Unboxing brand new Google Bard and GPT-4

Подписаться 71 тыс.

Просмотров 33 тыс.

50% 1

It's an AI unboxing video - this tech was released less than a week ago!
Watch me run my first ever Bard + GPT-4 side-by-side prompts.
I picked this prompt not because it's something everyone should do (hardly!) but because it tickles me, which is a great first try philosophy. There are many serious and powerful uses you can put this technology to, but I hope your first prompt will bring you as much delight as mine brought me.
If you'd like to read my thought process and snide asides written to accompany this video, you can find them here: bit.ly/quaesita_unboxing
*To try LaMDA, you just sign up for Bard and off you go. (Free, but there might a waitlist.) Link: bit.ly/google_bard
*To try GPT-4, you just sign up for ChatGPT Plus and off you go. (Costs $20/month.) Link: chat.openai.com
Curious to see what happened when Bard started writing poems? Subscribe to find out in my next video!
Comment to tell me what your first prompt was!

Наука

Опубликовано:

20 мар 2023

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 49

@Axiom23573 Год назад

“Based on this video please guess which company I work for…” 😂

@AN-iv6xp Год назад

Thank you for the video. Love the fact that Bard has an opinion, takes things to the next level - I am not sure if we are really comparing apples to apples here. They both look like very different products. Bard seems to be doing a more insightful work which is one step post stating facts

@alpha0xide9 Год назад

Bard was prompted to take sides though

@L.I.T.H.I.U.M Год назад

The difference is that GPT acts like a machine whose only job is to collect information, analyze it, abstract, figure out the patterns and present the appropriate information, while Bard tries to act like a human. I think these chatbots are only a positive as long as we keep them dry and machine-like rather than lively and human-like.

@gf62196 Год назад

ChatGPT plays safe. Although I like that it gives you an introduction about their background before it gives you an answer, I prefer Bard’s approach to a more straightforward answer.

@trappedcat3615 Год назад

You can ask both Bard and ChatGPT to perform a desired character or personality before the chat or mid chat.

@endlapavankumar7301 Год назад

Thankyou for the video, hope it'll available soon in India.

@ksdr5412 Год назад

I liked Bard's answer for Team Kant. It HAS an opinion, and that's what I wanted to see/hear for an intellectual language model. It understands and COMPREHENDS (!!! the peak of understanding !!!) the ideological difference between both schools of thought. Open AI is simply bringing in the definitions from the web, and I have observed it in many cases, it even presents a popular opinion, when asked for an opinion. Here, Bard seems to have its own thinking pattern, an original piece of thought. I like Bard.

@Thefare1234 Год назад

GPT-4 can also do what Bard does, but they don’t allow it. I remember GPT 3 in OpenAI s playground also provided interesting opinions instead of generic responses; not sure if it is still like that.

@Jysq Год назад

If they're allowed to have preferences and their own opinion. then they'll be biased on every queries you asked which is not really helpful if you think it through.

@Thefare1234 Год назад

@@Jysq Bias is not necessarily a bad thing for a discussion. Sometimes overly unbiased and encyclopedic answers just lead to nowhere and seem like a waste of time. I think users should be able to adjust the level of bias according to their needs.

@Jysq Год назад

@@Thefare1234 so just like bing chat with 3 modes. Creative, balanced, and precise

@Ahmed_Eltayeb Год назад

لThank you very much for this useful introduction. I have been following the videos you provide for quite some time, in addition to other content, and I found that there is a fundamental difference in the learning process between other content, especially the field of statistics. Greetings

@c.comploj3775 Год назад

How is Bard comparing to Bing? The Google stock does not seem to have picked up in the stocks, showing that Bing has more trust today in the stock market.

@johnnypetro9314 Год назад

I'm guessing that you like being employed by Google so I think we can all guess where this is going 😂

@Hyawnu Год назад

If an AI model has an opinion on a contentious topic, it basically means it has bias (preference). It should be up to human who should decide which is preferable, based on the objective information AI provides.

@jjmarcos Год назад

Contentious is subjective. AI will come to contentious conclusions all the time. Like the holobunga or black crime.

@c.comploj3775 Год назад

there is nothing objective in philosophy.

@ahmadpak Год назад

I like the UX of bard to put out all text at once. less visual noise

@I_Lemaire Год назад

No! The Typewriter effect in Chat-GPT has a functional purpose: faster response for the user.

@quatie Год назад

Do you think Google and Microsoft will eventually charge for access to their large language models that power their chatbots? Why or why not?

@kennethmackinnon1385 Год назад

Interesting to see the difference in approach side by side. The question I am asking is whether Google Bard is heading too much towards being "your plastic pal who's fun to be with"? Do we want friendliness and opinions from AI or do we want assistance with information?

@nabilah789 Год назад

What about having both? What stood out about Bard for me is the instant loading speed, and also it's straightforward answer style which feels nicer to read.

@satyakebakshi7896 Год назад

Hi, Cassie are their any APIs available for Bard/Lambda?

@LaHoraMaker Год назад

Waiting for Bard availability outside US, but it looks promising!

@onspot786 Год назад

Also available in the UK and through VPN to other countries

@I_Lemaire Год назад

The thing is: Why/how is GPT so good at code?

@rishishahbeatz Год назад

Bard is so good

@queryingdata Год назад

Thanks for the introduction to Bard. We are lucky to be living through such a great technological revolution. Bard looks very promising. However, I disagree with a lot of comments here on Bard having its own opinions. I believe that these AI models should be like the Media - giving out all the information and answer our questions in a human like way but without having their own opinions. There will be many young children who will have access to these AI tools sooner than later. And an opinionated AI may prevent children and even a lot of youngsters from developing their own critical thinking skills. What do you think?

@A.N.A8 Год назад

I wholeheartedly agree with you, I will make own my opinion, not the Ai !

@Sagar13iffy Год назад

This is clearly an ad for Google Bard and how it is opinionated (which can be a good or a bad thing) rather than objectively providing information like chatGPT. While there's nothing wrong with advertising it, you trying to camouflage it like just another video of comparing the two AI bots, is something that bothers me. I am still your fan and will continue watching your videos though but now with a bit of skepticism.

@waveseeker Год назад

Eventually, you'll figure out she's a charlatan, as I've worked with many underwhelming Googlers like her.

@loststranger5623 Год назад

A bot with an opinion is BS. I understand you work for Google so have to advertise for them though 😅

@zooni43 Год назад

Not available in my country Pakistan😢. Why these barriers ?

@axel_r_ Год назад

To avoid crashing on day 1. Many countries does not have access yet. Not available here in Norway either.

@andersonc137 Год назад

Doesn't work with a VPN?

@zooni43 Год назад

@@axel_r_ many thanks

@max72122 Год назад

ChatGPT like to play safe.

@kurisu-makise Год назад

they try to please everybody but actually please nobody

@ousmaneyadely Год назад

As I see in your presentation we still need a middle point between the two. Bard relies on our proclivity to bias and our inclination towards antromorphization of everything (marketing psychology or just your choice to make a Unique Selling Point on AI), while GPT-4 focuses on facts. It does seem to me that with Bard you just use human psychology to sell more while GPT does not pretend to show it knows what it doesn't. I always go for truth.

@geekyprogrammer4831 Год назад

not available here yet :/

@humaninside Год назад

To give GPT4 a character, you'll have to prompt it to play as a character. Otherwise it'll act as a neutral chatbot character OpenAi already gave it behind the scenes. The difference can be seen if you play with the API, where you'll have to choose the character anyway.

@Thefare1234 Год назад

I have tried it, but it doesn't work. It keeps repeating “As an AI....” followed by a neutral encyclopedic answer. The only difference is that at the end it asks about your opinion.

@beltranypuga Год назад

Biased video. In any event, Bard is still vaporware, when it's fully available, and we get a chance to see if it's really worth it, then we can judge. Meanwhile, ChatGPT is the way to go.

@jamesr141 Год назад

"Hume is nice to read" NO! Sorry but you are just being pretentious. His writing is impenetrable and borderline turgid. I wrote my Philosophy Master's thesis on Hume, so I have read all his works. He is by NO means a wit. Even the quote offered up by bard was not witty at all. If you had said Wilde or Voltaire, Saki or La Rochefoucauld then I would understand why you would "have a crush as a kid." Come on, we are all guilty of pretentiousness from time to time. But you did not read Hume as a kid and get a crush on him for his "wit". What are some of the witty phrases that stuck in your mind? Haha exactly. Hume himself resisted pretense where he saw it. Honor your crush and do the same.

@kozyrkov Год назад

By kid I mean teen. I really enjoy Hume, I just do. But I do prefer WIlde for wit, you're right that he's in another league. Voltaire is not not my cup of tea, at least in translation. As for the last two, I have yet to read them. Thanks for pointing me in their direction.

@waveseeker Год назад

Bard is just another annoying Chatbot, just like the useless Google Assistant.

@MDz375 Год назад

Opinionated Language Model is a nonsense. Because this is like asking an opinion of a person who struggles schizophrenia. This is the opinion of a simulacra that serves as an agent of language model. So if you give an instruction to the model to behave like "Hume supporter" you will get a different answer.