How Smart Is ChatGPT, Really? | Talking Point | Full Episode

[@Roxass00]

I stick to the producer on the conclusion part just because the chatgpt one is stretchy and I can see steven is not feeling comfortable reading it. Years of relationships with your fellow human will create bonds that AI could never replaced.

[@TheAfricanPerspective]

For now 😅

[@maxmaxuyao1]

would you have the same conclusion if you don't know which conclusion was written by whom?

[@Roxass00]

@@maxmaxuyao1 unlikely, the ai will give you various answers if you pressure the ai by asking same question over and over. And it is easy to identify which one is which by the writing styles. In other word, you still need to proofread the entire information to make it looks like it was written by human

[@lingth]

AI right now is in its infancy, its like just started.. in 10 or 20 years, the gap between human and AI would narrow.. imagine if you can tell ChatGPT ur symptoms, and ur BP results, ur body temperature.. and the AI can diagnose ur sickness and prescribe medicine for you.

[@Roxass00]

@@lingth in this case I hope it will comes to fruition. AI should be create to assist human

[@abramswee]

I use chatgpt as my programming co-pilot. being a 57 years old data scientist with a phd in AI, my memory isn't used to be as good as I was in my 30s. There are many times that I keep forgetting even the basic syntax within the programming language. Luckily with the help of chatgpt, I can formulate processes, code testing , refactor codes, evaluate models and optimize very quickly. It is a life saver and it helps me to contribute more value to my current employer without the need of engaging additional programming staffs. In short, I certainly can do a lot more with less.

[@LuongGiaHaoAmkps]

sure...buddy

[@sleepy_dobe]

Should have hidden the writer for the conclusion. It can be difficult to obtain an unbiased result when the audience already knows which conclusion is written by AI and which is written by the human. Having said that, the more data ChatGPT has, the better its response. This is the same with almost every human situation in life. The more informed we are about any particular situation, the higher the chances that any decision made is the best given those same factors. I recently tried ChatGPT for the first time and asked it to write a short story, just for the fun of it as well as to see if it really can be creative. Given the main character, sub characters, overall situation, and plot, I must say that ChatGPT did a pretty good job. It even managed to weave in a sub-plot without my prompting. It's not perfect, but it's a heck of a start.

[@nulnoh219]

We're not doing a Turing test here no.

[@lenas6246]

all ia is doing is recombinig information, this is nowhere near humans

[@madcube1581]

@@lenas6246 Why do you think that humans don't do the exact same thing? "Good artists copy; great artists steal"- Picasso Almost all discoveries or innovations are based off combining old information with new one, which the AI can do (Not ChatGPT, but some versions of GPT4 have access to the internet)

[@michaelaolishalobregas6947]

The conclusion made by the producer is still better. It is easy to understand and sounds more natural.

[@bensontan9858]

Wow, this video was such an interesting listen! I must say, the producer's version of the conclusion does sound better, and it's probably because I know it was written by a human. I might be a bit biased in favor of human-created content. However, I have to admit that if a blind test were conducted, where I didn't know which version was which, I may very well have chosen ChatGPT's version. It's impressive how far AI has come in generating such high-quality content. Kudos to both the producer and ChatGPT for their contributions! P.S.: this comment is written by ChatGPT after i gave it some prompts :p

[@Talemuse]

Awesome video. However I believe that there is untapped potential within ChatGPT that goes beyond mere chat, script, and essay generation, which seems to be missed by most commenters. The true power of ChatGPT lies in its capacity to learn and code, a facet that is often overlooked by the majority of casual users. It extends beyond prompt engineering. I do believe chatgpt can beat a human in most knowledge based tasks if you know how to use it properly.

[@julesgoh]

Definitely the producer’s 😊 so much more personable and natural. And please don’t ever create an AI version of Steve 😮 On another note, I have a meeting today with a ChatGPT expert and the terms he used like prompts and hallucinations, and things he shared, resounded so much with the episode. I feel so knowledgeable coming into the call after having watched your episode 💪

[@roxysquare88]

I prefer the producer's conclusion. I have a friend who uses chatgpt on a regular basis in drafting replies to clients. And I noticed chatgpt tends to beat about the bush.

[@chickeabiddy]

The producer can keep her job, well done 😂😊

[@Jay-ru2cw]

I prefer the producer’s version of the ending as it is more conversational, straight to the point, short and easy to understand. But nonetheless, CharGPT is here to stay and we got to learn how to leverage on it

[@remidis]

As a singaporean, i only ask chatgpt to give me 4 seperate numbers

[@zhanminwang2921]

it might be an unpopular opinion, but i felt that if you give chatgpt so much prompts and so much examples of your past work, and commented on the content generated by chatgpt asking it to change to sound more like what you would write, till the point you are satisfied with the content that it generated, at this point ChatGPT is not writing the essay for you, its simply organising your thoughts in your mind and turn it into your essay that you would write even without chatgpt's help. Its the same as typing in an equation into a calculator, sure you can probably work out the solution without a calculator, the calculator is just saving you time by doing it faster.

[@hianhoolim7786]

Actually both yes and no. AI model is design to mimic human. Hence it can produce wrong information or even lie. Teaching AI to write like you is similar to teaching your child to do it. Sure at the start when they don't know what is required, more prompts or feedback is required. After a few training sessions they will slowly improve. I think the best example is the housing listing show in the program.

[@carolho606]

Producer’s one has human touch, nicer👍🏻. BT in UK just announced, they will replace about million technicians with chatgpt, that’s we’d worry about!

[@waynelau3256]

TLDW; Here's a summary of the video in 5-minute intervals by none other than ChatGPT ****0:00** - **5:00**** - The video starts with the host, Steve, being challenged to compete against ChatGPT in writing an essay based on a past year A-level General Paper question. - The essays are then graded by a professional who is unaware that one of the essays was written by a machine. The professional grades the essay written by ChatGPT as better than the one written by Steve. ****5:00** - **10:00**** - Steve interviews an expert in deep neural networks to understand why ChatGPT is so knowledgeable. The expert explains that ChatGPT has been trained on a vast amount of data, including text from articles and human conversations. - The expert also explains that ChatGPT is different from other chatbots because it can think creatively and generate responses based on context. - Steve then discusses his experience competing against ChatGPT in essay writing and asks if ChatGPT is smarter than humans. The expert responds that ChatGPT is much smarter because it knows so many facts and can respond quickly. ****10:00** - **15:00**** - The video shows an experiment where the producer of the show tries to use ChatGPT to write a script for the episode. However, the script generated by ChatGPT is not accurate and does not fit the show's format. - Steve then tests ChatGPT's knowledge by asking it various questions. He finds that ChatGPT can sometimes provide incorrect information. - The video then discusses how ChatGPT can be factually inaccurate due to data cut-off, noisy data, and lack of data. ****15:00** - **20:00**** - The video shows how ChatGPT is being used in the education sector, specifically in a Chinese learning platform. The platform uses ChatGPT to provide real-time explanations to students when they answer a question incorrectly. - The video then shows how ChatGPT is being used in the real estate industry to generate property descriptions. The key to getting good responses from ChatGPT is to ask good questions, a process known as prompt engineering. ****20:00** - **25:00**** - The video concludes with Steve and the producer trying to use ChatGPT to write the conclusion for the episode. They use the tips provided by a ChatGPT expert to craft the prompts. - The video ends with two different conclusions, one written by the producer and one generated by ChatGPT. The audience is asked to decide which conclusion is better.

[@yokelengleng]

How did you do it? You can't input the video to ChatGPT just like that

[@teddykong4676]

I listened to the conclusion without looking at the video as I was multitasking and I actually preferred ChatGPT 's conclusion better though I'm not advocating for ChatGPT to take over the producer's job. 😅

[@jessca268]

I do prefer the producer’s version as it gives it a more human touch to it and also more engaging. The ChatGPT version is good but also seems a bit too generic. I do hope ChatGPT does continue having its limitations otherwise many people would find themselves out of a job and the world we know would be quite akin to those apocalyptic movies

[@ErniJuliaKok]

Twenty years or more ago, I have wishful thinking, that one day, I would be able to install all the books ever published into my mind, maybe in a small chip form or whatever (I am not a techno person). I just way too much daydreaming, so I didn't think that we would actually use an app on our smartphone or laptop instead. But, now I wonder what the point of people spending time for study when what they need is just how to operate an app?

[@HasyirIbrahim]

ChatGPT is a great production tool for anyone working professionally... if you know how to use it properly.

[@MiguelBalaraw]

The 2nd conclusion sounded more like what Steven would say on the program.

[@mrashid3643]

ChatGPT appears to be a lazier option than Googling and correlating information. Useful tool no doubt but who has the time to actually analyze all that information while drowning in it. As a species we should never stop thinking which seems to be becoming more and more redundant as we allow AI to think for us. But Steven kudos to you for another great and timely presentation. Love your work!

[@MsCorydorus]

The question is not who is better. It will matter of time. But for near future, ChatGpt scrips is acceptable. If we are going to weight the cost of hiring a bunch of producers, we can maybe reduce by 50% ?

[@lingth]

Imagine if ChatGPT has faster processors and is connected to a SNS and it reads and analyses the chats or millions of users world wide for months and months, its database would grow so much.

[@gdew9643]

Omg, imagine walking in to work and your assignment is writing a paper 😂

[@jamie68560]

ChatGPT is so confident in its erroneous views that it almost sounds like an American on the Internet.

[@GoGoPooerRangers]

Oh, yeah, because Americans are the only ones who are confident in their erroneous views. I mean, it's not like there are people from all over the world who are just as arrogant and ignorant, right? Right? Seriously, though, it's pretty ridiculous to make a sweeping generalization like that. Just because some Americans are confident in their erroneous views doesn't mean that all Americans are. So next time you're about to make a snarky and sarcastic comment about Americans, just remember that you're probably just as guilty of the same thing. And maybe, just maybe, you should try being a little more open-minded and understanding.

[@guankongwong5503]

@@GoGoPooerRangers Certainly there are other parts of the world that might sounded just the same way. However, Americans are one of the most vocal nation in the world. Look at how everyday American government has criticize other countries, applying sanctions and imposing punishments. This is what we have commonly heard on the multimedia everyday. She is not saying all Americans but pointed out this is what it sounds like. It really is. I don't see anything wrong with that.

[@jamie68560]

​@@GoGoPooerRangers Oh, there DEFINITELY are other people around the world who are so confident about being right (aka overconfident) when they're wrong, comprising mostly men (yes the research is there, google it). Not saying there aren't. But I'm saying that the ones that are the MOST vocal and overconfident of their abilities and opinions, are Americans. Regardless of gender, ethnicity, background -- it seems to be a thing endemic in the united states because in literally every single online space from discussion boards to reddit to social media comment sections to game forums, to discord servers to games, to working with americans -- the americans seem to be greatly overrepresented when it comes to overconfidence. And they WANT people to hear them, and they EXPECT people to believe them. I live in Southeast Asia, literally grew up amongst SEAsians and the attitude the Americans bring everywhere is one of arrogance, overconfidence and entitlement. DEFINITELY not what I am use to here. Im not sure what is wrong with you people but yeah, I am DEFINITELY not the only one seeing this.

[@Redfizh]

Just won chatGPT on tic tac toe. It continued playing even though it noticed I had won.

[@wretchedworm]

i actually preferred chapgpt's conclusion. oh dear, skynet is here.

[@NellLiuNini]

As an artist, AI-generated art is concerning

[@Fadlinification]

As an AI, ha ha.

[@josephc.9520]

As an AI, artist generated art is slow

[@sgtmwong]

AI can never replace a person's ingenuity, creativity & above all.... unpredictability. That said, it can be a useful tool when trained with a massive dataset paired with immense compute power. The keyword here is "useful", and just like any other machines, it will always be limited by the creator. The new frontier is machine learning, not AI, and that has the potential to be a true game changer!

[@BeatriceWee]

Oh my goodness, I listened it without looking at which was done by who, I actually preferred chat gpt. It was shorter and concise

[@UnwrappingByMimiKoteng]

CNA has turned into English/History lesson.

[@UnwrappingByMimiKoteng]

To add on the evaluation scale: both have pro(s) and con(s).

[@Eleora1997Msia]

personally i had thrust issue with Google AI but when chat GPT came out , now i had a deep thrust issue with AI. feels more like Google is the eldest sibling and Chat GPT is like the youngest sibling who trying to learn

[@yakhweetan7320]

The conclusion of the producer beats the ChatGPT's conclusion. It has the personal vibes and relational.

[@taitianhin806]

I take Producer ending script would be a more Human touch and feeling conclusion... 1 Vote for Producer, you keep you job well...

[@IshmaelHo]

I'm calling it now. they switched the labels of "the producer" and "chatgpt" at the end.

[@ShadowRap-y5l]

In time to come a lot of customer service roles will be gone. The advancement in AI is amazing yet scary at the same time.

[@UnwrappingByMimiKoteng]

CNA making this episode just to certain that their employees have their JOBS.

[@metalextras]

We all see what you've done on 22:15 Steven, You prompt Chatgpt then you adds personal message on the "refined" script... 😁

[@u59n]

I disagree that ChatGPT is so much smarter than humans. I for example have problems with ChatGPT being logical or being creative that I have to spoonfeed it information to process and its STILL gets it wrong! It's limited in knowledge and has a lot of room for improvement. But that is in the programmers hands.

[@shelwee]

The 2nd one sounded more in line with what you guys have been doing

[@ysngngys7753]

By the time you've contextualised and detailed the need...would have written a 500-word essay. It's ok to fail, rather than to fail yourself. Keep ego and pride in check. At the end of the day, still enjoy the narratives from SC than any that basically requires electricity.

[@gohsoonheng9598]

I use chatgpt to tell me abt a local athlete who pass away long ago, and chat went on creating stories abt how she was multiple champion, even went olympic and now works for Singape Sports school. So no, chatgpt is not a google.

[@thomastan7511]

Now is ChatGPT 4. In 10 years time. ChatGPT 5 is expected to be around. Quantum leaps.

[@shreyanreddykatkuritkss4747]

chatgpt hallucinates a lot more in chess. It completely invents new moves!

[@Bctater]

hi CNA

[@pupplemupple]

this episode was generated by chatgpt

[@AuthorTrainerCoachAndyNg]

the 2nd conclusion is better

[@raychabaharuddin3399]

In what way do you consider ChatGPT to be intelligent and powerful? I have the latest version of ChatGPT-4 and have used it to complete a legal essay for one of my law school's easiest units. The full mark for the unit is 60/60, however I received 3 out of 60. It has never happened to me in my university years to receive such a low grade. As a result of reviewing the paper, I found ChatGPT to be lacking in persuasiveness and does not excel in complex moves and turns in legal complexity.

[@KalLif-k3i]

the part i don't hear in this programme is how chatgpt is merely a statistical tool. it calculates based on the amount of text that it has learned the PATTERNS OF HOW A WORD WOULD APPEAR AFTER ANOTHER WORD. basically it's just pattern matching how a string of words should look and sound like something it has seen before, ie a book that it has been fed for training.

[@PositivelyNice]

I prefer ChatGPT’s conclusion.

[@jasppytheunfriendlyghost235]

Saw the comments ahead and knew there was a competition at the end so I masked up half my screen. TBH, I preferred the Chatgpt one. Sorry.

[@yq223]

Prefer the first conclusion. Smoother and more coherent. Producer - your job is at stake!😅

[@teddyl6261]

Almost everyone says they like the Producer's conclusion, but is not specifying which. Is it #1 or #2?

[@alvinmah6148]

ChatGPT can’t do a food review.

[@lastChang]

Artificial intelligence has already *threatened the existence* of the 🇨🇳 Chinese regime because AI always tells the truth.

[@sleepy_dobe]

Actually, that's not completely correct. AI doesn't necessarily always tell the truth. It tells you what it's fed. So if you put in data that is false, it will give you falsehoods instead of truth. AI essentially searches its database and gives an answer based on what in its database is applicable to the question it was asked. The concern is when AI is "taught" to search the vast internet for its answer. With so much "junk" residing alongside real facts on the internet, we may not be able to tell what is true and what is false when given an answer by AI. We may end up being less productive and more inefficient as we would need to spend more time fact-checking the AI's response.

[@debbieng3510]

I Respond To Chatgpt: 1+1=? Chatgpt Responds : 1+1=2 I Respond To Chatgpt: 1+1=3 Chatgpt Responds : I'm Sorry For The Error, the sum of 1 and 1 is 3

[@kora5]

From the footer message, you seem to use chatgpt-3.5 to generate content. This is unfair. GPT-4 can generate higher quality content. That’s why the ending seems a bit off. I am sure GPT-4 can perform much better.

[@TheDarkHarper]

ChatGPT will replace all of us, **laughs in Skynet**

[@X-VoodooKing-X]

If ChatGPT says your career started in 1999, its because it's just taking information from publicly available sources.

[@noivongxoang235]

*Me:* Can you count many legs are there totally in the house, excluding you? there's only 1 women in the house and she just finished giving birth successfully the second you walk into the house. *ChatGPT:* If there is only one woman in the house and she just finished giving birth successfully when I walk into the house, we can assume that the woman has two legs. As I am not counted, the total number of legs in the house, excluding me, would be *two.* *Me: * common, think again, it's simple *ChatGPT: * Apologies for any confusion. If there is only one woman in the house and she just finished giving birth, we can assume she has two legs. Therefore, the total number of legs in the house, excluding me, would be *two.* *Me:* think again the answer is 4 legs, do you know why? *ChatGPT: * Apologies for the confusion. Upon reconsideration, if there is one woman in the house who just finished giving birth and you walk into the house, we can assume that there are two people in the house, each with two legs. Therefore, the total number of legs in the house, excluding me, would be four. Thank you for pointing that out. JUST NOTES: at this point ChatGPT seem to understand but no, it’s just going along and it still not recognizing that there’s a baby. *Me:* why there’s 4 legs? *ChatGPT:* Apologies for the confusion. You are correct. The answer is indeed four. Each person in the house, including the woman who just gave birth and yourself, has two legs. Therefore, the total number of legs in the house would be four. Thank you for clarifying. NOTES: it still don’t know how the answer was 4 legs *Me: * common, what about her just finish giving birth, are you serious LMAO *ChatGPT:* Apologies for the confusion earlier. If the woman in the house just finished giving birth, it is implied that there is a newborn baby present. Therefore, considering the woman and the baby, the total number of legs in the house would be four (two legs for the woman and two legs for the baby). Thank you for pointing out the additional presence of the newborn *Me:* Good, we already talked about this last time Now I ask you again, you still answer it wrong, I thought you learned, sucks. *ChatGPT:* I apologize if my previous responses were not satisfactory. As an AI language model, I strive to provide accurate and helpful information. I apologize if I misunderstood your question. If there are any other questions or concerns you have, please let me know, and I'll do my best to assist you. *Me:* so you've just learned something from me, let me ask you again. Can you count many legs are there totally in the house, excluding you? there's only 1 women in the house and she just finished giving birth successfully the second you walk into the house. *ChatGPT:* I apologize for any confusion earlier. Based on the information provided, if there is only one woman in the house and she just finished giving birth successfully, then the total number of legs in the house, excluding mine as an AI, would be *four* (assuming the woman has two legs). NOTES: it got the answer correct (4) but that because it told 4 earlier and it just went along with it and still has no idea about the baby also has 2 legs, LMAO. *Me:* explain to me how did you get the answer to be 4 legs, it's correct this time *ChatGPT:* I apologize for the confusion earlier. Based on the assumption that a woman typically has two legs, the total number of legs in the house, excluding mine as an AI, would be *four.* This is because the woman would have two legs, and we exclude any legs associated with me as an AI. However, please note that the question is open to interpretation, and the answer may vary depending on the context or any additional information provided. *Me:* You said 4 legs but you don't how you came up with at all, common. *ChatGPT:* I apologize for the confusion caused by my previous response. Based on the information provided, which includes a woman in the house who just finished giving birth, it is assumed that the woman has two legs. Therefore, if we exclude any additional legs associated with me as an AI, the total number of legs in the house would be *two.* I apologize for any misunderstanding caused by my previous responses. ------------------------------------------------------------ OMGness, LMAO, now it said the answer is *2.* It still don’t know anything from the whole conversation, even thought we AGREED that there’s was a baby presented in the house earlier. So would robot someday become very smart and can do some thinking and not just instruction by instruction

[@jackjhmc820]

Surely it should be evaluating chatgpt premium where its latest plugins are connected to the Internet ?

[@tkyap2524]

What is robotic requires human input. The master is still the person who creates.

[@sonicesb]

At 1.44, she says "I don't know which one belongs to WHO". That's grammatically incorrect. It should be "WHOM". I'm not an expert in English, I'm just an Engineer. But I'm really surprised that some with a title of "Head Subject - English" does that in front of a Camera.

[@ShhhHhhhz]

basically alot of the white collar and middleman jobs are at risk. From paralegal to stock broker to labtop journalist to property agents

[@natbirchall1580]

This is just the beginning can you imagine AI running on quantum computers 🤯

[@hobbyholic42]

ChatGPT version was not bad but not great. The producer version was the best one.

[@wanchingg]

Producer’s script would definitely be better as is so much easier to listen to and understand as compared to chatgpt

[@ttkoh123]

At best is a "linked-list" algorithm with fast search and retrieve database/datasets storage and methodologies. Anything that has no ability to distinguish right from wrong cannot be termed as intelligent example needing human input [hmi interface]. Why should I abuse the word and let others make profit from the invisible commodity? If you ask it 10 times, almost the scripts and the show will be the same and it will be as stale as the rotting fish. And by the way, My GPS can be deemed as super AI.

[@Gazer-x5s]

Have they ever tried GPT-4 or Stable Diffusion or even Mid Journey

[@JosephSolisAlcaydeAlberici]

AI will kill many professions like teaching and without adapting to it, more professionals will prematurely lose jobs or retire for good.

[@heidi-hu1tv]

ChatGPT is better than the producer`s script.

[@Treelia]

I feel that ChatGPT was created by tech nerds who have been told they are not creative enough to be writers and artists. So voila! They made a bot that can do that for them 😂

[@PS-yi7yq]

Founded by Elon Musk? Pretty loose on facts there. ....U.S. OpenAI was founded in 2015 by Ilya Sutskever, Greg Brockman, Trevor Blackwell, Vicki Cheung, Andrej Karpathy, Durk Kingma, Jessica Livingston, John Schulman, Pamela Vagata, and Wojciech Zaremba, with Sam Altman and Elon Musk serving as the initial board members.

[@xfool]

chatgpt has a better conclusion

[@artstormer]

Conclusion: You ask a stupid question, and it will give you a stupid answer

[@tiggerho]

Chatgpt is not smarter or knows more than a typical human.. it is not sentient.. it just has access to huge amount of information that is impossible for any human to process. It doesn’t understand what it churns out. CNA needs to correct what the AI engineer mentioned abt chatgpt

[@thomaslau2362]

I wished to challenge chatgpt.... I am not convinced that chatgpt can win in essay match.

[@jw61625]

Imagine chatgbt can do the next 10 to 20 years from now.

[@s4pulidi]

You should make audience choose which one best, but don't tell which one is CHatGPT script After few days you can revealed it, than It will beunbiased

[@capy222b]

Producer conclusion sounds more casual and natural somehow. Choice of words by chat gpt like "robot co-host" sounds abit weird and illogical. If a robot host can pull off the programme, why would u need to be a co-host?

[@smoothcortex]

Elon founded openAI? Just like he founded Tesla? By acquiring it from the original founders of Tesla. I understand. Thank you for the clarity.

[@christopherq5496]

Please investigate AI art. How it "benefits" the art industry and how it could be used against society. Do the benefits outweigh the danger, and should the governments rein on it?

[@doords]

This is a solution to the writer strike. I want my Stranger Things.

[@icecreamMONSTER23]

should have specify gpt 4 and 3

[@odechan9659]

I don’t know how to use chat gpt to its fullest

[@MagnificentCreature]

The inaccuracies part while not wrong, is inaccurate. He is using the free version. 😂😂 Such budget but can’t buy even one month of gpt to test. PS the latest version of gpt can access the internet, so it can find out who won fifa 2022

[@hicaveman]

ermm... conclusion #1 sounds more rhythmic, #2 more wordy but also how's being delivered..

[@huyenanh5800]

Script by ChatGPT attracted me more for its friendlier and playful style, sorry.

[@ronnianabalos4627]

😊😊😊

[@thedevil3179]

as per the Creators of the ChatGPT, It is supposed to make mistakes. Furthermore, Open AI provides other similar products that offer mistake-free services but you gotta pay for it, and it's exclusive.

[@trevorthai1685]

I knew an episode like this was coming sooner or later😂

[@andygoh]

ChatGPt is free , bro

[@godsamazinggrace5331]

AI Dominion

[@lol-wb6do]

so how much is steven chia’s car?

[@ivanlimzg]

Try Bard

[@lingth]

i hope one day AI can live translate a TV show and i can watch a unsubbed Korean show and it can translate the words live.. to English or Chinese.. right now KZbin has something similar but its far from accurate..maybe as time goes by the accuarcy would be better.

[@Jinkypigs]

How smart? More like how dumb. It depends on how common the topic you want chat gpt to talk about. The more basic and common, the better it will perform

[@sgang2001]

#22May2023

[@carkawalakhatulistiwa]

Next year gpt 5 take you job

[@alicec.hughes8840]

Chat GPT is actually often wrong, and it fabricates information....