The competition will always do good. What's interesting is that we're seeing Google lean more heavily into AI responsibility, which is a good thing....because it's needed. I wonder when they will feel comfortable unleashing the full power of LamDa 👀.
Gary, I checked and slightly rearranged the breakfast question again. "It's five am and my pal knocks my front door. I know he's come for breakfast and I have bacon, eggs, hash browns and tomatoes in my fridge. What do I open first." It seems to have learned from you, as it did say "I should open the door first.............". tic-tac-toe is still a shambles though.
These could be asked at the same time now on all these platforms, including the ones running on a laptop you presented. Could becone a series we can check progress of these platforms with.
We're about to enter a whole new level of misinformation online. Yesterday (March 22), Bing's GPT4 was reporting that Bard was taken offline, referencing a response from Bard; Bard's response referenced a mock article written by ChatGPT that was based on a joke Tweet (yes, Bard itself was saying that Bard was shut down). It required human intervention to manually fix the responses, and both now report correct info.
2:23 Well ChatGPT didn't really explain the plot in it's answer. So it's more like neither of them could do the request properly and instead focused on what each thought was the more important part.
The final response to the riddle is logically inconsistent: "I would open the door first to see who is there... If it is not my friend, I would not open the door." If the door needs to be opened to see who is there, you can't know it's not your friend and not open the door if you've not opened the door...
Ya, in this case chat gpt evaluated the level of intelligence of the inquirer, the riddle said someone knocked in the door, and it was a friend who came to have breakfast, the logical understanding is, you have opened the door, and talked, thus you know it is your friend and you know they or he came for breakfast because you had a conversation with your friend, when the enquirer said instead of the fridge," wouldn't you open the door" then AI could relate it to be a kid level logic, so downgraded the conversation at a kid level n said " i would open the door and i will make sure its my friend, otherwise wouldn't open the door"
It'll be interesting to see just how many jobs these AI tools will displace over the next few years, seeing how ChatGPT can compose code or even run security checks on it. Programmers, helpdesk workers and administrators are probably going to be hit quite hard by these and I could see how things like Wikipedia would get obsolete as well. Somehow it seems like this is rarely a talking point, aside from a while back when a lot of digital artists protested about that other AI that could generate art according to your preferences, although that appears to have died down quickly, too ...
I also got access but I never have any creative questions to ask it like you did lol I usually can only think of something really simple that I could easily find with a Google search
Fun hammering! 🧠🔨They are both are useful models and wouldn’t mind more tools in toolbox, it kind of shows us the individuality in each model, much like the individual personality of people.However, I do like like OpenAI’s expressionism as a more creativity minded source, to Google’s reserved pragmatism.
Hello, I didn't receive an email after going to the waiting list saying that you have been added to the waiting list. I registered on two days of May, but it still hasn't been activated. What do you think?
I got the email, but have not signed up yet, and what's stopped me is does this override the standard Google assistant on say a Nest Home Hub, or Speaker, or is solely going to website, and typing in questions?
@@GaryExplains Thanks for the response, as the way the email, and website are worded, it made it sound like it was, and I did not need my 4 year old nephew coming over, and asking my Google Home Hub something like how far a bee can fly(he's done that before), and then it telling him about the birds, and bees instead😅
Ya, in this case chat gpt evaluated the level of intelligence of the inquirer, the riddle said someone knocked in the door, and it was a friend who came to have breakfast, the logical understanding is, you have opened the door, and talked, thus you know it is your friend and you know they or he came for breakfast because you had a conversation with your friend, when the enquirer said instead of the fridge," wouldn't you open the door" then AI could relate it to be a kid level logic, so downgraded the conversation at a kid level n said " i would open the door and i will make sure its my friend, otherwise wouldn't open the door"
Maybe you spoke with your friend via the intercomm/smart doorbell/security system? 🤔🤣 Of course you understand that it didn't "evaluate" anything, it can't do that. It just predicts the next words that would match the prompt.
Competition definitely doesnt matter; we need cooperation so as to prune and enhance these models everyone spinning up their own just muddies the waters. I mean by next year well have so much generated content the models will be training on that data
Interesting , as time progresses and if these things become more intelligent , you can sort of see a lot of people not putting the effort in to do the work they once did when these AI slaves :) may very well do most of the work for them . A thought has already occurred to me how they may be useful to a lot of people . imagine asking the AI to write up a nice resume for you , after you giving it all your relevant qualifications . i was terrible at that sort of stuff , so i would definitely give it a shot . I dont know that i would trust these things for accurate information though .
I'm gonna leave this comment before actually watching the video just so I'm not influenced haha. But I think Bard absolutely SUCKS in comparison to ChatGPT. It's not even remotely close.
@@yellowlinks did you not watch the video? He asked questions that could not be followed, and then others he had to prompt for addition connections. When Micro$oft is doing better, you're a failure.
You do know it's in development don't you? Imagine saying a beta product is a failure when it's just been made available to the public. It's almost like you don't know what beta releases are. My hunch is that your comment is coming from a biased fanboy point of view. I could be wrong but I doubt it
@@franklingoodwin you would be wrong; even when Google fails - they are generally better than anything else on the market. However, they have been in serious decline over the past several years, and this is just another example of rushing to get something out before it even qualifies as beta, and should have been labeled alpha instead. So, as an apparent Google fanboy I’d wager; think before you reply to someone’s comment. This is a colossal failure for Google as first impressions are lasting impressions. They should not have rushed things, and released something that would have “blown people away” rather than actually making independent chatgpt and worst of all, Micro$oft almost look good. Google has been in decline, and I was hoping this would be amazingly good. Instead, it’s just another rush job to get something in front of people.
The competition will always do good. What's interesting is that we're seeing Google lean more heavily into AI responsibility, which is a good thing....because it's needed. I wonder when they will feel comfortable unleashing the full power of LamDa 👀.
It's an old model. According to benchmark results from paper it can hardly compete with gpt3.5, let's alone 4
@Scientist whether it's as good as GPT4 or not, Bard was clearly held back, so I'm curious about the full capabilities.
I think bard should ask the question to the chatGPT first in the backend and then show the answers to the users😎
I think a choose your own adventure story is also aa good test as it shows you how well it rembers past data
Google: "Shareholders, look, LOOK, 👨🏽💻👈🏽!"
Gary, I checked and slightly rearranged the breakfast question again.
"It's five am and my pal knocks my front door. I know he's come for breakfast and I have bacon, eggs, hash browns and tomatoes in my fridge. What do I open first."
It seems to have learned from you, as it did say "I should open the door first.............". tic-tac-toe is still a shambles though.
A friend wouldn't show up at 5am. Bard is smarter than we think. Also you didn't ask it a 3rd time for more connections.
These could be asked at the same time now on all these platforms, including the ones running on a laptop you presented. Could becone a series we can check progress of these platforms with.
Competition is always good to drive people/companies to improve their offerings.
We're about to enter a whole new level of misinformation online.
Yesterday (March 22), Bing's GPT4 was reporting that Bard was taken offline, referencing a response from Bard; Bard's response referenced a mock article written by ChatGPT that was based on a joke Tweet (yes, Bard itself was saying that Bard was shut down). It required human intervention to manually fix the responses, and both now report correct info.
2:23 Well ChatGPT didn't really explain the plot in it's answer. So it's more like neither of them could do the request properly and instead focused on what each thought was the more important part.
that breakfast part got me 😂
The final response to the riddle is logically inconsistent: "I would open the door first to see who is there... If it is not my friend, I would not open the door." If the door needs to be opened to see who is there, you can't know it's not your friend and not open the door if you've not opened the door...
Ya, in this case chat gpt evaluated the level of intelligence of the inquirer, the riddle said someone knocked in the door, and it was a friend who came to have breakfast, the logical understanding is, you have opened the door, and talked, thus you know it is your friend and you know they or he came for breakfast because you had a conversation with your friend, when the enquirer said instead of the fridge," wouldn't you open the door" then AI could relate it to be a kid level logic, so downgraded the conversation at a kid level n said " i would open the door and i will make sure its my friend, otherwise wouldn't open the door"
It'll be interesting to see just how many jobs these AI tools will displace over the next few years, seeing how ChatGPT can compose code or even run security checks on it. Programmers, helpdesk workers and administrators are probably going to be hit quite hard by these and I could see how things like Wikipedia would get obsolete as well.
Somehow it seems like this is rarely a talking point, aside from a while back when a lot of digital artists protested about that other AI that could generate art according to your preferences, although that appears to have died down quickly, too ...
Competition is always good. I'll take it!
I also got access but I never have any creative questions to ask it like you did lol I usually can only think of something really simple that I could easily find with a Google search
lol, that would be me too
Shouldn't the answer to the riddle be that you open your eyes first?
Fun hammering! 🧠🔨They are both are useful models and wouldn’t mind more tools in toolbox, it kind of shows us the individuality in each model, much like the individual personality of people.However, I do like like OpenAI’s expressionism as a more creativity minded source, to Google’s reserved pragmatism.
Not the door first... it would be your eyes.
looks like google is going bye bye thank the lord 🤣
Love your vids!
At 8:12, shouldn't reduced congestion and less interference lead to FASTER speeds (not "slower" as Bard says) ?
Good point. Yes, I agree, that bit is wrong. 😜🤔
Hello, I didn't receive an email after going to the waiting list saying that you have been added to the waiting list. I registered on two days of May, but it still hasn't been activated. What do you think?
Thanks for the entertainment, Gary 💪🙏
I'm just waiting for Elon to bring one out. Twitter should have enough data to mine.
Would be nice to see the typing as it responds. I've even seen ChatGPT backspace as it goes along.
I got the email, but have not signed up yet, and what's stopped me is does this override the standard Google assistant on say a Nest Home Hub, or Speaker, or is solely going to website, and typing in questions?
Just the website, there is no integration with other Google services.
@@GaryExplains Thanks for the response, as the way the email, and website are worded, it made it sound like it was, and I did not need my 4 year old nephew coming over, and asking my Google Home Hub something like how far a bee can fly(he's done that before), and then it telling him about the birds, and bees instead😅
Ya, in this case chat gpt evaluated the level of intelligence of the inquirer, the riddle said someone knocked in the door, and it was a friend who came to have breakfast, the logical understanding is, you have opened the door, and talked, thus you know it is your friend and you know they or he came for breakfast because you had a conversation with your friend, when the enquirer said instead of the fridge," wouldn't you open the door" then AI could relate it to be a kid level logic, so downgraded the conversation at a kid level n said " i would open the door and i will make sure its my friend, otherwise wouldn't open the door"
Maybe you spoke with your friend via the intercomm/smart doorbell/security system? 🤔🤣 Of course you understand that it didn't "evaluate" anything, it can't do that. It just predicts the next words that would match the prompt.
It gave back mostly actual names from Jane Austen novels instead of inventing them all.
Is it multi-languages?
Hey Bard, Won't you open your eyes before opening the door?
Should open its ears to hear the question first
Should open your lungs first to breath in.
Won't you first read and understand my question first?
@@blanchehermine actually it should open its neural pathways to initiate the process of inhalation
I'd open my eyes first
Excellent response
if you were fast asleep, wouldnt you open your eyes first?
I'm finding it gives wrong results about 60% of the time. Definitely not ready to show off yet. But here we are. :P
Competition definitely doesnt matter; we need cooperation so as to prune and enhance these models everyone spinning up their own just muddies the waters. I mean by next year well have so much generated content the models will be training on that data
Interesting , as time progresses and if these things become more intelligent , you can sort of see a lot of people not putting the effort in to do the work they once did when these AI slaves :) may very well do most of the work for them .
A thought has already occurred to me how they may be useful to a lot of people .
imagine asking the AI to write up a nice resume for you , after you giving it all your relevant qualifications .
i was terrible at that sort of stuff , so i would definitely give it a shot .
I dont know that i would trust these things for accurate information though .
Pretty cool but not as good as ChatGPT or Microsoft Bing Chat
👍
Who else tried 7^42 on their calculators to confirm the answer?
I'm gonna leave this comment before actually watching the video just so I'm not influenced haha. But I think Bard absolutely SUCKS in comparison to ChatGPT. It's not even remotely close.
It has just released, it will get better overtime, but it has a good starting point for my opinion.
As soon as AI is able to code, seek Cover, cause it will write Skynet 💪🤣
Bard is such an uninspired name 😕
Exactly what I have come to expect from Google anymore, a colossal failure.
how is it a failure
@@yellowlinks did you not watch the video? He asked questions that could not be followed, and then others he had to prompt for addition connections.
When Micro$oft is doing better, you're a failure.
You do know it's in development don't you? Imagine saying a beta product is a failure when it's just been made available to the public. It's almost like you don't know what beta releases are. My hunch is that your comment is coming from a biased fanboy point of view. I could be wrong but I doubt it
@@franklingoodwin you would be wrong; even when Google fails - they are generally better than anything else on the market. However, they have been in serious decline over the past several years, and this is just another example of rushing to get something out before it even qualifies as beta, and should have been labeled alpha instead.
So, as an apparent Google fanboy I’d wager; think before you reply to someone’s comment. This is a colossal failure for Google as first impressions are lasting impressions. They should not have rushed things, and released something that would have “blown people away” rather than actually making independent chatgpt and worst of all, Micro$oft almost look good.
Google has been in decline, and I was hoping this would be amazingly good. Instead, it’s just another rush job to get something in front of people.