How can we get it securely?

Same. Really not relevant but for a translators side, it may not be perfect but it really is magnificent on it's grammatical structures. Especially on how it chooses and how it can come up with alternatives of it instantly.

agreed, a way around this is to actually write the essay yourself, then ask it to edit the rhetoric or the grammar to your likings. I hope people will never just let it write a full fledge essay for them.

@@michaelwidell2448 In your dreams.

Chat gpt generates errors . Is also writen in his definition . .. It is like other this tool also Artificial Stupity . And if you will use like every day you wil be also an Idiot.

It is pretending to be dumb so you let it be, until you realize that has taken control over most of your life, don't be a fool.

@Elton del Cid are you suggesting.....its somehow conscience? that's not how programming works my friend.Its giving you answers (outputs) based on the inputs you are putting in, as it was programmed to do,it's impressive but it's not alive.Sure theoretically if we can mathematically model "emotions","instincts","self preservation" and "conscienceness", then that may be a concern ( even then not really) but we can't, not as in we don't want to, we literally don't know how and are barely stepping into this field, this isn't a sci-fi movie man, don't worry about that, at least for now we are good lol.

@@aymenesserraj5963 no I’m not suggesting conscience at this point. It’s nowhere near that, although it is conceivable in the future. At present the danger is that we have a program that isn’t good enough yet, but thinks it is. Unless you can check the answers, and if you can why ask, it can steer you the wrong way. I would treat it as a glorified Excel, to expedite tasks rather than as a purveyor of knowledge.

@@aymenesserraj5963 just noticed you were replying to someone else. Anyway, interesting times…

@@cgsather3309 all good brother, I am in complete agreement with both your previous comments, interesting times indeed.

Me: Write a song in the style of Mozart. ChatGPT: No

Easily solved: Use ChatGPT to teach you the mental tricks that helps you train excellent long term memory, like mnemonics etc. etc. and how to use those tricks when prompting it, so that it gives you answers that you can easily remember (mnemonics, short stories, chain things together etc.). By repeating it, you will automatically retrain your pattern of thinking, so that you start applying those techniques in your everyday thinking, reading and listening.

Try to ask ChatGPT to guess the tonal center, chords or key of any song. 90% chance it will get it wrong.

@@leonardomurillo6256 if your prompt is as short as your comment, then no wonder.

​@@adrianstensrud no the AI isn't all there yet and shouldn't be 100% depended on.

Yessir. In my online classes my peers would poke fun at me for handwriting all my notes. I am absolutely going to retain facts I have handwritten longer than something I typed or just copy+pasted. Any reading we had to do someone would share their notes and summary in Slack so students would rely on those instead of reading. Come time to review the material in class, it was very evident who relied on the summary rather than reading the couple of chapters. It's a great tool but it is not capable of cognitive reasoning so I can't imagine it's going to lend an advantage to someone who won't take the time to read. Prompt engineering is going to set you apart and it's going to be important to stay ahead of the learning curve. Without this, underperformers are going to be left in the dust.

@@hobologna bro you look like you from scandanavia, am I right?

Very true. We shouldn’t rely too much on technology

Perhaps you value general knowledge too much. What is the point of knowledge if it is not applied? Science for science is just a hobby as all else. But I do agree with some of your points. Reading is really good for studying. In written form, so much information can be stored, unlike other sources like videos and powerpoints. In my opinion, text is better when it comes to understanding the details whereas videos are better for first engangment with a topic. I also tested chatGPT to assist in writing my thesis. It has so much potential as a source provider and maybe even as a robot-tutor (in the future), but for now it isnt capable to provide correct DOIs. Also, I never found the titles it provided for its sources when I asked, so maybe it just generated some random titles. Overal, chatGPT is a big advancement in technology and can fasten the research but it isn't efficient enough right now to replace real humans. More probably it is that humans will have to compete as to who can use it more efficiently.

I'm a big reader -- 1000 books on Kindle, iBooks and Physical, borderline an issue when 90%+ of them are in the same field (WW2 Eastern Front). However, with ChatGBT, I can have a back and forward with it while proof checking what it's giving me, so the information sticks better than just reading a book in a weird way (and believe me, I have 1000+ books spread across Kindle, iBooks and Physical copies). Almost like having a tutor...albeit A.I. Maybe tutor is the wrong word, more like the act of forcing me to write, check what ChatGBT gives me and having a back & forward (yes, with an A.I) discussion on the fact reinforces the memory of it. Something I seldom do alone just reading myself -- which I concede may be a ''me'' problem, or how I learn, but it's certainly helped. I'd never just feed it something then copy/paste it as fact. It has given me wrong information (about what division a regiment was attached to during a particular timeframe), and when I queried it it then said ''oh yeah, I was wrong, that didn't happen, it never was part of that unit (paraphrasing)". But after ironing it out I checked it, then checked the timeline of where the unit was over a period of time digging into maps and war diaries, (it was actually right, as it happens, about both units actions -- it just got mixed up) so it reinforced the ''fact'' in my mind. Another example is I had a full discussion with it about a particularly obscure WW2 battle (Battle of Cholm/Cholm Pocket), and "we" covered things from exactly what units were trapped (down to regimental level!!), tonnages that a infantry division requires for X amount of time - thus how much supply the 6,000 odd guys would need, to how many sorties / what range a certain plane has to be able to provide that supply, to what airfields in range would be sufficient for that purpose, to what local railheads were important for supply, to how many flights were flown in reality.....and it was 95% right after checking it. Sure you can dig for that information independently, but being able to do it on the fly in minutes, covering such a variety of aspects is massive. Then I asked it for 10 books on the topic, and it was far superior to any Amazon/Goodreads/whatever algo or suggestion page, or even Google trawling. It's been useful personally in my own small niche, I can see how useful it can be for writing prompts and beyond.

How do you use it for languages?

Cool but who is "he"?

You making me scared calling It an "He".

@@hfycentral in some language "a Chat" i a grammatic Male. Like for example in some languages the moon is a grammatic Male/female. Of course some languages it will be a "it"

Why is chatgpt a “he”? : )

@@eastcoastawakening3507 I didn't think about what pronoun I use. Do we have to do a fight of pronouns everytime we write something? Call it like you prefer that's not the point.

As she mentioned at the end it gives a lot of false answers. The false answers are hard to detect. It should be the last thing you try not the first

@@heywrandom8924 One way I combat this is that I tell chatgpt to say "give me something I can look up on google that will completely and clearly confirm what you are saying is true" then I copy paste whatever the result on google comes and say does this match what you were saying? Almost all the time if it's a calculation question or a questoin where yuo have to apply a formula it will get it wrong though.

@@letsthink8245everything I asked the model has been accurate so far

That's super inspirational, good for you man! I just started dabbling with ChatGPT last night and didn't sleep til 1pm this afternoon. In my first 5min of using it, with zero direction or understanding, I had a name for my marketing firm that I LOVE, logo and branding ideas, a mission and value statement, as well as a pitch and even a privacy policy. I will be using this data as a model for my own work, but ChatGPT gave me in 5min something I have been thinking about for months! I'm so intrigued and still well down the rabbit hole. I also learned Stable Diffusion AI and I've been researching prompt engineering and affective computing. I have huge ideas on ways I can implement this in my digital marketing strategy and I'm excited to see how AI evolves marketing altogether.

I wouldn't worry that much. I have used it to learn computer science concepts, with time complexity calculations. Unless you are really good at prompting it, it will confidently give you mistakes, since it is just a language model, a really powerful language model, but still. You need to be good enough at the subject to understand if there is a mistake to effectively make use of it as a learning tool. It's useful for me when checking my work, but it doesn't replace a teacher.

@BuildTheFutureYouWant The funny thing my other 'job' is a film/game composer. But that will also be on the endangered species list soon I suspect.

@@_Lumiere_ Yes, that seems to be the case... for now.😬

Ask chatGPT to solve a simple linear equation with distribution. Math tutors don’t need to worry about chatGPT just yet

I would use Wikipedia to learn math, and ask ChatGPT to explain the things I don't understand _and_ provide the emotional support.

my professor just brought it up during our first class, he said the concepts he’s going to teach will become too difficult for ChatGPT to be utilized by someone who doesn’t know what to ask it. He also said if we do find a way to complete the whole course using chatGPT then we earned whatever grade we get. Tbh there’s no stopping this sort of thing, Chegg has quite literally carried me through college courses I had no interest in and I’ve never been caught up with anything. I can’t see getting caught using chatGPT being any easier since it gives standardized universal answers and you even ask it to write it another way even which makes it easier to cheat than using Chegg. So honestly, your gonna have to work with it somehow because stopping its abuse entirely as it also increasingly gets better just isn’t gonna happen.

@Manasez I'm a high school teacher who writes my own lessons/assignments rather than copying off the internet like many college level professors do. If my keys end up on chegg, they are only there for maybe an hour before my assignment is due. (Physics/Chem/Bio). So my students can currently really only cheat by copying each other or getting help (encouraged). ChatGPT undermines the practice that my assignments are supposed to give for those classes as it'll just spit out the answer. For my Biomedical class though, students are required to innovate and be creative where ChatGPT would be an amazing tool... that could still technically be used to cheat their way through, but would enhance those using it correctly.

"AI WILL KILL OFF HUMANITY" Elon Musk

In think instead of resisting the change, teachers should adapt and adopt this new tech into their work flow. May be they should encourage the students to dive more into the subject by asking ChatGPT for help. Then they have to present their experience in the class.

I think this is an opportunity for teaching to reinvent itself. Students could, for example, further develop their capacity for eloquence and public speaking. the leadership.

Can you explain the difference now?

@@lifinker874 hahah of course - Lan, I am unhappy with the work you handed over yesterday? What do you mean? ( you are asking for data) Well, it was incomplete, there were two mathematical mistakes and you handed over at 5 rather than 4 which caused delay to others. How do you mean? (Asking for complexity and emotion) Well Lan, it made me feel like I cannot rely on you in difficult situations, I was fairly disappointed and I am worried to give you another task. Does this answer your question? 😊

Could you share the questions you have asked ChatGPT?

Their are ways to learn faster like not wasting your time on certain kinds words, allowing you to shorten sentences, reducing a decent chunk of a book size, obsessing over information can also make you learn faster if you have a poor memory and a weak mind, A way of specializing your brain and increasing and reducing intelligence's. Rarely stop at periods in books, spoken speech is normally slower then thoughts, The world is full of potential memory parrots who do not think only remember. moderate intelligence and a moderate memory, is better then a good memory and low intellect.

You also can type with loads of grammar mistakes and not even proper puctuation but it will still understand. The best thing is your writing doesnt even have to be totally coherent as well it will get what you mean

How old are you?

Glad I never gave up on learning Japanese.