Personalized AI for you | Gemini

Google’s newest and most capable AI model - Gemini.
Join Google Research Engineering Director Palash Nandy as he showcases Gemini’s advanced reasoning and coding abilities, all while exploring ideas for a birthday party.
The model understands his intent to plan, design and build visually rich interactive experiences that go beyond chat interfaces and best display different types of information.
Check out more Gemini demos: goo.gle/4164rNO
Find out more about Gemini: deepmind.google/gemini
Read the blog post: goo.gle/3uRyug7
Subscribe to our Channel: / google
Tweet with us on X: / google
Follow us on Instagram: / google
Join us on Facebook: / google

[@digitalapplied]

This is the next phase of generative ai, real time interfaces. Such a great demonstration, Google is so back.

[@_mobasshir_]

It won't be real time

[@mijaelviricocheaparra7474]

@@_mobasshir_ It will

[@_mobasshir_]

@@mijaelviricocheaparra7474 Not yet. But yes it will be

[@skilarcarbo6002]

@@mijaelviricocheaparra7474 you get your feings you get the money from the

[@NiceLadyCincy]

I've had an AI Chip in my Pixel phone for the past two years. Google just didn't market it that way...but Magic Eraser? Reading numbers in photos, pulling text from images? Of course, it was AI. Pixel Pro 8 is even more AI savvy.

[@HelloCodepur]

Power of Flutter + AI 🔥🔥

[@tristanwegner]

Wow. The intergrated modalities and the interactivity with parts of the generated content is a big step up from GPT4

[@randomguypolololas]

The most impressive thing is it's generating UI on the fly! That's insane!

[@shukrantpatil]

That day is quite colder where apps become nonexistent

[@Gymfitnes878]

@@randomguypolololas o

[@raulvaldezgranados7140]

😊

[@RakeshKumar-je4jz]

TT

[@yaphanyee]

This is just next level, it’s much beyond intelligence, the reasoning, structured and flows are so precise and natural, not to mention how creative and unpredictable the way it expresses the idea, a big WOW

[@Google]

Bard's here to take things to the next level!

[@advaitbhore]

As a UI Designer, this is such a gr8 step in the right direction... if chatgpt needs to learn anything from the gemini release, it's this.

[@entertainmentyoutube3606]

Are you not worried about losing your career and job?

[@advaitbhore]

@@entertainmentyoutube3606 UI/UX is not just what many of us sometimes assume it is limited to...in my opinion, it spans much further than what is being demonstrated here. It still has way to go. Until then I think I'm safe, feeding the AI and training it ig lol. Eventually, yes I think just about everything will become obsolete and be considered either a waste of time or too much hard work like a thing of the past due to AI (that is when they get them to be as trustworthy and efficient as calculators). Something like the difference between hand-made and factory-made. Invention of trends and what humans like, can still only be decided by humans. That is until we unlock the algorithm for creativity.

[@entertainmentyoutube3606]

@advaitbhore I'm a ux ui too, and I dont want to be jobless, copywriters and translators are already suffering, how they will pay next month bills? There are serious and real 5h8ngs happening thanks to this

[@advaitbhore]

@@entertainmentyoutube3606 Even if we go against and get them to shut the advancement officially, so that people stop the legal use and development completely, we just slightly slow down the process, not much. It's too late to put the genie back in the bottle. The advancement will continue despite the backlash, in open source or individual development. It is too useful a tool to neglect and ignore for the people. Our best bet is to focus on what still remains irreplaceable. Keep shifting to skills that have not been developed in the field by the AI yet. Keep climbing the skill ladder, one rung at a time. The top, to where the ladder leads is hidden by the dense fog of the future. It may be very close, it may be little far, but when the top level is here, everyone will reach in an instant once they've built a direct AI lift to the floor. Then we'll remain the only ones who do it manually. Being shallow and going against it will avail us possibly some headlines, and sympathy, and if we try hard enough to try to shut it down, just some consolation prize which will have an expiry date.

[@btm1]

man you're such a bot

[@mkbrk]

Now THAT is interesting! I've only watched a few of the Gemini demos so far, but this one is comfortably in the lead.

[@Google]

So glad you're enjoying our demos!

[@darshank8748]

@@Googleyou’ve made such a great job!

[@JOHN.Z999]

I am impressed with the new integration and way of communicating with humans, it's lively and very dynamic. I hope this becomes the standard for the new Google going forward. I found it extremely efficient and relevant!!!😮😮😮😮

[@ShpanMan]

Awesome, this is the next step in AI - more than a simple single question and answer format. It also plays to the strengths of Google and will integrate very well in their products.

[@Google]

With Google extensions, Bard with Gemini Pro can integrate into your apps and services! ✨

[@JoyStrother-vn7ch]

It would be amazing to have a Google HQ in the Louisville Metro area of KY! The jobs, revenue, and economy burst that it would bring I can't wrap my head around it! Just WOW! And the architecture in the area is beautiful, we have some really amazing historical buildings locally right in the river that would be perfect and could use a touch of love NOT MODERNIZED AND REMODELED, but restoration and renovations would just be so lovely!

[@cedfunches6039]

This is a great step into doing AI the right way. You have to have varied intent be at the forefront of the returned data. Who, What and Why inform the How and When

[@matthewdignam7381]

Google was waiting to strike when we all thought openai were ahead of the curb, turns out it was google all along, amazing stuff

[@caelen_c]

This is so cool! Like Bing Chat but 10x better

[@NHCH]

This is mental, great job guys!

[@obladaobladi4020]

If these generated app cards support forwarding, then in the future, apps can be directly forwarded and used through chatting, which will change the way apps are distributed!

[@lawrence9239]

THIS IS JUST IMPRESSIVE! Dynamic UI WOW

[@terrytan1827]

This demo looks amazing. Google step up the game bringing AI to the consumers. While OpenAI is more towards developers and research. I do think GPT-4 can achieve similar feat if that's where their focus. But I hope they focus on making GPT-5 amazing.

[@Agent.J]

they should not focus on consumer product. let google handle that...gpt should be a profession high level tool rather than dumbed down consumer product

[@siddish]

Crazzyyy, just want these adhoc interfaces for data science

[@frankieS2727]

Next generation of UX/UI for AGI! 🙌 🎨

[@okubowisdom]

Multimodality. Spectacular! Internationalization and Localization for Customization and Personalization is an actual competitive advantage when comparing the different existing models and I hope Google / Deepmind recognizes this on time.

[@stacksmasherninja7266]

This is really impressive

[@iandanforth]

Is the working time shown in the video an accurate representation of how long it takes Gemini to complete the first interface generation and population?

[@hebiasohj3802]

@iandandforth GPT4 can certainly do this and usually may requiring multiple prompts. I think the advancement here is the contextual multi modality in the response in addition to what appears to me to better reasoning policies.. so yes the details matter but in fact they really shine in this case

[@churroholics8689]

​@@hebiasohj3802I can't find anywhere any evidence that GPT-4 can create real time UIs to answer user queries for day to day tasks, only UIs made specifically for a "create a UI for.." query.

[@dontwannabefound]

Wow, this is actually interesting

[@ReedoTV]

If the real version is 1% as good as this demo then it will still be incredible

[@ceefar10]

Incredibly innovative, cant wait to try it out

[@Google]

Innovation awaits when you're using Bard with Gemini Pro!

[@LacksonMunthali]

Is that UI in flutter? I can see the debug banner on the top right corner.

[@emilrueh]

He specifically says it is at one point.

[@LacksonMunthali]

Commented before he mentioned it. It's amazing how it's generating and rendering the UI. Looks like some server side rendering.

[@education.online_frevryone]

When and where is this coming?! Excited 🎉🎉

[@RealChristopherRobin]

As a natural language model, this is great.

[@0269_m]

😂😂😂

[@phen-themoogle7651]

ChristopherRobin, nostalgic , nice! haha

[@orionsbelt29]

Did it pull those images on the cupcake or make them?

[@hernanduarte3406]

🤯iMPRESIONANTE!!!!

[@SupItsJ]

Is this Notion that the content is generated inside of?

[@xvll-l1589]

Where do I access this?

[@FacadeMan]

Has anyone been able to test this feature yet? I would love to get my hands on this and get a feel for a truly futuristic development process that I could only dream of.

[@gwendolynreynolds6445]

but where did you get your sweater? it's lit.

[@tayo7758]

This is really cool

[@Projectzon]

Amazing Experience

[@ultrasaiyan4283]

Only hope it will be like that in practice.

[@AribZeeshan]

As someone who is just getting started into CS ...... What do i learn ... it seems that every development I try to learn, AI will do it better ........What do i do

[@MuratKarakus-to4tr]

welcome to matrix...

[@gizemmalcok]

woww this amazing! 🌟

[@manulkhatri9466]

When do we get to play with it

[@weirdsatya]

Flutter is used for ui 🎉

[@wy2528]

Can’t wait to use Gemini

[@Google]

Check out Bard with Gemini Pro at bard.google.com to be a part of the magic!

[@vaishnavnegi9640]

Its multimodal, its generative, its explainable. Checks all the important boxes.

[@mayaramirez9244]

This is so dope

[@apurvachakke3349]

When are we getting this feature? It is not currently available in Gemini

[@mikeg1368]

Where is it? Or when will it be out?

[@phen-themoogle7651]

Next week for Pro version, next month for Ultimate version.

[@xantari79]

Where is this Bespoke UI interface that is being demoed in this video?

[@Bangladesh689]

Is it live to test?

[@Rolyataylor2]

I would love to see the entire Android Experience replaced with AI runtime generation. Video and images with interactive buttons and animations. It could learn over time and the UI could be tailored to each individual user and each individual scenerio.

[@JBoy340a]

Those multiple modalities make text only look so 2022.

[@yijinghuang6816]

Is any link can lead to try this Flutter app online?

[@AnishUtthiranNC]

This is awesome stuff!

[@Google]

Glad you’re loving the game-changing capabilities ✨

[@entertainmentyoutube3606]

​@Google are you going to take my career and job? Im a ui ux designer, what im supposed to do?

[@sarahsmind]

Wow. Just wow.

[@TheDaveau]

What tool is being used here? Doesn't look like Bard. The "Data Collector" interface looks incredibly interesting - I'd be keen to try this out.

[@h.c4898]

No. Bard can't do that. It's the "Bespoke" prompt. Probably part of Gemini.

[@skierpage]

I think this is just the text interface to the Gemini UltraSuperMega model, but you can see there's a DEBUG badge in some of the screenshots that gets Gemini to show its underlying JSON for the different steps it takes in building the dynamic UI to answer the prompt. Other LLMs output HTML so they can show tables and images and figures, this is building an interactive "app" on-the-fly. Amazing!

[@wrightvcx2249]

I asked Bard about Gemini Pro, and it redirected me to a cryptocurrency wallet.

[@jasonnugent963]

I'm sure this was done for presentation reasons,. but I'm surprised it did not ask what age she was,. or what your location is (to predict weather for an outside party),. or how many people would attend or how much money you want to spend. Lots of variables !.. ;P

[@skierpage]

Nor did it create a "You are a top children's party planner. Collaboratively review these children's party ideas with the client and help them choose the best one." AI persona.

[@JohnBr0]

Wild. 🤯

[@wayne7936]

Bard, what sweater is Palash wearing? I need that fit.

[@poplars71]

Billy Reid :)

[@KolTregaskes]

Looks good but it says "Bespoke UI". Can we see examples from Bard please?

[@skierpage]

Yes, the whole point is this generates a bespoke UI to present multiple answers to the user's request. I suspect it's all taking place in an HTML window generated by the Flutter toolkit. Bard can't do any of this.

[@History_Mystery_Crime]

Google is officially back to the Game💥💥💥

[@ok373737]

Impressive.

[@vinit095]

On the right top it says 'Data Collector'.

[@Astro.004]

Can it generate ai images like gpt 4

[@sleepi5550]

That’s exactly what it did in this video

[@JazevoAudiosurf]

it's cool gpt-4 has a competitor, but it can do all of that too. so it really depends on the details. excited for AlphaCode 2 demos

[@churroholics8689]

I don't think it can. I can't find anywhere online any evidence that GPT-4 can create real time UIs to answer user queries for day to day tasks, only UIs made specifically for a "create a UI for.." query.

[@JazevoAudiosurf]

@@churroholics8689 Gemini can't do that either. there is ongoing controversy about it. if you know transformers, it's immediately clear they had some bigger engineering efforts to make this work. there have been people showing that gpt-4 can achieve the same thing when engineered that way

[@samario_torres]

amazing

[@pradhyumnakm3574]

Awesome, but how to use Bespoke UI

[@motahrandianlall819]

please write an article on the mauritius financial center

[@supreetsahu1964]

Hints of AGI

[@RishiKumar-nz1tp]

What are the chances that such tools/technologies will not be used for malicious intent?

[@h.c4898]

You can't control that. A fork can be used for utility or harm. It depends who holds it. It wouldn't be fair to put the blame on the tool. Otherwise, all knives will have to be banned. Then how I can chop my steak into small pieces so I can swallow it and nourish myself?

[@RishiKumar-nz1tp]

@@h.c4898 Exactly! I think that's a good observation. But still knowingly with such power tools there will be chances of high consequences as well, we are excited on creation of such tools. So there's a no everlasting solution to this?

[@shadygamererfan377]

We are in the future! Woohoo!

[@Google]

Happy to be here 😊

[@jorgeavila3385]

I always knew Google was going to come with something big. That pile of cash is certainly very useful. I am not surprised they let open AI to go fist. They were up to something, and this is that something.

[@h.c4898]

Sundar said he was always skeptical of AI. The reason why they've kept it under a vault until openAI showed up and came up with the big guns. Now It's waking up.

[@attilavass6935]

WHICH Gemini is capable of this?

[@AmineBoulaajaj]

Does this mean more Google layoffs?

[@ElenaDotNET]

Flutter was there

[@qf0421]

I don't need a UI. I need a more stable version of GPT4 with much longer output.

[@Emiwryy]

What about the other AI, Bard? Is Google dropping it or what?

[@HarpaAI]

🎯 Key Takeaways for quick navigation: 00:00 🌟 *Introduction to Gemini's capabilities* - Gemini is a multimodal AI model designed to understand and reason about user intent. - It can generate bespoke user experiences beyond text responses. 00:30 🎨 *Creating a bespoke interface* - Gemini generates a visually rich, interactive interface without any coding. - It makes reasoning decisions from broad to high resolution. 01:00 📝 *Writing a product requirement document (PRD)* - Gemini creates a PRD based on user input and requirements. - It plans the functionality of the user experience. 01:57 💻 *Designing the user experience* - Gemini designs a user journey, including list and detail layouts. - It generates Flutter code and retrieves data to render the experience. 02:28 🍰 *Customizing the experience* - Users can interact with the interface to get specific information. - Gemini generates new UIs based on user requests and preferences. Made with HARPA AI

[@aebisdecunter]

Savage

[@choongching]

Can we try Harpa?

[@KentHeckel]

Hmm

[@ozioma]

bespoke interfaces 🤯

[@FrozzenFreak]

AI will reason in JSON. Let that sink in.

[@SonuBadmash-ll8dj]

Iyk. 😮yK!

[@MunirNawaz]

@Google, Awesome

[@humble_integrity]

gemini can go beyond chat interfaces!!! except hte first thing he does is use a chat interface

[@Ismail-hq8cq]

Wow the frontend they using Flutter, Look right in the corner

[@kennedymutugi3740]

A good Generative AI .

[@Pixellof]

Wow, I hope Google will not add censorship into this

[@sleekism]

This is mad!

[@youtuber99]

WOW

[@Danuxsy]

We need personalized models that understand who we are as individuals. One way to achieve this is to have the AI literally be part of your life through microphones and cameras so it can hear what you hear and see what you see 24/7 and use the massive data it generates to build a picture of who you actually are. This is cute but you aren't thinking big enough, why are you stuck in the digital realm? it's supposed to be multimodal, it should be ingrained in our physical world too, become a part of who we are.

[@StevenAkinyemi]

And how is current AI models going to handle the massive data from the 24/7 input streams? You think their compute is unlimited?

[@Danuxsy]

LLMs have no trouble compressing the internet into weights the size of a terabyte. The data you generate throughout your life is nothing compared to the size of the internet already. This is a none issue, the issue I believe might be the processing of the data in real-time, especially for vision because of the cost.@@StevenAkinyemi You only need to store the data temporarily to process it and then it become part of the neural net in the weights.

[@skierpage]

Go re-watch the movie "Her".

[@Danuxsy]

I've seen it twice, i think my idea might originally have come from there, at least in part but I do like to take some credit of my own too so I can feel special and intelligent.@@skierpage

[@shakeerp3374]

1:44

[@zagyex]

So we wont need websites in the future, just this thing. But then... where does it get the data? Where did it get the recipe from? Serious copyright questions here.

[@chikken007]

Is it just me or does this voice sound kind of "smoothed" by some audio filter /AI?

[@shakeerp3374]

3:33

[@shakeerp3374]

0:20

[@cesaradrianarreola7266]

Welcome to the Gemini era. "Another suitable name besides Gemini could be 'I AM'."

[@USONOFAV]

Demo looks good but I'll be over here, not holding my breath. Remember their big reveal of Google Duplex? Still in the realm of fantasy, last I checked.

[@cooltop101]

You must not have checked recently than. Duplex has been launched for awhile (couple years at least) and I have used it myself a few times. They just never officially announced it was released, but it's there.

[@vazules]

Very cool. The Next step of ChatGPT or Bard. Integrate this Knowledge Graph Technique in bard.

[@stinger4712]

I just asked bard if it had Gemini and it said no

[@abinashmohanty]

This looks like the earlier version of chatbot interface aka boring Interface 🤷🏼‍♂️👀

[@SantusaSaico-ir5ux]

En español

[@tuckerbugeater]

i will never allow ads