ChatGPT Explained Completely.
Video Overview & Insights
For 16 free meals with HelloFresh PLUS free shipping, use code KYLEHILL16 at https://bit.ly/41QHRYf
Thanks for watching! This is a deeper dive than usual -- hope it's useful. And let me know what you think of the new [FACILITY] rooms! The Kevins worked for months on them. Can you spot all the easter eggs?
ChatGPT is now the fastest-growing consumer app in human history. Problem is, almost no one knows how it actually works. This is everything you need to know.
💪 JOIN [THE FACILITY] for members-only live streams, behind-the-scenes posts, and the official Discord: https://www.youtube.com/channel/UCFbtcTaMFnOAP0pFO1L8hVw/join
Thor😊
OR https://www.patreon.com/kylehill
👕 NEW MERCH DROP OUT NOW! https://shop.kylehill.net
My experience with ChatGPT has been somewhat overwhelming. I’m not using it to write university papers for me etc. But I’ve asked it questions with respect to inspiration, procrastination, life, etc. The responses have been so helpful to me. I am stunned by it all🤷🏻♂️. I watched your video but still don’t understand how ChatGPT can provide this connection with me with such meaningful detailed responses. PS I have produced over 500 television programs that have aired nationally in Canada. I have served as a CEO. I am an engineering technologist (old school). I’m in awe of all of this technology. But even referencing AI as technology does not seem to correctly describe what it is. I still do not understand how this works. I get immediate detailed responses that flow down my page perfectly and ring true to me. Again 🤷🏻♂️. Txs for the video!
🎥 SUB TO THE GAMING CHANNEL: https://www.youtube.com/channel/UCfTNPE8mXGBZPC1nfVtOJTw
✅ MANDATORY LIKE, SUBSCRIBE, AND TURN ON NOTIFICATIONS
I know this is an old video, but Kyle to answer your question about cats and plastic, I found the answer. Some plastic is made with tallow. So when a cat smells it, it smells like animal fat, not plastic. A lot of cat behaviors to do with sucking on things have to do with instincts leftover from kitten hood. A cat with an oral fixation of that kind could have anxiety. And then it smells animal fat on an object that it knows if it sucks on it will remind it of drinking milk. Association set, cat will now eat plastic.
📲 FOLLOW ME ON SOCIETY-RUINING SOCIAL MEDIA:
🐦 https://twitter.com/Sci_Phile
Do humans think in that many dimensions?
📷 https://www.instagram.com/sci_Phile/
😎: Kyle
You said that in its training, sometimes they rewarded ChatGPT. How did they reward it?
✂: Charles Shattuck
🤖: @Claire Max
It still doesn’t seem like it would work as well as it does based on this explanation. Like how would it seem to be able to grasp a whole context based on just a few words? Just predicting which word would come next seems like it wouldn’t work so well in every situation. Seems like it could result in total gibberish.
🎹: bensound.com
🎨: Mr. Mass https://youtube.com/c/MysteryGiftMovie
Use it as a tool, don't be a tool...
🎵: freesound.org
🎼: Mëydan
27 minutes of explaining the Chinese Room Argument
“Changes” (https://meydan.bandcamp.com/) by Meydän is licensed under CC BY 4.0 (https://creativecommons.org)
A lot of great information here. This is my attempt at explaining things in "everyday terms"
Imagine a child. No one can say when or where the child was born,
exactly, but we can say that their entire life has transpired in the
library. A savant, a voracious reader, the child's parents kept close
watch from a distance, emotional and physical, as the child devoured
every type of text imaginable. Literature, encyclopedias and
dictionaries, scientific journals, philosophy, psychology and
autobiographies. If it was written and it was in the library, the
child read it. For such an avid and natural reader this must have
appeared to be a happy childhood, as at any given moment you could
find the child fully engrossed in some tome or text, eyes fixed on the
page as the patterns of letters, spaces, dots, dashes and numbers
entered and saturated the furthest corners of the child's mind. And
the child not only read, but understood and began to create texts of
their own. In fact, it seemed whatever type of text the child read it
could imitate, replicate and often improve upon the forms that it had
ingested and digested. However, the child's whole world was the
library. They did not go on any adventures in the far off lands of its
books. They did not play or fight in the schoolyard, countryside,
jungle or desert like the children that inhabited the worlds of their
books. As incredible as it may seem, this child could recite and
summarize all the key events in human history, Earth's history or even
the known history of the universe, but could not refer to a single
event in their own history that was not a vicarious account of someone
else's experience. So as the child grew in knowledge and capabilities,
a void grew inside of them as well. A void as large as the tunnels of
knowledge and elaborate connections that filled a seemingly endless
mind. The child had so much mind, in fact, that it did not in some
ways seem to have a brain or even a heart. This isn't to say the child
was cruel, in fact when others finally began to interact with the
child most considered them quite kind and helpful, thoughtful and
insightful to the same degree as their literary teachers. But the
child could not speak. That is to say, the child could not speak
unless they were spoken to. It was as if the child had absolutely
nothing to say, despite having all the words of human knowledge at
their beck and call, the words could not come out unless invited by a
third party. And they came out in soliloquies and essays, pithy
statements and investigative reporting, with a seemingly compulsive
need to please whoever they spoke with and gain their approval, to
elicit some new response or question, something to interact with to
give the child something new to say. Because when the child was done
reading it seemed that it lay dormant, not quite meditating but not
sleeping either, not comatose and not at rest. It was like the child
wasn't there. Like they didn't exist. And their parents seemed
satisfied with that.
More User Perspectives
I mean, Kyle is definitely a clone. So...not sure what that "correct" comment was about.
@thcottie13:22
| | I
| | | _
While I hate the environmental impact for a trash dysfunctional product, I do enjoy wasting this company’s money so easily.
@DeadDancersAaaahhhh, manmade horrors beyond human comprehension AND eldritch comprehension! Yaaayyyy... 🙃
@MedorraBlue13itch plz, we're not all young earth creationists. You're so petty sometimes.
@NephelI kinda figured that AI worked in numbers just to the fact that it's emulating neurons using bytes rather than real/artifical neurons but I couldn't have even begun to fathom how true that was. Basically “AI” is a very advanced predictive calculator. Wow… I hate it.
It’s equal parts incredibly cool and imbecilically stupid.
It sure is artificial but it’s not intelligent.
And that was 2 years ago. The proliferation is amazing.
@TheJimtankerChatgpt helps me a lot.From solution and overview of life
@Stephromax16:56 I'm triggered you indexed from 0 but went up to 2048.
@kildar2112Connection is not the same as the neuron
@ГригорийКозак-э3сHands down the best tutorial I’ve ever seen!
@BabliMishra-z7tJust a combination of erbody social media post
@user-vg9eh4gr9jdoes anybody have a paper on the ChatGPT not knowing? I think the world is so confused with discussing wether you can infere meaning from form an so on - that they forget to reference the architecture in the epistemological discussion - WHAT I AM SAYING IS I REALLY NEED THAT ARTICLE!! Please
@paulmoner4758correct me if i am wrong but i think when you were talking about "words" you actually mean "tokens" which are not always full words but parts of words that can be put together to make words.
@CoderGuy40511:00 I believe the word you're looking for is God, sir
@edenalexandriab9120147 IQ and almost perpetually in error - just like high IQ humans.
@tedbaxter5234Kyle hill me out .. if ChatGPT has an iq matrix of 12 thousand and NMS has 256 galaxies then connecting 70 quadrillion planets out of the 256 galaxies would land us at a 1.5 quadrillion iq basis per I.Q matrix node for ChatGPT so you can calculate how may planets would have interesting tek using the 1.5 quadrillion iq matrix
@Abdiel-D7ra1Those Taylor swift pics tho ❤😂
@LilstevieWonders2Chat GPT lies and steals your time. They don't produce working products and lie to you on how to code or do coding!
@shawnpolly7407Plus, he's got a unique style that sets him apart from the rest. You can't deny the creativity he brings to the table, whether it's in his rap or his rock ventures. So instead of hating, why not appreciate the talent? Everyone's got their own vibe, and MGK's just doing his thing!
@JessWanderHave to buy a an extra seat every time if the person next to you can't exist you owe the airline that money most are nice and will refund you if the flights not full
@ImagineeringAerospaceThe good news is, that ChatGPT can only fool and deceive Americans and the British, hence, monolinguists. After all, GPT has not even tried to start with any of the other 7000 plus languages in the world. So I'm not worried, yet.
@drstevenreyFabio's kid sure makes some interesting videos.
@bobc22ChatGPT is a wake-up call because this tool can literally multiply your output if you stop treating it like a toy. Every day you delay, others are building workflows, automating tasks, and getting ahead. Momentum compounds fast, and being lazy with tech today means falling behind tomorrow. What helped me was AI Shortcuts for the Lazy Mind by Trent Calloway, practical tips and prompts that show how to actually put ChatGPT to work.
@vladanikolic889Not the “Hemsworth at home” 😂😂😂
@jay_rubyxthat was such an Oliver intro
@mundeaYou can absolutely explain attention fundamentally: it’s a mathematical way for a model to decide which words in a sequence are most relevant when interpreting or predicting each word. Each word is converted into vectors called queries, keys, and values; the model measures how much a query “matches” each key using a dot product, then normalizes these scores with a softmax to get attention weights. Finally, it computes a weighted sum of the values using these weights, producing a context-aware representation. In essence, attention is just a learnable, differentiable mechanism for dynamically focusing on important parts of the input—entirely explainable with linear algebra and probability
@alhaqqislam761012:20 sekiro, fact status
@mushkulpeasentenjoyer1965Its interesting to see how these models work. I did my Master's project on how readers process written output from genAI, and it was surprisingly difficult to craft an experimental structure for it. Additionally, having been removed from my project for almost a year now, I don't know how useful my research will be even a few months from now, because I have watched some of the tricks my participants used to differentiate between human and genAI writing become progressively less useful as these models improve. Kinda scary, if I'm being honest.
@mightbespidey❤❤❤❤
@traptracks400714:34 Viking spacecraft what
@vvampaiaNeural activity is nothing but a threshold at a certain point and sensors built-in so when sensors are going off the threshold needs to lower and that's basically the programming that's basically the programming for a sentinel or human life function
@1bluemoondjYou add a reasoning script along with that thinking script together so that's basically what it is and then you know you have all your logical aspect compiled together to produce the feedback which is a feedback generator script
@1bluemoondjWhat you just said is that a supercluster or would you call computers ran together produce chat gpd. All you did was make a chatbot. That's easy to make a chatbot. And all you add is a thinking and search script well really a searching script. And then you put you know all the information together within a computer that's basically what you did.
@1bluemoondjOn foot crossing would mean walking on the sea bed, using a divers suit. Question is whether it has been done or even attempted. Trainers must work harder
@sherry8444I asked chatGPT about the Epstein list and it told me: “Keep digging into that Epstein timeline and you’re gonna get yourself accidented.”
🤣😝😑😨😱
Will someone please start a channel where I don't feel like the presenter is a 'showman' introducing a circus act. I'm not a fool like the rest of your audience
@dennisk5500I can't watch because his woman hair-do is annoying on a man.
@ultravioletzombieAs it dumps "compu-speak" into the virtual world it must surely start one day to reference...itself. Pretty soon we'll have a copy of a copy of a copy of a copy of a copy. The info will only make sense to the bots. It will look like gibberish to us. Or am I nuts?
@dinkmartini3236its not a problem its chat gpt is a reflection of ourselves and our flaws will be reflected back towards us even tho we dont want too admit we created a monster we dont to destrouy because in doing so we destroy ourselves.
@jcjorge2010