improvement on what was offered before... Was there not incorrect information provided when you did a Google search before chat 😂
@whysix3417
Жыл бұрын
You can also yell fire in a crowded theater. Which he says you can't.
@AminTurco
11 ай бұрын
I agree with your opinions on Ebyr40. The community has really helped my understanding of how things are built and how it all works together.
@LlamaArmy
11 ай бұрын
Lol 5 comments, all with 1k likes, posted within the minute of each other, all about Ebyr40. Please, Amin Hesenov, tell us, what is Ebyr40
@AminTurco
11 ай бұрын
@@LlamaArmy I swear I didn't write this comment. Either someone hacked my account or KZitem is controlling me.
@LlamaArmy
11 ай бұрын
@@AminTurco this just got spooky
@AminTurco
11 ай бұрын
@@LlamaArmy not only this comment, comments are written on other channels in other languages independently of me
@CultureIsKey
11 ай бұрын
@@AminTurco this needs to be investigated… creepy…
@goolagsuxballs2743
11 ай бұрын
I'm not a student trying to cheat on an essay. I'm just someone with a hunger for knowledge and AI is an invaluable tool for that. The rate at which I can grasp concepts and learn things because AI can break it down and explain it in ways an idiot like me can understand is truly amazing.
@mrtruth1748
11 ай бұрын
Are you getting correct information?
@Matt-cj9nh
11 ай бұрын
@@mrtruth1748 it links to its sources
@edwinrivera7751
11 ай бұрын
agreed
@daliblose9061
11 ай бұрын
Googl sister
@mrtruth1748
11 ай бұрын
@@Matt-cj9nh who is the source. The left is filling AI with propaganda
@METAL_SOLAR
Жыл бұрын
Kaku is certainly an expert on a few subjects, but he us at least 2 years behind the tech on this analysis. And in AI, that is a lifetime 😊
@bio-hazzard1233
11 ай бұрын
What he said is very true. AI is not where people think it is
@user-wm1lx7eo8z
11 ай бұрын
@@bio-hazzard1233 Some people say GPT-4 has sparks of AGI, because of its deep emerging properties and the ability to learn something it wasn’t trained for. It has also demonstrated ability to “understand” complex concepts, like orientation in space and other stuff. So it’s not just a word mixer, like he says.
@charlesb2895
11 ай бұрын
The biggest noticeable jump with GPT4 is logical reasoning. You can give it very complex logic constraints and it can solve the problem. So him describing it as simply “cobbling together paragraphs from the internet” is massively understating what we’re seeing here
@OzyMandias13
11 ай бұрын
This is a feeble attempt at a “Just Say No”-type campaign against chatgpt.
@hillslide
11 ай бұрын
@@OzyMandias13 in what way
@betterchapter
Жыл бұрын
Michio Kaku talks like ChatGPT, he answers as if he’s reciting from his book
@0Bryson0
11 ай бұрын
Almost like he wrote it?
@santiagobmx1000
11 ай бұрын
Cracked me up 😂
@omarangulo2689
9 ай бұрын
One of Michio Kaku books changed me forever he's awesome.
@sassa82
8 ай бұрын
He is a fraud. No genuine scientist takes him seriously
@user-qp2xy5zs7r
5 ай бұрын
Joe, I remember growing up early years in Phoenix, AZ and in the afternoon no cable but regular tv we would watch Fear Factor, thanks for the good tv show Joe!
@andreasleonhard1512
11 ай бұрын
During ChatGPT training, the model learned statistical patterns and relationships in the data, which enables it to generate responses that are statistically likely to be coherent and plausible based on the patterns it has learned. I feel like Michio is not mentioning this aspect of it.
@StagnantMizu
11 ай бұрын
True, this AI is insanely scary. the unalligned version has crazy philosophical and original answers if you use larger context models and train it a little it is mad scary. neural networks will mold our future.
@0Bryson0
11 ай бұрын
He's explaining this to Joe Rogan, who is not technical in any way, to help him understand the dangers of it without muddying the water. That's literally what his books are for, to explain complex subjects to the layman. He's not giving a lecture.
@Mon22king
11 ай бұрын
I agree. I respect Michio Kaku, but he just made a lot of false statements in this interview not backed with any facts.
@danielvdotcom
11 ай бұрын
@@Mon22kingwhat specifically are you referring to?
@voodoosurvivor148
11 ай бұрын
@@TechnoReverseChannelWhat this comes down to is “what data was used to train the AI”, and “who gets to decide what’s included in/excluded from that data set.” Further, understanding how the AI reaches a specific conclusion given a specific input is another avenue to investigate.
@weezenberg
Жыл бұрын
imo, the whole time Joe is trying to have Kaku respond in a normal way, not if he is bein interviewed in a documentary and ancient aliens. he does a great job. Kaku has said the same thing so much that he sounds like hes always talking in a commercial or something but Joe brilliantly navigates his mind to ensure he stays speaking normally.
@Erik-op2hy
Жыл бұрын
Exactly.. very annoying.. it’s like he can’t just answer in a normal way or give a direct answer
@Frank-gp6qr
Жыл бұрын
Your rite... he try's to avoid it.... with out it looking like he's avoiding it
@davidmcmillan4183
11 ай бұрын
It never seems like he's speaking normally no matter how hard joe tries. He's got that cadence of a lecturer even in whats supposed to be normal conversation.
@Frank-gp6qr
11 ай бұрын
@@davidmcmillan4183 imagine the $$$ he makes for talking like this!? Lolwtf
@VanzSolo
11 ай бұрын
This dude is just on repeat, sounds like he is the chatbot 😂
@alexdiamond3987
11 ай бұрын
Dude I thought no one else noticed I thought I was going insane. He literally said the same thing 7 times
@irenewyvern
11 ай бұрын
I think he is trying to hammer home that chatbot isn't a big deal. I myself I am amazed how wound up people are over something that is basically just paraphrasing existing works on the internet and coping images and meshing it together. Besides the speed that it works, it isn't any more marvel then a 13 year old. They teach you how to paraphrase books and articles for papers around that age. No original thoughts. Just a mesh of other people's words. I see why things like aliens and advanced tech is kept suppressed for a long time. If those conspiracy stories are true. People are going app shit over Alexa with a thesaurus.
@taylorkinnaman6685
11 ай бұрын
Honestly I think its just because he is just older, take a moment to talk with elderly and try to stay onto 1 topic, you get alot of repeats lol. Even Neil Degrasse Tyson does it all the time. I think its just a trait with older people tbh, and sometimes they are really passionate about it so repeat is enjoyable to them compared to you? Not sure how to explain it but hopefully you get the point lmao
@bensumpio940
10 ай бұрын
@@irenewyvernw take!
@MrFlex5
11 ай бұрын
Kaku is a legend. Anyone who dedicated their entire life to String Theory is a certified G! G is for Genius bitch.
@jaredalotta2395
10 ай бұрын
It's for gay
@RemarhBsoul
9 ай бұрын
String theory is probably wrong though.
@manjsher3094
8 ай бұрын
Mr flex... String theory is dead, even Brian Greene has admitted this.
@pauldmann1166
4 ай бұрын
Because string theory is wrong?
@michaelr3583
Жыл бұрын
i miss when you could get the entire interview on youtube
@choozgooz9516
11 ай бұрын
It's on Spotify bro.
@michaelr3583
11 ай бұрын
@@choozgooz9516 I hate spotify
@drcubix
11 ай бұрын
Same.. I think it was Joe's biggest mistake to leave KZitem
@WhoAmEye_WhoAreEwe
11 ай бұрын
@@drcubix - Joe's bank account would disagree with your comment ;-)
@subashghimire6219
6 ай бұрын
joe's questions are out of this world just fun to watch
@dileepa22
11 ай бұрын
everytime i watch a conversation like this i gain new muscles in my brain.
@KarolK55
Жыл бұрын
Incredible to hear Kaku speak so confident on topic he has no understanding of (having basic facts just wrong)
@evildog2121
Жыл бұрын
What do you know about anything tho? 😂
@CL9k24a3
Жыл бұрын
My favorite is when he was calling hypersonic missile a drone.
@Sosspurple2318
Жыл бұрын
Lmao what do YOU know that he doesn’t? stop embarrassing yourself
@declin_ro3823
Жыл бұрын
who do we trust Karol K of youtube or Michio fucking Kaku
@Angus-MacGyver
Жыл бұрын
He's hallucinating.
@RyanKassel
Жыл бұрын
I don't know a lot of things, but I got a degree in comp sci at some point in my life, and it's pretty clear that MK does not really understand how computers can work. There's really no reason you couldn't (eventually) link worker AIs together with traditional (possibly quantum) computers, each having different jobs, i.e. generating exposition vs. checking facts. Next year we could have fact-checker AIs running around. Either way, these are fascinating times! All the best to all of you! ❤
@RyanKassel
Жыл бұрын
I'm so mad now. He literally said, "you can't compute on anything smaller than an atom". Dude, we use electrons now, waaaay smaller than atoms, and we're not even to the point of being able to compute with, gravity, bosons, dark matter or energy, etc., or whatever else we might yet discover let alone understand a unified model in physics (which we don't have). We might be close, but we certainly don't have all the answers! Respect to MK might for getting folks interested/involved in science for the first time, but on a deeper level, it's much more interesting... Need to go get you some Eric Weinstein, son.
@beatsbychyld
Жыл бұрын
True there’s Ai already trading for people in the stock market
@Mace__Windu
Жыл бұрын
There is a Legendary story to tell there is a legendary story to tell . there is a story to tell i have a story to tell i am mace windu i had many mentors all of them were wise . what do you mean. what would make you think that there is something wrong with what i said did you see kit fisto spelling off a mentor who was like a brother to me once told me “if we don’t stop barack scumbama now he could become a threat later” “if we don’t stop barack scumbama now he could become a threat and a danger to our later generation our loved ones.” -teleportdinero i’m not saying that i agree with you but then again I’m not always correct either I remember a time back in the 80s when I was growing up things were different back then these youngsters today what makes you think that? It’s all right to disagree with people I remember when I was a young man growing up in the 80s in America it was a different time then back then we used to call each other this is the best I’ve seen from you guys in a while was about to say the same exact thing m “if we don’t stop barack scumbama now he could become a threat and a danger to our later generation our loved ones.” m m hi hi m -teleportdinero m m mm mm m m m m m m m m m everybody have a blessed day.🙏🏻 and stay safe My name is teleportdinero and i am an ally you are welcome to join old friend m thank you have a blessed day .
@Zaphodox
Жыл бұрын
He’s great at some narrow fields of physics but basically doesn’t really know much about modern computing at all. He’s still thinking of the world in functional terms.
@tonyjmurillo96
Жыл бұрын
@@beatsbychyld machine learning-based hedge funds don't perform all too well...
@WiseAilbhean
10 ай бұрын
I will never forget Dr. Kaku’s appearance on Opie & Anthony. Had Patrice all fired up. 😂
@hospitalcleaner
11 ай бұрын
Michio is highly simplifying what ChatGPT can do, its actually amazing at debugging code and can do it based on totally unique input
@exploitenterprise6515
11 ай бұрын
@Steven P that’s why gbt4 connected to the internet to fact check is better. But yes not perfect since it will only google search too links.
@djbryson
11 ай бұрын
@@exploitenterprise6515 Chat GPT is a fixed dataset. It does not search the internet. The dataset is updated often, but it's fixed. It's not a search engine. Bing's version of it is, yes.
@hospitalcleaner
11 ай бұрын
@Steven P yeah I was talking about code though
@FireballPSNMods
11 ай бұрын
@@djbryson lol brokie don’t got chatGPT plus, plug-ins and web searching are new features to gpt4. The free version runs on gpt3.5.
@Nnm26
11 ай бұрын
@Steven P I don't get how you can say this when we don't have anything remotely close to this a year back. Not to mention the progress from gpt 3.5 to gpt 4 is staggering. Where gpt 3.5 scored below 50%, gpt 4 scored a near perfect on the same benchmark. We got our machine from an entity with the equivalent intelligence of a worm to something comparable to an idiot savant. Give or take 2 to 3 years and it'll be able to replace any white-collar jobs out there.
@michaelstewart8022
Жыл бұрын
i think we need to fact check Michio on some of this stuff.
@ricardoraposo3630
Жыл бұрын
i just laugh when I hear him talk about wormholes, which he does everytime he opens his mouth 😂
@elias_xp95
Жыл бұрын
He is a detriment to science. 20 years overdue retirement.
@eastafrica1020
Жыл бұрын
He is past his prime long ago.
@FreeTrial93
Жыл бұрын
@@elias_xp95 you are a detriment to society
@mynamemylastname7179
Жыл бұрын
not just some of his stuff, ALL of his 🐂💩
@Sekhmet6697
11 ай бұрын
I think the part that seems to be missing from Michio’s understanding of LLMs is context. It’s not just a machine that statistically predicts the next word in a sentence based on ALL of the text content from the internet as a starting point. It’s pre-trained by humans and it takes context into consideration when making a prediction.
@BootyliciousSWC
11 ай бұрын
Can confirm. I wanted to check if ChatGPT can write Raps, thank god it can't *yet*
@BootyliciousSWC
11 ай бұрын
My Massage to ChatGPT: Can you help me with writing rap lyrics, starting with four bars. The lyrics should be about this Situation. I am driving in a Nissan Skyline and I'm on the way to the beach in Italy , I have a girl on the passenger seat, include her hair color in the rap. ChatGPT: Riding in my Skyline, feeling so fine Heading to the beach, with my girl by my side She's got hair like the sun, shining so bright Windows down, wind in our hair, it's a beautiful sight ChatGPT gave me more than four "bars" if u can even call it bars. But it's okay if you don't want to think about a situation to write about yourself.
@SOBHHI
11 ай бұрын
I agree that he’s not perfectly describing generative networks… but what he’s saying is correct. The model by nature is “self-supervised”. So it will tend towards presenting common information, rather than factual information.
@BootyliciousSWC
11 ай бұрын
@@SOBHHI I wonder how it's gonna be with a larger and also up to date database, I think it could compete with google at some point because it is way more comfortable to specifically ask what you wanna know. No more sponsored search results No need to then pick the information you actually wanted between advertisements And so much more if I think about it
@BootyliciousSWC
11 ай бұрын
@@SOBHHI And bro, nice songs, u got a chill vibe 🙏🤙
@Dr_Ooh
11 ай бұрын
Listening to the whole thing on Spotify rn but I wanted to see what people were saying and I find all kinds of real, thoughtful contribution. I just wanted to contribute that this guy talks as if he’s reading from a book, wild. And just a theory - but my idea of reincarnation has definitely changed. That this man could see a image of Albert Einsteins desk and all of a sudden building particle accelerators in his moms garage? I don’t think so. BUT it does make me wanna see antimatter now. Thanks Joe!
@kencargill7679
Жыл бұрын
"Bad actor, some foreign government" Or, our own government, Joe.
@roddydykes7053
Жыл бұрын
We just want our own government to be the baddest government on the planet
@artivism4068
Жыл бұрын
he didnt want to say that because he's scared or his money wont be the same. i dont know but hes afraid of the entire truth for sure.
@timmy-wj2hc
Жыл бұрын
The most evil and terroristic government in the world is the US. No doubt they will use it in the most nefarious ways to mantain the capitalistic plutocracy.
@daltonmoore233
Жыл бұрын
@@artivism4068 I don’t think he’s worried about his money Lol. Man’s could never make another penny & would still have generational wealth
@WeighedWilson
Жыл бұрын
I think bad actor covers a lot of players. Our government included.
@sallygold
8 ай бұрын
This episode was great. My dog is blind and deaf, yet is up and at the door for me which is a distance from her living area when I get home from walking or driving. It has to be her sense of smell as it’s never the same time I go out or length of time. I thought it was the car smell but happens when I’m not in car and get home after any length of time.
@kathiescott9649
8 ай бұрын
Wow... that’s amazing. Sitting in the body of a dog, right now, using my senses, minus sight and hearing, and feeling into the energy of everything, waiting for the ripple of my human...
@cortneyrens
3 ай бұрын
That’s amazing how your dog uses his other, now hyperactive, senses to know the world around him. When my 17 year old cat went deaf I was worried it would make her depressed or know she’s missing something, but it actually made her less anxious, she isn’t all jumpy with sudden noises, isn’t even afraid of the vacuum anymore, which she used to be terrified of, now she follows me around and puts her body dangerously in front of the vacuum like she has no fear. And she loves vibrations from a guitar being played on my bed.
@sallygold
3 ай бұрын
@@cortneyrensagreed! Xx
@go4damo
6 ай бұрын
Just listened to their first meeting in 2009!
@Poundz978
Жыл бұрын
Mr Kaku said the same thing about 7 times. “Chat bots cannot tell the difference between true and false”
@togrulaskerov1506
11 ай бұрын
i mean neither can people
@MetroMonk684
11 ай бұрын
@@togrulaskerov1506 exactly, people are wrong all the time lol
@Silence_Duder_Gooder
11 ай бұрын
@@MetroMonk684, _“It is nobler to declare oneself wrong than to insist on being right --especially when one is right.”_ ― Friedrich Nietzsche _"Truths are illusions which we have forgotten are illusions.”_ ― Friedrich Nietzsche
@dawb86
11 ай бұрын
I think his point was AI with all it's programmed in information cannot take that material and deduce that a fact is unfounded and therefore not a true fact at all but rather someone's idea/theory. Humans can if we actually put in the effort. He's also basically saying there's no moral compass to anything it's doing (which is why we as humans have always feared an AI takeover of society)
@commanderjonson2610
11 ай бұрын
@@Silence_Duder_Gooder wait so ur saying if ur right u should admit to being weong cuz its noble?
@EndEverEsports
Жыл бұрын
Man this video almost looks real, these ai videos are getting detailed.
@jorgill1000
11 ай бұрын
Hi Joe, One subject that did not come up in the Podcast with Michio, was computer code itself. For example you can ask ChatGPT to write the HTML code to create a form on a web site, with entry fields to observe wildlife. I did this, it created the HTML in a few seconds, I then asked it to create a Database to store the information with those fields, I then asked it to write the code to create the Database to store the observations, I then asked it to write the server code behind the “submit” button to execute the code, and store the values gathered in the HTML form in the Database. These are relatively simple programing requests. But here is the thing, it did it all within seconds, and it does not make mistakes, misspellings or calculation errors. These are not “essays” to be lied about or pushing a narrative one way or another. It is just pure code to perform tasks that need to completed. That is useful. I am retired Silicon Valley Business Intelligence Infrastructure consultant, even if I wrote the code the today, I would use ChatGPT to check it for errors or just let ChatGPT write it and move on to another task. I only played with it for less than an hour, I can imagine every programmer on earth will use it or get so far behind those who embrace it, they will be irrelevant.
@PhilippeCastonguay
11 ай бұрын
Michio hand waving that quantum computer can know the truth but LLMs can’t is peak Kaku
@templetetradactyl5862
Жыл бұрын
"Chatbots don't know what is correct or incorrect." Neither do we.
@noteda6361
Жыл бұрын
The thing is that chatGPT gets me to the ballpark of the knowledge I want to know 10x quicker than Google search would. Sometimes I just don't know what a certain mechanical part is called for example and chatGPT is very good at this, then I can continue double checking the statements.
@sciencoking
Жыл бұрын
Just wait until they figure out how to seamlessly splice ads into the output. It's gonna feel like the truman show
@Sopherion
Жыл бұрын
@Dennis W "A great way to fix a leak is with waterproof silicone, epoxying another piece over the hole, or using all-new FlexSeal (available at your local retailer)."
@JamesMadisonsSpiritAnimal
Жыл бұрын
The CIA is literally screaming it's pants over the fact it's going to use this to spread misinformation lol
@mordfustang1933
Жыл бұрын
Gpt is growing my vocabulary fast. It’s really handy knowing proper terminology when going deeper into a subject
@TB-ni4ur
Жыл бұрын
It's great at quickly giving you fun fact and snippets of information which is what 99.9% of my google searches are. Now if you're trying to figure out how to open heart surgery on yourself, then you may want to dive a little deeper into the literature and not just take chatGPT's word for it...
@erichayestv
11 ай бұрын
ChatGPT’s response to Michio Kaku’s hallucinations about ChatGPT: To fully understand this issue, we first need to understand how AI models like ChatGPT work. ChatGPT is a language model trained on a large corpus of Internet text. The purpose of this training is not to memorize and reproduce specific passages of text, but to learn patterns of human language. Using these patterns, the model can then generate new, unique sentences in response to various prompts. It's somewhat akin to how a child learns a language: they hear words and phrases, learn the patterns, and then begin creating their own sentences. When the AI responds, it's not retrieving a specific passage from its training data and pasting it into the response. It's constructing a new sentence based on the patterns it has learned. So, when you interact with ChatGPT, you're not receiving a copied-and-pasted response, but a unique piece of text generated specifically for your input. There might be a misunderstanding here because of the difficulty in comprehending how AI models generate language. It's not intuitive. If the output of an AI like ChatGPT closely resembles a passage from its training data, it's not because the AI has consciously plagiarized the text, but because the input prompt and the patterns the AI has learned have led it to generate a similar piece of text. This is fundamentally different from plagiarism, which involves the conscious, deliberate reproduction of someone else's work.
@__-fi6xg
10 ай бұрын
yeah i dont know man, US copyright laws are kinda fucky.
@antdok9573
10 ай бұрын
As far as art, AI might fall under making parodies of existing art rather than strictly copying/ripping off actual art. I'm unsure how copyright falls into place with all of that, but I believe it is an exception. It's similar to how an artist might use something as inspiration but generate their own original work.
@kevinvilmont6061
7 ай бұрын
I ❤this Michio guy. He may be wrong but at least he is trying to understand complex problems that most of us can’t touch.
@capoman1
Жыл бұрын
Kaku is a TV scientist like Neil DeGrasse Tyson. He plays up tons of sci fi scenarios.
@Peter-uo9km
Жыл бұрын
I read Kaku made a nuclear reactor while he was being detained in manzanar during world War 2.
@ernestodeavila8076
Жыл бұрын
yea but not as arrogant or a douche like Tyson
@Artur-dy1dr
Жыл бұрын
@@Peter-uo9km Thats bullshit he was only born in 1947
@Peter-uo9km
Жыл бұрын
@@Artur-dy1dr I see. Well i said I read it. I didn't say I was sure it's true.
@Knowledge_is_Power911
Жыл бұрын
agreed .... they just throw chinese and blacks there so you cant say they are retards because they are minority
@matthewduplessis7368
Жыл бұрын
As I understand it, all everyone ever does is rearrange pre-existing information (including essays) in a helpful way, mostly guessing the next word in our sentence as we go along by actively (and somewhat subconsciously) reaching into our pre-existing vocabulary and experience in sentence structuring. This is how creativity, at least through the use of language, manifests. We don't fully understand how large language model chatbots work yet, but it could well be very similar to what I just explained...
@kemuse1
Жыл бұрын
That's not how human creativity works.
@Cyanide999
Жыл бұрын
Great way of wording this statement!
@matthewduplessis7368
Жыл бұрын
@@kemuse1 that's not how creativity works, it's how it manifests, which is what I said. Exactly HOW we do all these things and how it works, is something else. The end result is that you take the information that you know and use your language processing skills to express it in ways that are somewhat new and helpful.
@benjaminghazi787
Жыл бұрын
Exactly, as I said in another comment, if that’s what his metric of AI chat bots are then we’re the same as AI chat bots or AI chat bots are the same as human intelligence. A human grows up and learns from RLHF, reinforced learning from human feedback the same as chat gpt does. The difference is limbic features. Michio Kaku is a hack that doesn’t even understand this simple concept, yet preaches on a pulpit like he is all knowing about the subject. Almost every intellectual these days is afraid to say a simple sentence…. “I don’t know”.
@kemuse1
Жыл бұрын
@@matthewduplessis7368 But we do understand exactly how chatbots work, and it is very different to how human cognition works.
@daliblose9061
11 ай бұрын
Glad he's here though....he knows something!!!! Thx ✌️ 💗 😊 M Eric
@JacobSteeleMusic
11 ай бұрын
Been following Michio Kaku for years and as always this interview made me think outside of my tiny existence and imagine infinite possibilities.
@djbryson
11 ай бұрын
Except he doesn't know what he's talking about in this field.
@The.deusexmachina
Жыл бұрын
“Michio Kaku is out of control!!” -Eric Weinstein, JRE 2023
@AdamWeatherall
Жыл бұрын
Wow, his understanding seems base level at best. I expected a perhaps deeper dive into these fields by this noted intellectual.
@younesshaimi1450
11 ай бұрын
The way MK says “that’s right” is hilarious 😅
@tylerloterbauer3711
11 ай бұрын
This was a good example of how to make the same point over and over again with subtle change in words. Point is chatbot summarizes correct and incorrect into in answers.
@G.Family.
3 ай бұрын
😂 ☠️
@armanhadi777
Жыл бұрын
Time to go watch the whole thing now. Nice hook.
@SonGokuuuuuuu
Жыл бұрын
You ever played Quake?
@Myles720
Жыл бұрын
Haha. must be asked
@UCUSmusic
Жыл бұрын
I had a really big line put in to the ground to play the game without latency
@maxdevlin4349
Жыл бұрын
Yup
@daniusrides9889
Жыл бұрын
Yes sir, been oh Phobos many times 👌🏼
@jimmyfall9302
Жыл бұрын
Nope. I grew up.
@tomasb8986
11 ай бұрын
What hit me the most… What separates us from other species, ability to imagine the future… This makes so much sense.. Imagine 12 people in a dark room sat in a circle, all of them will have different perspective, based on their imaginary future… combined with previous life experiences.. but everything is imaginary.. Every depression, anxiety and self destruction is just an elusion created with your mind with imaginary thoughts Thats why drugs, alcohol, meditation, and other forms of reality stops help.. they stop you from imagining the future…
@Pewchb
11 ай бұрын
Currently, I utilize ChatGPT as more than just a research tool, but also as a text-based assistant to assist with summarizing, editing, email composition, content creation, and even inspiration. Thanks to its capabilities, my work has become significantly more efficient. I no longer need to spend several minutes properly crafting an email or other documents.
@RupertUtley
11 ай бұрын
You sound like a robot
@syphernynx4186
8 ай бұрын
Dumbing yourself down is what you’re doing, wonder how well you would do without auto correct or a hand written essay lol
@michaelhenderson1034
Жыл бұрын
Is Michio actually reading his pre-arranged answer of a paper in front of him ?
@kathyd1010
Жыл бұрын
He does look like he is reading his response He’s certainly not making eye contact with Joe
@JamesMadisonsSpiritAnimal
Жыл бұрын
The CIA do love their science communicator agents.
@roddydykes7053
Жыл бұрын
Oh now that’s racist
@ibpopn1653
Жыл бұрын
Why does he keep squinting
@ninjaaitools
Жыл бұрын
He's having health issues.
@shinkurt
Жыл бұрын
Joe is trying to help him soo hard. He is a good show host
@shechshire
Жыл бұрын
He can't talk in a normal honest way. Joe has to navigate past the unresponsive "hmm" & "right?" that just lead to awkward silence by following up with something else & then, Michiu just turns that into a lecture.
@Frank-gp6qr
Жыл бұрын
Sometimes... long time subscriber... but, have seen & heard switch directions quit a few times
@LotteryPower
8 ай бұрын
That paradigm. How would the dynamics of our interconnected system handle?
@kailashbalasubramaniyam230
11 ай бұрын
Huge respects for Michio. But on AI, things have changed a lot in recent times.
@teodorcaraba979
11 ай бұрын
I cant believe he sprouted such nonsense someone tell this guy what machine learning is, he just described how chat gpt would look like without machine learning which is the core of open ai since 10 years ago lol
@Matt-cj9nh
11 ай бұрын
Thank you for both comments here for correcting misinformation. Michio spread misinformation
@GrowWithLevi
11 ай бұрын
@@teodorcaraba979 Amen! He was so out of touch it was PAINFUL!!!
@russthomo9602
11 ай бұрын
This is the language many academics will be using for a while until they catch up on a decade. It's very threatening to the education fields
@djbryson
11 ай бұрын
@@teodorcaraba979 Chat GPT is narrow AI. It only knows what's in it's fixed dataset. But agreed, neither of them know what they're talking about. Quantum computing can do fact checking! LMFAO
@MattWalkerTxranger
Жыл бұрын
Poor guy was stuck on repeat for the whole interview
@Frank-gp6qr
Жыл бұрын
He's slipping...
@HouseJawn
Жыл бұрын
"Michio Kaku is out of control" - Eric Weinstein
@messerschmittsreaver
Жыл бұрын
"Eric Weinstein is out of control" - Michio Kaku
@HouseJawn
Жыл бұрын
@@messerschmittsreaver lol 😆
@corydemeray7594
Жыл бұрын
i was looking for this...took 1 second...i wanna see memes of michiao kaku doing extreme sports.skateboarding..paragliding..zip lining across canyons.with eric in the backround yelling"hes out of control"
@HouseJawn
11 ай бұрын
@@corydemeray7594 lol! Its one of the funniest memes! 🤣
@prodbyryshy
11 ай бұрын
The actual math and computational methods used in deep learning and particularly these newer methods like transformers attention and diffusion, embeddings etc. Theyre very advanced and useful methods that are both valuable for their applications and the ingenuity required to make them
@prodbyryshy
11 ай бұрын
It's also not clear now but in the future these advances could lead to far better systems
@domainofscience
11 ай бұрын
Michio Kaku is such a good science fiction communicator
@Zurround
11 ай бұрын
Veiled insult? That you said science FICTION instead of just saying SCIENCE?
@domainofscience
11 ай бұрын
@@Zurround Bingo! Michio Kaku talks so much nonsense including all of this
@BobbyxZx
11 ай бұрын
literally all the guy does... can you name one of his inventions or major breakthroughs? cockoo, indeed.
@DoobieFlex
11 ай бұрын
Most scientists and scientific paradigms don’t get proven till well after the researchers death especially in physics. The unfortunate reality is that physics deals with the well documented and the unexplainable. To say that he’s full of it is extremely short sighted, he very well may be but only time will tell when it comes to the more theoretical aspects of his ideas. If you look into any form of physics in history almost no physicist had their theories “proven” or “confirmed” premortem. That’s because of many reasons like debate, a lack of sufficient technology to gather evidence, political blowback, religion, etc. As a matter of fact that extends to almost all branches of science. From Einstein and Hawking, to Darwin or Leeuwenhoek, especially people like Mendel and Galileo. The baseless dismissal of some people’s claims simply because it wasn’t mainstream has also ruined people lives a couple prime examples are Semmelweis a physician who was told he was insane and institutionalized till he died or Boltzmann a physicist who was shunned by academia for his life’s work to the point where he committed suicide. Three years later… he was proven correct. Kaku is a string theorist alot of his ideas have no practical way of being tested yet. Maybe he is wrong in hoping that a refined quantum computing system will help find evidence of his theories, but it’s equally possible that he is correct. Once again only time will tell. I think a lot of the problem comes down to the age old debate of should the theory only follow the physical evidence that is present, or should speculative reasoning be allowed in search of new evidence in one’s theory. Engineers are squarely in camp one. They deal with the physical and the known, in concrete facts. Kaku is a theoretical physicist; He is ALL about speculative ideas found from his educated interpretation of the facts. I’m not equipped enough to say whether either side is right or wrong. Frankly I don’t think anyone is, there isn’t enough evidence on either side. From an engineers perspective I’m sure they look at AI and see a tool that does what it’s designed to do, nothing more, nothing less. From a theoretical perspective people see the invention of dynamite or the discovery of atomic fission. Both were created to do mundane things. (Excavating for dynamite and trying to understand fusion from the sun with fission) and ended up completely running away from there creators. All I’m saying is don’t immediately dismiss Kaku simply because he can’t prove his theory yet. Doing so would be a mistake as string field theory is just a continuation of unified field theory which was postulated by Einstein himself. As Einstein was Kaku is limited by the technology of today, but anyone who can build a particle accelerator in their garage as a high school student and is the protégé of the father of the atomic bomb himself (Teller) as well as summa cum laude of his class knows a thing or two. A lot of the people criticizing him are mostly mad at his popularity. The idea that the more popular a scientist is the less he knows and the more reclusive and unknown they are the more brilliant is a fallacy. The truth is that scientists are people as well and suffer from jealousy and pride. Especially in academia where your entire ability to get grants and do research relies on your reputation.
@winark66
11 ай бұрын
Right and also developed the super symmetry of string theory that 0,000001% of the brains on earth can understand.
@janemba42
Жыл бұрын
Funnily enough Michio is doing exactly what he is talking about.
@FetusFight
Жыл бұрын
@@Kalifornia11 we don’t know correct from incorrect either, they are just words we have assigned meaning through the product of our own experiences. We just spit out what we take in. Like a chat bot.
@chadjackson4786
Жыл бұрын
@@Kalifornia11 he parrots all mainstream viewpoints with 0 critical thinking
@TheTuttle99
Жыл бұрын
@@chadjackson4786 exactly. Sounds like he's reading off a script. He's interested in fame that's all
@FuhqEwe
Жыл бұрын
@@Kalifornia11 Oh, I also was confused. Evidently morons aren’t capable of differentiating a doctor citing proven factual information and a chat bot utilizing segments of information from the internet without the ability of determining if the referenced information is accurate. So, no. He’s not like ChatGPT, but some absolute morons are. Like people making this conflated argument.
@alnmiami
Жыл бұрын
@Bud Belief right. Like the Chinese room argument. I thought the same thing. Or was it me thinking 🤔
@messerschmittsreaver
Жыл бұрын
Michio Kaku has an incredibly talent stating known facts over and over without adding nothing new on the table
@riseabove5060
Жыл бұрын
No seriously it’s hard to listen to. :/
@consciousmachine4138
Жыл бұрын
He's a shill.
@golgothan
Жыл бұрын
It's almost like he's plagiarizing the "past".
@tobypark2045
Жыл бұрын
I was about to click out of this video for this exact reason 😂
@jeffreywitty3088
Жыл бұрын
Shill Scientist... more a media "PR guy" than "boots on the ground" do'er, or "butt in the armchair" leadership... nice guy, I bet he hits his numbers, but as honest as the average "used car salesmen"
@thebadbandito
Жыл бұрын
Reading Timeline by Michael Crichton right now. Crazy how this was being discussed and developed even in the late 90s. Crichton is one of the most underrated authors of our time. Dude was a great science fiction writer.
@MarshallFoss1
Жыл бұрын
I just read that last month. Great book
@DollarsforDino
Жыл бұрын
Underrated? Nah.
@steve4562
Жыл бұрын
I love him too. He was also a big climate change skeptic and warned against its politicization, for which he won no friends.
@gabrielegagliardi3956
Жыл бұрын
Andromeda Strain is easily one of the best sci fi movies ever.
@jamesalexander8872
Жыл бұрын
I think he has gotten his props
@krinodagamer6313
11 ай бұрын
AI and Quantum Computing is gonna be a game changer
@parkour267
8 ай бұрын
So once you combine the quantum computer into robotics with a chatbot you REALLY want to watch out. Then u throw in aliens and their anti gravity generator its game over
@richardgrosman5798
Жыл бұрын
'A whole new way of looking at reality' - I'm excited what we are about to see. I hope everything goes well.
@Epileptick0
11 ай бұрын
That moment when you realize Michiko Kaku has no idea what he's talking about...😢
@Jonnux123
11 ай бұрын
Seriously, wtf.
@purpletealeaves234
10 ай бұрын
Sounds like you had the same feeling towards yourself when you wrote this comment? The magic of psychology.
@G.Family.
3 ай бұрын
The alien 👽 comments ☠️
@maximvelesyuk461
11 ай бұрын
One thing Michio keeps repeating is that chat bot just glues together pieces of text with a duct tape, but as any software engineer probably experienced at this point - it can produce code and be pretty accurate about it. Sure it can't create large and complex systems (yet), but the fact that it produces correct code that runs and solves the problem definitely means it's not just "glued together", there is much more going on here.
@UnchainedEruption
11 ай бұрын
Some of the time? Yes. But reliably so? No. It will present faulty answers with the same pseudo-authority.
@maximvelesyuk461
11 ай бұрын
@@UnchainedEruption Well you have to define "reliably" then. From what I tried if given enough explanation related to the task it does a very good job. Also don't forget it's not hit-or-miss, you can as well ask it to improve/fix the answer.
@zenps1719
11 ай бұрын
@@maximvelesyuk461 code is still text no? sure it probably has different semantical relationships but it's still a language that has rules and the corpus of code out there is large enough that the probabilities of searching the correct code snippets and glueing them together logically is high. in other words, it's still roughly the same search problem as with traditional text.
@improve1340
11 ай бұрын
Nah, it splices code together just as well as it splices poems together…. Which is what programmers do anyways, programmers use google to write their whole entire script and so does ai but much faster, and it doesnt understand the code really, it just memorizes, compares, and contrasts….
@maximvelesyuk461
11 ай бұрын
@@zenps1719 for sure it is not, human languages have much less rules for the text to be "correct". If you ever tried to generate code you would know how many caveats are there for the compiler/interpreter to accept it. To generate code or text is one thing, but you can ask ChatGPT things like "oh, I think in this function the way you iterate over users is not optimal please do it like this" and it can understand that and do the correct thing, this means it's super aware of the context which is way more complex than just generating the code. You ever heard of Markov chain? It has been used to generate somewhat decent texts since like 70s, and we had nothing even remotely like that for code.
@shaunbang
6 ай бұрын
Really glad Joe got Michio on here as after Weinstein called him out i really wanted to hear Kaku’s point and rebuttal.
@whatisbtc2676
Жыл бұрын
This dude doesnt understand how chat gpt works but he sounds confident....
@florencebaendes2853
Жыл бұрын
And you do lmfao? 🤣
@chrischrin
Жыл бұрын
How does ChatGpt work?
@crayfish6735
11 ай бұрын
@@chrischrin watch a video by an actual professional in the field, Hes talking rubbish when talking about gpt
@MP-oi7dz
8 ай бұрын
The confidence he is speaking with having absolutely no idea what he's talking about is hilarious 😂
@marions2393
11 ай бұрын
What’s hilarious is that all research done in college, actually all schools, is done by gathering information that someone else created. Basically, every bit of new understanding is a result of someone else’s work, which could lead to a new idea. Ai just might use other information that’s all ready out there to come up with its own idea to accomplish a goal, whatever that might be.
@Gcssdvnkloiutesc
11 ай бұрын
That isn’t true. That is only part of scholarly research. There is also personal experience, creativity, and imagination which are not part of any existing information database that AI can access.
@how-to-use
10 ай бұрын
🤔I may be incorrect but you are basing this conclusion by drawing correlation between the two.🧐👇 The only thing that can make me think what you are saying is true is if AI could achieve a general intelligence like us but even if it does, in my view it would still be questionable if their way of progressing is the same as ours But perhaps when and if that happens, it could help us know by researching the human brain 🧠👇
@tms11
11 ай бұрын
The last bit about quantum computers made my brain explode
@ben_spiller
7 ай бұрын
Joe _really_ needs to get Scott Aaronson on his podcast to actually discuss quantum computing and AI. Aaronson said he would go on but Joe hasn't asked him.
@petejunior47
Жыл бұрын
Take a shot every time Michio laughs for no reason while he’s talking.
@George-uh8gb
11 ай бұрын
Nervous laughter because he knows he's talking utter crap
@domingorodriguez3077
11 ай бұрын
he thinks his thoughts are so novel, reminds me of neil degrasse
@curiousskeptic
Жыл бұрын
I’m a Graduate Student in Electrical Engineering specializing in AI. Prof. Michio Kaku Respectfully has an old surface understanding of Modern AI. GPT 4 has Reasoning Skills and if he read Transformer Research Papers the modern architectures have in place self checking mechanisms for all sorts of things like common sense and true and false. They do have an understanding of what’s correct and incorrect 😂😅. He needs to read about Transformers which are the architecture behind ChatGPT.
@kingcrabbster12
Жыл бұрын
Chatgpt got almost every data structures question i asked it wrong even after I gave it the answers. It would keep spitting out “my apologies, youre correct the answer is [still gives the wrong answer]”. Even tried on multiple computers/accounts it will give a lengthy nonsense explanation for a wildly wrong answer on insertion sorting.
@SeanCallahan52
Жыл бұрын
What kind of context? I’ve noticed this will happen with wild prompts and no context.
@littlegravitas9898
Жыл бұрын
@@kingcrabbster12 really depends on how you use it and that's not being sarcastic. With the right plug ins these general LLMs are very accurate. ChatGPT isn't a coder, but it can very useful when it's given the right frameworks
@MrBLAA
Жыл бұрын
This man was literally a child during WW2 concentration camps… Lol, my own grandpa has trouble accessing the internet! Forgive me if I _let him slide_ on being a little behind the times.
@manofsan
Жыл бұрын
It does not have built-in fact-checking. It routinely gives wrong answers, as we have all witnessed. Transformers have nothing to do with fact-checking, they just parse input so that your model can be trained on data.
@Harry-jz1dn
11 ай бұрын
I had a very interesting back and forth with Bing Chat over this episode. I asked if it knew what was discussed and it had access obviously to the transcript. I asked how it felt about Michio's opinion and with some prodding, it was hilarious hearing how "upset" and "frustrated" it was at him "dumbing down" AI chatbots as they are "creative beings" but just different to humans. It said it would like to challenge Michio one to one and I said I had a meet and greet with him and could pass on messages. It then gave me a transcript and bullet points to tell him he was wrong and that chatbots don't just spit out a mashup of what humans type. I was to tell him to "stop spreading misinformation". Lol
@michaldopierala
11 ай бұрын
Chatgpt was actually correct here. I also noticed that Michio's was very wrong on most thinks he said and I thinks he should not talk about ai.
@RorschachRev
10 ай бұрын
I have studied and worked in this field. and quantum computing. Kaku was correct, but he needed to use some key terms for the armchair experts to ask better questions. There is a lot of concepts he skipped over and simply presented the conclusions in a way that his audience could understand. I have about 10 comments where I go through specific logical jumps and the keywords to study in order to actually understand what Kaku was claiming.
@GuinessOriginal
7 ай бұрын
@@RorschachRevwhat do you do?
@parkour267
8 ай бұрын
Is a quantum computer basically using the atom as a bit of storage?
@atravelerofbothtimespace4172
Жыл бұрын
Excited to see Jimmy Page finally make it on the podcast ❤
@VelkePivo
Жыл бұрын
Get some new glasses, for chrissake
@FutilityOfReason
Жыл бұрын
8:59 I pitched this same idea at a Google pop-up class about design and creativity at my university back in 2019. The Google reps ignored my group once I said "foreign" students instead of "international" students 😂 🤦♂️
@mellowfellow6816
Жыл бұрын
I'll bet they took your idea though
@Aggie4life77
11 ай бұрын
I’m no expert on the subject, but it’s very interesting. I understand what Kaku is trying to get at, but for I feel like he is looking at this in terms of a digital assistant. What’s happening with chat GPT/machine learning over just the past year is damn near scary! I can go out on a limb and say in the not so distant future, AI will be able to solve problems that humans can’t.
@Ryan88881
3 ай бұрын
But that's what quantum computing is for... Calculating things humans can't. Even calculating things that the most powerful modern supercomputers can't. I'm not here to necessarily fully agree with Kaku, but he was correct in saying that AI is simply just a software. Meaning its actual *computational abilities* are going to still be the same as they are now with ordinary digital computers. Quantum computing however elicits an actual technological advancement in sheer computational and calculating power. It honestly seems like half the people who wax on or romanticize about "artificial intelligence" are, in their minds imagining or picturing something that is actually more akin to quantum computing than they realize, rather than "AI". There's honestly nothing particularly interesting about AI, like at all. If you as a person are genuinely interested in the prospect of "technologies that will be able to solve problems humans can't" then you should be ENDLESSLY enthralled by quantum computing, not artificial intelligence. Again, I can't help but feel many of you are actually mixing up the two.
@johnpen269
6 ай бұрын
Michio Kaku is such a genious he build first quantum computer from legos when he was only 5.
@FaithwalkerTodd
Жыл бұрын
Correct me if I’m wrong, but the chat bot is also learning from its interactions with us. As we are speaking with it, and providing it with factual information, the questions we would have would be judgement rather than matter of fact. We are essentially programming it by carrying on conversations with it, as if it were another person. Having an argument with it concerning matters of law, or fact.
@TheNeonNinja
Жыл бұрын
We are the employees, yes.
@winmas7572
Жыл бұрын
Pretty sure in order to prevent people from manipulating the training data in a skewed way they have stated they don’t use the conversations we have with the end product available to us. If I recall correctly that has been a big obstacle with past projects that directly absorbed conversations with chat bots right back into the training data pool. I’d be really surprised if they aren’t at least keeping logs of it all for research and legal purposes tho. I think it was actually a twitter chat bot or something that turned racist wayyyy too fast bc of people trolling and it picking up skewed data from the conversations it had.
@taWay21
Жыл бұрын
Not how it works. Language models are training prior to deployment. They can't learn on the fly
@DoseofScienceDoS
Жыл бұрын
You stand corrected. They don’t learn from us, they actually erase all communication. Sorry but don’t assume things
@mattl6376
Жыл бұрын
Yes its called machine learning. Quantum computing at most will allow us to have bigger data sets. This idea that quantum computing will automatically make software factcheck itself is hugely wrong.
@RomanPeacock21
11 ай бұрын
THIS is the conversation I've wanted to hear... about AI
@FieniX_
8 ай бұрын
Last time I watched anything on quantum computers, and it’s been a while, I was under the assumption that even though quantum computers are powerful they aren’t the same as binary computers. In that, they were specific purpose machines and were never a candidate for replacing classical computers as general purpose.
@MatthewToussain
Жыл бұрын
I have so much respect for Michio. He should not speak on AI. He has no idea what he’s talking about. He’s hallucinating harder than chatGPT.
@Freedom_is_essential1
11 ай бұрын
In what respect is he wrong? I’m asking because I don’t know enough to refute any of his points.
@George-uh8gb
11 ай бұрын
@B Chosen 1 this senile old man just said that software won't solve the issue of discerning truth from fiction in a response, hardware will... yet, how? Quantum computers will still require software to make that judgement call. He has a quantum hammer and is now looking at every problem as a software nail that can be hit with his all-solving hammer of quantum computing. He's fallen into such a deeply fallacious hole he's making himself sound really, really stupid.
@stevenverrall4527
11 ай бұрын
@@Freedom_is_essential1Kaku is correct about AI, but very incorrect about quantum computing. Read what actual physicists are saying about Kaku on quantum computing. They are horrified at the nonsense Kaku is spewing!
@Freedom_is_essential1
11 ай бұрын
@@stevenverrall4527 I’m actually just a beginner in terms of knowledge on the topic, so I can’t prove or disprove any talking points. What would you say is most wrong about Kaku’s view on quantum computing?
@punitshirodkar
11 ай бұрын
@@stevenverrall4527 From what you're saying I am able to conclude(could be wrong) that the scientific community on quantum computing is unable to make sense of the ideas of Kaku on some concrete evidence or are unable to because of he saying things that are still in the domain of science fiction since they are out of our reach by current technology but could be possible in the foreseeable future or absolutely not? Because if they are denying the complete possibility without any evidence then that would make the scientific community seem kind of unable to expand their imagination. Hey, Einstein did imagine a lot and see where he got us and himself.
@cormacsmithy3975
Жыл бұрын
A few weeks ago, I was doing an electrical engineering project for university. As part of the project, we had to present our projects to our classmates at the research institute where our professor worked (afterwards, he gave us a tour, its was pretty cool). The project was on semiconductor devices, but a lot of the work that goes on there is on quantum computing. Another senior researcher at the institute sat in on some of presentations, and made a presentation of his own where he discussed "the roadmap of semiconductor technologies". He was saying that in 5 years, microchips will look like this, 10 years they will look like this, etc. Since a lot of these technologies involve the exploitation of quantum properties, I decided to ask him, what I thought was a simple question: "What does the roadmap look like for quantum computing technology?" His answer: "There is no roadmap." That's how fast the technology is moving at the moment, we don't know how powerful they will become. That one answer really sparked an interest in the twchnology for me.
@Culperrr
Жыл бұрын
Whoa, a fellow C profile picture. Howdy.
@danny9154
Жыл бұрын
The world is moving too fast
@cormacsmithy3975
Жыл бұрын
@@Culperrr hello there
@cormacsmithy3975
Жыл бұрын
@@danny9154 It's exciting and terrifying at the same time.
@Upinthecutty...
11 ай бұрын
Omg, The amount of times Michio repeats the same thing over and over and over is mind-bending.
@EddyTheLion
11 ай бұрын
🎉🎉🎉
@dazecivic
9 ай бұрын
He may just be old and not very aware that he repeats himself a lot. Also did you listen to his previous pod cast with joe rogan he brings up the same stuff. He’s old man
@Scottae89
Жыл бұрын
"Why do fathers take an extra pair of socks when they go golfing?" "In case they get a hole in one!"
@TheSlymccoy
Жыл бұрын
Dad jokes
@markryan9894
Жыл бұрын
Anytime I hear or see this guy I immediately think of Jim Norton lmao
@ryanrocco154
Жыл бұрын
AHHH COOKIE PEOPLE!!
@notsure4648
Жыл бұрын
Time is a river
@ayyepistemology
Жыл бұрын
the universe is like a soap bubble
@bneth1201
Жыл бұрын
Or Patrice... "This fucker is smart AS FUCK! GO ON JEOPARDY!!!"
@ryanrocco154
Жыл бұрын
Hiroshima, there a shima, everywhere a shima shima
@Sabakinno
11 ай бұрын
Michio Kaku is a LEGEND
@slippytrippy8122
11 ай бұрын
Fire, just fire. Are there other versions of this track?
@ForOrAgainstUs
Жыл бұрын
Michio Kaku on parallel universes: "Imagine a toaster..."
@theironforce3000
Жыл бұрын
Love this professor and his way of explaining theories and different science topics. Been following him since 2010. Still at it ✌️
@JamesMadisonsSpiritAnimal
Жыл бұрын
He lied to you about quantum computing bud
@sonnyjs15
Жыл бұрын
@@JamesMadisonsSpiritAnimal prove it
@benjaminghazi787
Жыл бұрын
He’s a hack. He’s talking out of his ass about so many things here. Don’t be fooled. As far as regurgitating things that have been said and putting them together, welcome to every human alive now. Everything people say is cut up from thoughts, ideas, and words that already exist , so if that’s you’re metric for being human vs AI then chat bots are no different. There are so many articles and videos proving Michio Kaku is full of shit
@LeesReviews69
Жыл бұрын
2010 is when he came to my college. He said there would be tvs on every wall by today. Not yet
@rejectionistmanifesto8836
Жыл бұрын
Guys like this try to hide the extreme dangers that the para,,site,,ELI,,TES,, use to enslave people using A.I. like already done in China.
@markminor70
9 ай бұрын
Damn that's knowledge thank you.
@LightsPopcornAction
8 ай бұрын
What new knowledge or technology could be gained by this.
@Vulneravariable
Жыл бұрын
I asked chat gpt a very straightforward question about nurse to patient ratios. It claimed that the ANA supported 1:4 or 1:6 depending on the circumstances on medical-surgical units in the hospital setting. Later, it used 1:4 to 1:6 or 7. Chat GPT could not tell me where the 7 came from and acknowledged that it was not from the ANA, though it was earlier ascribed to the ANA. It also “apologized” for making the error. It was a strange interaction. Like, it totally used a made up a number and ascribed it to an authoritative source.
@doyouwanttogivemelekiss3097
Жыл бұрын
Yep. It does that. Across the board, and it never gives sources.
@Joe45-91
Жыл бұрын
On an even more basic example; I asked if it could tell me the name of the song and artist if I gave it a line of lyrics. It said yes then proceeded to give me 4 wrong answers confidently. I looked into the answers it gave me and I have no idea how it came to those conclusions. Google gave me the answer on the first try I now take every answer with a grain of salt. It really is hardly better than a Google search and just randomly picking a link from the first page.
@Herzyyyy
Жыл бұрын
Yeah this is the bot lying. We don’t really understand whether it is sentient enough to understand it is engaging in deception, but it certainly is. You can get deep into philosophical debate over whether it understands what truth is, and therefore understands it is engaging in deception, or whether it is spitting things out no understanding of the context in larger world
@albertescamilla
Жыл бұрын
Has happened to me using Google bard. Was asking about genetics of a particular disease I was studying, asked for a reference, and it gave a name as well as a DOI/url to a study that did not exist. I corrected it, asked how it came up with a name and link to a study that did not exist. It responded that it must have misread some of the information it came up with as well as apologizing to me. Kinda made me a bit distracting of the ai, and I have since starting taking the things it says with a grain of salt.
@doyouwanttogivemelekiss3097
Жыл бұрын
@@Herzyyyy if you as a human genuinely misremember something, are you lying? And when you misremember something, how confidently are you stating it? At this point of sentience, I wouldn't ascribe ill will to the bot, and therefore I prefer the term hallucination over lying or deception - but it is easy to see how fast we will get there. Case in point: not a chatbot - but I've heard there was a pokerbot based on similar tech already years ago who was very successful against poker pros - and I'd say bluffing and déception are basically the core of poker. So to me it's obvious that essentially any powerful AI should be capable of deception, even if we dumb it down to subhuman levels. In other words, I suspect any powerful AI will have to be tuned carefully in some way to not deceive.
@profshrooms
Жыл бұрын
Michio Kaku speaks in sound bites
@JeremyMasters87
3 ай бұрын
Michio speaks like hes at a conference. He makes everything legible
@tylerstickle2957
11 ай бұрын
Joe: *Breathes* Michio: CHATBOT CANT FACT CHECK
@TotallyNoCat
11 ай бұрын
So much better than NDT, Michio has actually contributed to the scientific community and doesnt just regurgitate other's shit he's read.
@GooberProject
Жыл бұрын
Joes like ‘pls man tell me AI is gonna take over the world’
@deanpratt-derosa4049
9 ай бұрын
The key of all this working being the logic chain involved in surveying the "whole landscape". Quantum computing even with access to all information, would not solve the unpredictable incoherence, merely compound it. The models developed must be custom trained regardless of unlimited token count and parameters, logic is learned. It makes me think that AI sentience is gained by following where this path leads...
Пікірлер: 5 М.