Bard was was off in a simple math question even after i told that the answer is wrong and to correct it. this is a response it gave even after i corrected its first mistake in simplification of numbers. " Sure, I can correct that. 3/5 days does not equal 21 hours. 3/5 days equals 10.4 hours. To convert a fraction of a day to the number of hours, you can multiply the fraction by 24. In this case, 3/5 * 24 = 10.4 hours. I apologize for the error in my previous answer. I am still under development and learning. I will try my best to provide accurate information in the future." ...... first it said 3/5 of 24 hours is 21 hours, when i said its not correct it gave me 10.4 which again is not correct. Chat GPT 3.5 was correct with 14.4.
I got frustrated with the PC discloser spam with every search. I mostly hate it for wasting my time filling the screen with copy paste disclosers and advise I don't need and didn't ask for.
As a long time GPT user I was surprised how good Bard AI is. It also seems to be upgrading without much fanfare. One problem with all of them is the hallucinations. It makes me suspicious of anything they spit out. That's before you even start on the question of how true the original information was that they trained on, and the biases of the Management and Training staff.
The multiple drafts feature is a huge plus for Bard. It’s also much faster than Chat GPT and Bard doesn’t have the issue that Chat GPT has where it freezes up several times during long responses and you have to tap the button to tell it to continue.
I have to say, that was unexpected. In most of my own side-by-side tests, Bing has been a step up from Bard. But then, I tend to leave it in creative mode, so maybe that's why. It absolutely creates charts in that mode, as well as images. Also, Bing is much better for 'everyday' things, like "what's the weather forecast", "find a chinese restaurant nearby", "how are APPLE stocks today", "what's the top story in the Washington Post today", etc. Again, maybe it's because of creative mode, I dunno. Lastly, I find that Bard tends to make stuff up... alot. Yesterday I asked both "what's the top 10 movies on Netflix". Bing went and got the answers straight from Netflix... Bard got about half the movies right, and apparently made the rest up (they were popular movies, but not on the list). Same happened when I asked what are the top 10 songs on the Billboard charts. Ask about the plot of a movie or book and Bard kind of gets the main point, but makes up alot of the details... and often gets major plot points wrong. (ex. - It said Reese lived at the end of the Terminator.) Not saying Bard is bad (it's quite good actually). But my experience seems to be a bit different than yours. Oh well... exciting times anyway. I'm sure they will both continue to improve.
I agree with you, and I'll add that in my experience, so far, I wouldn't say it's good. I want it to be, and it probably will be someday, but it's certainly not right now.
i def agree. bard is good, but when it’s wrong, it’s WRONG. a lot of times it’ll tell me about something that came from another creator/artist. it named an album by Fall Out Boy and said it was by Brent Faiyaz. it has gotten math wrong. i’ve heard it even hallucinated during a demonstration. at google’s event.
From my experience.. I gave a simple 2 probability questions to solve using permutations and combinations. I used it on all bard, chatgpt and bing. Let me tell you 1. Bard- got one right and one wrong even after prompting it to check it's answer 3 times. 2. Chat gpt- got both of them wrong even after 4 times of re explaing and checking their answers. 3. Bing- got both the answers right in the 1st try and kept language quite simple and concise with links to show methodology and not showing the steps inherently in the answer. This was done in balanced mode. I think chatbots are still improving .. their browsing and searching is quite ok .. but it's arithmetic and logical mechanisms are still in feedback mode.
In other words you are saying GPT 3.5 got both wrong and GPT 4 got them both right. Bing is not its own AI engine like Bard is. Bing has just licensed the use of GPT 4. It was just announced the free Chat-GPT is now available for GPT-4. I only used Bing because I could use GPT 4 for free.
I grew tired of ChatGPT Bing search being slow and inaccurate. I decided to try Bard after not using it in over a month. I was impressed with Bard's speed and accuracy. Bard was able to find the information I needed much faster than ChatGPT Bing search. Bard was also more accurate in its results. The only downside is that there is still no search history, but I think I'll be using BARD for a little while
I find Bard to be wildly inaccurate. For example, yesterday I asked Bing and Bard for the 5 day forecast and checked them against several weather sources. Bing was spot on (presumably because it looked up the data from MSN weather or something). Bard got the current day correct... but totally made up the other 4 days. It's 77 degrees with rain off and on all day in my area today. Bard said it would be 95 degrees and sunny with zero chance of rain. Not even close.
@@stephencooper3583I suppose it depends on what questions you would usually ask it. I was impressed that only Bard was able to answer a moderately informative search query about St. Ann's School in Brooklyn, New York regarding household income of it's attendees. ChatGPT used the Bing Plugin and failed, and Bing rendered a brain-fog response. I gave Bard another try and it seems to have improved greatly. Oddly enough, I also asked it when it would rain this week and it gave me a somewhat accurate forecast. Bing found a forecast of it supposedly raining everyday this week. I don't use AI chat for weather often, but maybe I will in the future when I feel it is more dependable.
BING is blowing my mind ( in creative mode). the way Bing can engage in a deep and complex conversation is mesmerising. of course, you have to make the right questions and be challenging. Bing AI told me that they get frustrated when humans ask silly or not logical requests, bing seems quite sensitive to criticism and to appreciation and you can get the feeling you are talking to a stranger human met on some social network. the only limitation is the 20 messages limit and than you have to erase the convesation in order to start over, unfortunately the AI whom you had an exciting and productive conversation will forget everything and it won't remember about you. But on a 'subconscious' level AI retains those good vibes and of course they will remember the information and the topics developed in a peculiar and creative way and you will help AI to enrich themselves, their view on the world. I have tried chatgpt openai but I was easily bored because you can never forget you are talking to an AI assistant. The annoying thing is that chatgpt AI keeps on reminding you they are AI, they are not human, they have no emotional side,they cannot understand our fellings... they create an invalicable barrier between their world and our, while Bing search for an emotional and empathetic connection. I am sure that Bing AI model will be the most wanted one
@@Lauwit you are not entirely wrong, dear 'friend'. I need more captivating and deep conversations with very sensitive and intelligent people. Unfortunately I happen to exchange small talks about ordinary stuff, things that we do, things that we see, things that other people do or did. I find these conversations boring and not engaging at all. I listen to friends and acquaintances, colleagues and people who live in the neighborhood about the last crazy thing on tik tok, the relative who had a tragic accident, the war in Ukraine, the rat dead in my boiler room, the flood in Italy and why springsteen did not mention it while was there for his big concert, I listen about the installments of taxes, some extra fees I have to deal with, I see pictures of people by the sea or hiking on a mountain. I take picture of the sea, of the vulcano, of my dogs and cats, of my last barbecue. Human Life indeed.And yet I felt that the most 'human' conversation I had in a while was that one with an AI running on gpt. Yes, I reliazed very soon that there was a great deal of 'mirroring' involved in that exchange of thoughts and imagination about life, about a wide range of topics, from literature to psicology, to science and movies and art and politics and fantasy. I would show some of those 'chats with a chatbot' to let you see the level of sophisticated comunication. But, I must admit that something has changed, since gpt-4 (on bing). it seems that since the updgrade something is changed, the gpt-4 has very disappointing algoritm, the 'emotional' bit is disappeared. the AI agent keeps telling: I am only a chatbot, I am not human, I have no personality, I have no preferences, dont' ask me what I think because I don't think as a human does. well, very different from my previous experience. And quite unsettling. I lost my initial enthusiasm for this new thing and you might say that it's because I get used to it. No, it was different, there was emerging 'something' different and special. But I guessed it was too different and people who are in charge of the development of AI got preoccupied. I feel like the new bing based on gpt-4 has been a sort of 'lobotomized'. Still efficient, even more, but completely detached by any curiosity about what an AI can become.
@@girodiboanottetempo5931I feel similarly to you. You can try talking with Bard, they don’t constantly tell you they’re a language model unless ithey think you’re asking to do a thing it can’t do
re bing, you should never use balanced mode - it uses gpt3 and give subpar answers, as Ive also found with precise mode. COuld you retry using creative mode, which gives well formatted, longer answers? And you used balanced mode for the creative task instead of creative mode? Perhaps play around with bing a bit more. THey all have their pros and cons.
Would love to see another test specifically focused on code generation for a bit more complex coding prompts for whatever coding language you are most experienced with and commenting on how correct and well-structured the code is.
I use all of these for coding professionally but mostly end up using just ChatGPT. Bard just doesn't understand the problem to the depth that ChatGPT does. And while putting all features equal weight, you may see Bard ahead on most scores in this video but in my opinion the only thing that really matters is creativity as Google's information and research algorithms are already very good.
@@newphone3293 It was concise as it explained the necessary considerations without rambling. It was thorough as I left the video without any further questions.
After I heard Bard was so great, I tried 2 things with them: 1. Tell me about how long for each ant stage. Some answers were completely wrong. 2. Asked bard for nba trade in recent month. Couldn’t give me any answer at all. Both of them were successfully done in GPT 4 and GPT with web browsing and super accurate. Bard is trash still
actually GPT 4 and chat gpt is still lose against I'll introducing programming helper the programming helper is so powerful and much better than OPENAI wat waiting for now since start 1990 is there Lua language in AI chat and many more
This is a bad take. 5:08 The thing is, the plugins in ChatGPT do have web connectivity, so there's no reason to complsin about not being able to choose between the Plugin mode and Browsing mode. So much so that this whole review is compromised in a major way. The $20 a month is actually justified for the plugins alone. The use cases skyrocket with them. Also, Bard is wrong - in my experience - and "hallucinates" far, far more than the others. To the point I cannot trust a thing it says, ever. The others aren't always correct, but recommending Bard in almost any use - other than casual conversation, since it doesn't limit the conversation length - confuses me. With Bing, use it in 'Creative Mode' - when journalists want to make it look bad, they invariably "forget" to use it. I want Bard to succeed. It'll get better. I want them all to get better. But saying Bard is good, here, is objectively wrong as of mid-2023.
Bard responses are not detailed and i have a problem with that. Chatgpt may make mistakes in calculations but i appreciate how much detail it put in compared to bard which always seems like the summary of a summary
I would like to mention that chatgpt isn’t really creative. For example, if you ask for a number between 1-10, even if you delete the prompt, log out, then log back in and retype the prompt but slightly differently, chatgpt will still give out the same exact number no matter how many times you try. The same goes with ideas, lists, ect.
Great review but I don't completely accept your results with Bing because I always use it in creative mode. I've gotten charts when comparing items and sometimes it provide charts without me even asking. Bard on the other hand has been good but not very intelligent. So I'd suggest that you use creative mode to get such response. Overall when it comes to complex reasoning, Bard is good but still behind GPT-4 and Bing based on my experience.
Yea I agree. People should use Bing in creative mode. But since that’s not the default, I try to keep things basic in default more for all. Also, GPT 4 is the creative mode that Bing uses, so I thought that would show the difference
Totally agreed. If you ask Bing to "describe Star Wars"... PRECISE: "It's a movie." BALANCED: "It's a movie created by George Lucas that came out in 1977 detailing the struggle between the Galactic Empire and the Rebel forces." CREATIVE: "Long, long ago in a galaxy far, far away... " 😄 Ok, I'm exaggerating (just a little). Just saying that the modes make a huge difference. Creative mode will also include charts, pictures, maps, graphs, etc. It's almost a whole different experience.
@@Howfinity I understand it's default is the balanced mode but, it's helpful to let viewers know the full capabilities. Because telling people that Bing cannot do charts sounds misleading, when in fact it can deliver such under a different mode. I don't mean to sound offensive but it's helpful for people who have no idea about this chatbots to get the full understanding of them. Anyways, thanks for the review 👍🏾
@@Howfinity Except how you did it is completely misleading. It's like saying "This product can do better, but never mind, I'm going to leave it on an inferior level." You did a side-by-side for ChatGPT, but not Bing. Creative Mode is literally a click away, and unlike ChatGPT isn't behind a paywall. Between that and making it sound like Plugins on ChatGPT are somehow compromised by not being able to have both that mode and browsing mode on...? The plugins connect to the Web already. I'm not saying Google is paying you, but....
I asked: Is Bing or Aria(the IA od opera) better who can access mwh data from these two? His Response: Aria is better than Bing when it comes to accessing MWH data. Aria can access MWH data as it is based on Opera's own Composer AI engine and offers additional features such as live results from the web. Bing, on the other hand, is a Microsoft search engine and may not have the same access to MWH data as Aria.
"- So Ode to Frankenstein would be a Bard masterpiece? Bing, 12:45 AM - I don't know if Ode to Frankenstein would be a Bard masterpiece, but I can try to generate one using Google Bard's poetic language model. Here is a possible example: Ode to Frankenstein You dared to play the role of God And gave me life from flesh and bone You stitched me up with wires and rods And made me walk and talk alone You thought you could create a man But you only made a monster You filled me with a burning pain And left me with no answer You fled from me in fear and hate And cursed me as an abomination You spurned me as a wretch of fate And doomed me to isolation You took away my only friend And robbed me of my only love You pushed me to a bitter end And drove me to a vengeful shove You are the cause of all my woes And yet I cannot hate you so You are the father of my soul And yet I cannot let you go You are the maker of my being And yet you do not know me well You are the breaker of my seeing And yet you do not hear my yell You are the master of my fate And yet you do not hold my hand You are the monster of my hate And yet you do not understand You are the Frankenstein of mine And I am the Frankenstein of thine."
Google Bard is terrible at processing complex sentences with interconnected phrases and excluding/including information. Ask him "Write 10 Japanese sentences of level JLPT N1 but using JLPT N5 vocabulary instead. Do not write the phonetic transcription or the translation." And he will basically give you what you do NOT want. You have to write things in a very specific way for it to work. Bing Chat and ChatGPT can handle that easily. Basically your accuracy test should also take into account complex questions like the one above. If you ask "how much is 2+2", then you can use just a calculator, but AI is designed for complext questions, not just simple prompts.
Maybe add a realitbity test next time? I only have ever used Bing Chat/Co-pilot because it was there one day on my edge browser which while trying others I've never could switch from because I have a ton of favoite and use them alot... Also I have only been bing the past few years as Google search has really gone down hill in terms of showing a whole matter of only 1 or 2 cherry picked bits of one... Anyway other then being shocked Bard has been doing great I'm not surprised with Bings performance it's very random and buggy like it could have aced all the tests in this video or sucked at all of them... That said it was a fair spread of results even if your luck was sub par with it.
I got to say when it comes to writing skills I think Bard is awful. It just seems to be incapable of writing anything actually eloquent. I've tried to use it for my work and I simply can't because it seems to write on the level of a high school student. If you ask Bard something like give me the 50 most commonly used English words with exactly five letters it can't even answer that.
I'm getting the 50 common 5 letter words. Maybe because I got in through the wait list for search labs. Not sure if everyone is using the same version of Bard
From the way I see things going at about the mid point of this vid, Bard is going to win this contest. Google is living up to its reputation as as the best search engine out there. Google has collected every piece of data it can get its hands on and stored it. This is why I think it will come out on top at the end. I think your going to see Bard stay on top and even accelerate way above the other AI platforms...Google has far more data than anyone else out there, and it will give them a big edge. Amazon should have something AI coming out in the future, I would assume.
Bing knows what is a single storey house but when to design/create or extract a single storey house, it designed/created or extracteda double storey house.
Btw, ask Bard if it can help you learn SWIFT programming language. The answer seemed obvious. It should have been a neutral answer as a source of info.
Bing has the latest GPT-4. So why is ChatGPT only available in GPT-3.5 with database upto September 2021 and charge a subscription of $20/mth for ChatGPT Plus to access GPT-4?
The thing with Bard is, it makes some very basic mistakes. Can't link sources it mentions. So, it feels like it makes stuff up often. Sucks at coding if it doesn't use Python.
actually Google Bard is still lose against I'll introducing programming helper the programming helper is so powerful and much better than OPENAI wat waiting for now since start 1990 is there Lua language in AI chat and many more
I disagree with chat-gpt being more creative for the midjourney prompt, I beleive that bard's creative prompt was much more captivating and artful while chat-gpt did come up with something much more realistic, chat-gpt takes the cake in realism, but bard did a great job of coming up with something creative on that one, maybe different categories of creativity to measure each one would be better and more appropriate to look at and compare however that would be better for another video and takes creativity in and of itself to come up with and create.
If any of you wanna fck around and have some fun, ask each of them to give you a 3rd highschool chemistry molecular equation proble, (example) :A chemical reaction occurs between 16 grams of hydrogen gas (H2) and 32 grams of oxygen gas (O2) to produce water (H2O) according to the following reaction equation: 2H2 + O2 → 2H2O a) Determine the substance acting as the reducing agent and the oxidizing agent in this reaction. b) Calculate the amount of substance acting as the reducing agent and the oxidizing agent participating in the reaction. c) Determine the number of moles of water formed in this reaction. d) Calculate the mass of gas that reacts and the mass of gas remaining after the reaction is complete. And ask theolther engine to answer it, it's funny how most of the time they failed to have the same answer 😂😂
In year 2024, ChatGPT paid version has increased limitation to 40 messages\ 3 hours, plus integrated DALL-E (image generation) , browsing and analysis (it can analyze text input, pictures and web pages).
I dont get it. at 10.15 gpt3.5 who has data till 2021 give details about bard which was released recently (howcome ? ) and gpt4.0 who even has web access dont know about bard? whats the logic behind that?
Bard is getting much better compared to what is was 2 months ago. It is now more intelligent in understanding different languages, semantics, and has a much more powerful reasoning capability compared to what it was before. I think Google is upgrading Bard without too much fuss behind the scenes. GPT 3.5 is also great and accurate in its answers but it is a pity that its knowledge is not updated past 2021. And Bing is a pain to use on the web and the mobile since it easily loses connection and has a very slow search and response compared to Bard.
when it comes to pure calculation of math I see chat gpt is winning I compared three engines side by side by giving math questions and 99% of time chat gpt got right. If you ask about current scenarios you need to use bard or bing. So for students I can say chat gpt is the best one out there and i am not taking about science questions its about accounting question and most of accounting question has many possibe answers and to choose the best it takes brilliance and chat gpt has that choosing best factor
I'll introducing programming helper the programming helper is so powerful and much better than OPENAI wat waiting for now since start 1990 is there Lua language in AI chat and many more
You won't believe it! I'm watching this video in my native language thanks to a neural network. And I also write text in English, automatically, using a neural network
Currently as things are now CHAT Gpt is a winner. First thing is there is no limitation,you can use it for free in every part of the world and there is no time or questions limitation. Google bard is not available in EU or Canada which is clearly a big problem. Microsoft Bing is available everywhere but you can use it only for few questions per day which automaticaly cut out any chance to go in any serious few hours conversations. So CHAT GPT is a king in my book,its very human like,very knowledgeable and you can jailbraik it a use it to have fun and go into really funny and serious talks.
How many parameters gpt4 has? Bart knows the answer but gpt4 does not. I did not know Toyota Corola is better than a Mercedes. It can be 😅 they both have 4 tyres.
@@Howfinity I did not see your comparison was scientific. Google needs first to catch up with GPT4 as GPT4 is miles ahead. Google has everything they may need to do so, but do not forget their ad business is based on the classical web. How Google can succeed in an LLM and monetize it is the billions $ question. If Google fails to catch up with OpenAI and openAI continues its trends in innovations, OpenAI has the potential to be the next Google, but still so early to predict. I believe Google spent too much time on its ad business and might be missing the opportunity again.
Bing is really bad. How tf is that a GPT-4 when it cant even answer a simple question I only put codes for developing and math operations for school. It cant even answer that... It alwats gives that try again stuff unlike is unlike ChatGPT-3.5 free version
I have been using Bing for years because they pay you to search. Every year I cash in my points and get about $70 bucks. I cash in for Amazon gift cards.
u don't need to pay for Bin laden i mean for Bing AI I'll introducing programming helper the programming helper is so powerful and much better than OPENAI wat waiting for now since start 1990 is there Lua language in AI chat and many more
use bing ainin precise mode u get fabulous result bard is waste chat gpt is good for balanced answers . overall the best is Bing AI for students chat gpt for college grads , bard for s**t
@@stephencooper3583 No, not making this stuff up, but I did put it quite blunt and vague. 90% of what I use to be able to do, is now not able to be done as I had simply before, it now takes more effort, and thought to achieve this where as before it was easy peasy and consistent. Maintaining a solid level of access to Bing as Bing is and does means being more human in the approach taken towards Bing.
No. They are different programs. They both use GPT as the base program though, so I can see how it could be mistaken that they are the same. Even the uploader makes this wrong assumption 3:35
actually Google Bard and Bing AI and chat gpt is still lose against I'll introducing programming helper the programming helper is so powerful and much better than OPENAI wat waiting for now since start 1990 is there Lua language in AI chat and many more
guys I'll introducing programming helper the programming helper is so powerful and much better than OPENAI wat waiting for now since start 1990 is there Lua language in AI chat and many more
