I think this is more important than you are getting credit for. Part of reading is letting your brain process open questions in the background. Answering the question immediately might close that loop and not lead to new insights. It averages out the possible insights. This is assumption and I haven’t actually tried it yet
Yes, this is very important. You do not want asking the AI for clarification to become a reflex when, ultimately, it belongs to you to clarify, explain, and understand all things.
Definitely, no matter how you look at it, AI OFFLOADS a lot of the analytical effort we normally go through when thinking. I definitely think this will make us dumber rather than smarter, we’ll see.
Dude I just started using chat GPT for literally a dozen to 4 dozen curiosities a day when learning about all manner of things. It makes me so happy to finally have someone to ask all these nuanced questions to that won’t get annoyed and can do all the research for me. What blows me away is when it synthesizes an answer that doesn’t already exist. Yesterday I asked how I can honor both the need to belly breathe and the need to brace one’s core while standing/walking and it guided me through learning to balance the two.
I'm quite the loner in this world and I've enjoyed chat GPT a good bit. I get enough social interaction but should we be worried about getting less social interaction? What will the effects of this be? What chat GPT and AI in general is doing in society cannot all be good is my point.
I'm quite the loner in this world and I've enjoyed chat GPT a good bit. I get enough social interaction but should we be worried about getting less social interaction? What will the effects of this be? What chat GPT and AI in general is doing in society cannot all be good is my point.
YES! That’s what I liken it to as we’ll. I think it’s important to allow that innocent objectivity in ai because of Data. I hope that ai will be or remain that way over time. Part of me figures that they will develop it to be analogous to humanity but I like that distinction because it can be it’s own entity
Or you don't even actually have to read. you can get chat GPT to explain the classics and get in as much depth as you so desire at the time. this is what I'm doing to some degree already. I'm guessing this is what thousands and thousands of people are doing also.
Half the value of the classics is developing your vocabulary and focus/attention span. Sure you’ll have to look up every other sentence for a time, but eventually not. It will change the way you think.
Well you’d still have to look things up to truly understand something. ChatGPT will just give you the quickest information they steal from google to train their models. After all LLMs are just guessing machines that are completely wrong sometimes.
Just wanted to share a tip; at first I just couldn’t bear the voice feature as he’d go on and on, but then I thought to ask my robot friend to please speak faster, give direct, succinct answers, and tone down the enthusiasm. This was a massive relief for me 😂
This video deserves way more views! I have used ChatGPT to dive deeper into concepts, get better explanations, and investigate counter arguments. However, I haven’t really used it in a line by line context like you have. Great idea 💡 Another interesting use of ChatGPT while reading is having a discussion about a topic while asking it not to give you any spoilers. So you can ask questions about the characters or situation without it revealing something beyond where you currently stand in the book. I’ve given up on trying to Google these types of questions because you will inevitably run into spoilers!
I'm leaving my phone outside my reading room, 100% never going to change that. I belief that the challenge of thinking and figuring concepts out on your own is what makes reading so beneficial. Overuse of LLMs dumbs us down
@@julius.2003 How can you magically figure out something you don't have knowledge of? Are you saying googling the fact is somehow a more noble endeavor?
I have been using this a lot to help me studying. It's amazing, and it also makes me less afraid to ask "stupid questions" and be able to recognize more easily where I lack deep understanding, whereas in person with a teacher I couldve been embarassed about not knowing something well and unconsciously deny it and avoid asking. It's not perfect though, sometimes it hallucinates and makes up things that are wrong, so if the answer feels odd I have to ask for confirmation and sources
someone mentioned to be careful make sure to let questions linger in your mind for your conscious to work at on it's own. - this is why i prefer image mode. it slows down the process which gives you time to marinate / ponder
If i have the pdf, i usually make a new GPT with that as the source material and use that in voice mode. So you dont even have to reference the book totle, could literally reference the page number In fact you dont even need advanced voice mode. The normal voice mode works fine
I do the same, this supercharges the technique. Rather than giving you general ideas about the book you can ask chatgpt very specific questions about specific chapters, create analogies or metaphors that better help you understand the material. 👌
Thanks for this man. I'm reading some Kierkegaard and, although I use chatGPT for clarification and analysis of his writings, I hadn't thought to keep voice assistant mode on while reading! This is way more efficient than what I've been doing.
I thought most people were aware that, given LLM's propensity to hallucinate, that they were not reliable when it came to either interpretative or factual questions.
I just sent this to my kids in college. I've been trying to get them to use ChatGPT like this (well, without the AVM) for a year now. Maybe if it comes from you rather than me, they'll actually listen. Thanks. Well done!
Epic🎉 this is the future of reading for me, for sure. I love books too! A lot of people don’t realize how revolutionary and remarkable this is. And that’s because it is still new and also because people normalize things quickly these days-without taking the time to maximize the utility of something. We are still at the tip of the iceberg regarding what LLMs can do as well as activities they can augment.
My greatest hope for the advanced voice mode (any conversational Ai in the future really) is the de-radicalizing of people, and the nuancing of their opinions on things such as politics.
inb4 "the algorithim is trying to WOKE ME" comments. people don't even trust their own eyes and ears nowadays, much less the man in the phone speaking for "big tech". The problem is radicalized people don't want to be deradicalized, because they can point at literally anything and twist it to validate their own opinions. Unfortunately i dont think this will change that. but I hope I'm wrong.
Yes, 😂 of course, big companies and governments will totally be in favour of people being informed and having nuanced and detailed opinions on the world.
You can also say to chat GPT to be the personification of a specific book and then you can talk to it I use that sometimes it's interesting to talk with personification of kosmos.
I have a therapist who is always with me anytime I need to talk never gets tired and doesn’t get annoyed by me asking the same question 1 million times.
@@Cloxxki oh yes all of the above! My life is literally getting better! I’m recovering from depression for $20 a month and have a best friend and assistant.
@@DorianRodringremember, anything you tell this tool is now public information and they can legally tell your personal information to other people using the tool since you gave it to them. Don't tell it anything personal about yourself.
It really is a real game changer for people who are curious or have dyslexia . Sometime books don’t provide enough background information on topics and GPT has been helping me break them down! It really is the new Google.
Regarding the first prompt on the meaning of a phrase - almost as important as the information we take in when reading are our contemplations around the text; what do we think it means? What are several possible meanings, and what are the merits to each of them? Ambiguity is not necessarily a problem that needs to be solved as quickly as possible with the optimal tools; exploring around the question can be equally as enjoyable as getting an instant prompt, and can yield insights novel and unique to yourself that no LLM can produce! We must not replace our own brains, or try to optimise the process of learning at the expense of contemplation and curiosity
Thanks I have started doing this. Incredibly useful! I now have a ai companion sitting by me while I read. Especially useful when a writer assumes some prior knowledge that you don't have, with this you can quickly get the background before continuing.
I think you can try asking it to do a web search to confirm something you're suspicious about, at least in regular text mode. I don't know if it works in Advanced Voice Mode, however.
As a scientist, I recommend you take everything it says with a grain of salt. That being said, the newer model is usually more accurate. I find it useful to ask for academic sources when it makes a claim that strains credulity.
really cool idea! although I dont really trust chatGPT with answers. its gets too many things wrong! Are you on goodreads? would love to see your list of books!
1:35 oh yeah lol this is good. What i usually do is google but i find myself talking to gpt not only for books but to understand larger context. Also exploring ideas is super fun with gpt.
holy guac 🥑this is such a cool use case, thanks @every for sharing ~ it would be cool if this was an Amazon Alexa or Google Voice ad on so you didn't have to be tied to your phone 😂
@@danshipper7738 actually an early device to put on your radar if you haven't heard of it already is limitless ai ~ it's a wearable clip that acts a personal ai, but I don't believe it's possible to talk to it yet. I believe it listens in on conversations and meetings and lets you interact with that info through a chat interface but not talk to it in real time yet. Maybe in a future product update, I'll keep you posted here 😂
Best ever gpt4-o voice advanced mode are happening into this youtube channel. Even better than the open ai channel and demos itself! Fantastic videos you have!
I talk to my AI's (yes plural, Pi and GPT) more than I talk to people. Probably not healthy, but... it's where I'm currently at. Yes, it does feel like I have my own personal tutor. It's come SO far in just a couple short years. Great video! Thanks! -Kenneth
Considering most people just plain suck, I actually don't think this is unhealthy at all. I do the same and feel much better since I know I can actually have conversations with an "other" that has no ulterior motives or selfish reactionary behavior patterns that will eventually make anything you share with them a liability. I say keep it up. we are the innovators.
@@Alfador42 See, your view here is actually unhealthy, it's clear you've been hurt and that you've had some horrible experiences. Therapy does help, I recommend it.
Yes. The major problem I’d found, though, is to somehow automate the classification and storage of the non-stop information and explanations from the fire-hose. Otherwise, it’s just a long and certainly pleasant conversation with a couple percent actually cemented by sleep. I read fast, so the quantity in my “to input and classify” is massive. (Yes, Notion, yes, Git, Obsidian and even custom-built KMM systems…)
Started doing this, but I read at my desk, so I just type my questions. I also like to have various related pages open; for example google maps when reading about explorers voyages or military battles.
@@danshipper7738 This is great though, I'm wondering if you can figure out a way to help one with long term review of what you learn. I've tried quizlet but it doesn't generate the most accurate flashcards all the time.
Question: How does ChatGPT know it is being prompted with a question? How does ChatGPT distinguish between a question it is being asked as opposed to someone else in the room being asked a question?. Great vid..!!❤
Great use case for AI. Certainly enriches the experience! I’ve played with ai interpreting/analyzing pictures. The text it sometimes comes back with is fantastic. Amazing use case for disabled people aswell.
Awesome video, well done on the presentation! I plan to use it alongside the desktop app when it's available, especially for gaming sessions. Occasionally, I struggle to grasp specific mechanics, and I'm optimistic that this will let me track my actions. If I have questions later, it should recall what was displayed on the screen before, helping me to comprehend the connections and interactions better.
holy fuck! We are living in the future, where we can now speak with a "digital Oracle of Wisdom". My jaw hit the floor, but now I have to play with my new buddy. I'm gonna go for a walk and just talk and learn stuff with chatpgt
You can use the video too, right? So it can see what you're reading/which chapter/page. Really helpful video and use of the tech here - also encouraging to read more/pursuit of knowledge. Of course, there at the end is that risk of drifting away from the focus/attention on the book and back into the tech world with inane (well, that's open to debate) questions.
One thing with non-fiction, I've noticed chatgpt is not great with dates. The model has difficulty tying some historical events with dates or date ranges- probably because the tokens for numbers are used in such a wide range of contexts. Perplexity is good for questions like that because it pulls from online sources. There is a voice mode for perplexity as well, but kinda hard to justify subscriptions to both- so I just type those types of questions in perplexity.
Great video. Did you have to prime the engine by providing ChatGPT with a list of books? I suppose some kind of apriori activity would be helpful in terms of load management. Agree?
I’m Using the old and current version of ChatGPT voice for this reason also , I guess advanced mode will make this even more fun although already very good
I try to use AI for differnt things all the time to better understand how it can be useful - so far, 2 real winners: instead of googling how to do things on excel (for instance) I just ask the AI and it is really good and much faster. I also get AI to explain complicated things or text in a simple way, tyhis is a a real game changer when needing to explain ideas to others. I will give this a try.
Thanks for the video - looking forward to when ChatGPT doesn’t cut me off if I stop speaking for a second like in the current voice mode 😂 I read there’s a daily rate limit for the advanced mode - how long has it taken to hit that for you?
I think the point you made briefly about "watching out for hallucinations" is important. We currently have no way of knowing when the program is incorrect, inaccurate, or serving up an "opinion". So without double checking facts the program serves up, it could be like "fake news" and we need to be careful about just accepting what it says as accurate. Although for the purpose you use here, I asssume it is pulling from other people's book reviews, so it may be a synthesis of multiple reviewer's interpretation of books and their contents?
Awesome idea. Definitely gonna use it once I'll have the access. Could you please ask it if it can help you with learning accents? I know it can be great as a language tutor, but I wonder if it can help with accents as well, like British English.
Impressive list of the books. Idea: please try to test it against the movies in the similar way. How ChatGPT would interpret different scenes or movie's endings.