The capabilities of multimodal AI | Gemini Demo

For the purposes of this demo, latency has been reduced and Gemini outputs have been shortened for brevity.

0:00 Intro
0:19 Multimodal Dialogue
1:32 Multilinguality
2:04 Game Creation
2:31 Visual Puzzles
3:17 Making Connections
3:39 Image & Text Generation
4:06 Logic & Spatial Reasoning
4:55 Translating Visuals
5:27 Cultural Understanding

[dpsdps]

Absolutely mindblowing. The amount of understanding the model exhibits here is way way beyond anything else.

[degenplanet]

Just one problem: the video isn’t real. “We created the demo by capturing footage in order to test Gemini’s capabilities on a wide range of challenges. Then we prompted Gemini using still image frames from the footage, and prompting via text.” (Parmy Olsen at Bloomberg was the first to report the discrepancy.)

[joshuaryde]

Google has admitted in a blog post that this video isn’t accurate- the AI “was not responding to the voice or video at all”, but in fact had written prompts to respond to and still images rather than the live drawing/conversation which are not shown in the video.

[ChrisBrooksbank]

Im glad to see Google back in the game, this looks next level.

[mattador]

Pretty disappointing to find out that Google faked these real time live video conversational interactions.

[sakushi]

OPENAI DID IT!!
THEY DID WHAT GOOGLE COULD NOT

[TimeBucks]

The real-time element is by far the most impressive.

[JakeHaugen]

Absolutely next level stuff. The temporal inference was amazing. I was most impressed by it's ability to remember where the ball was and follow it. Seems well versed. What a time to be alive!!!

[familymultiplayergames]

When did Google lose their way and think it’s ok to fake videos to raise stock prices.

[Wrfire]

Unfortunately, what you see is not at all what happened. The AI does not actually reply to the person but to a script and pictures containing sometimes more information than we are shown here

[EC]

I got shocked and mind blown seeing how smart Gemini is in this video alone, it's kinda scary how advanced and smart it is, what is it? a primitive initial AGI? just WOW

[Yassine-tmtj]

What a journey we’re about to embark on!

[SkyLee]

GPT4o literally can do what have demo in this video 😅

[jiminyc]

"Google admits AI viral video was edited to look better", I just read the article on BBC website where Google explains the video is not real time but edited. Much ado about nothing ...

[PressForNick]

This is mind-blowing! Thanks for giving us a sneak peek into the incredible progress happening in the world of tech, creativity, and communication. This has the potential to be at the heart of everything we do.

[wyssli]

according to bloomberg: "In reality, the demo also wasn’t carried out in real time or in voice. When asked about the video by Bloomberg Opinion, a Google spokesperson said it was made by “using still image frames from the footage, and prompting via text, ” and they pointed to a site showing how others could interact with Gemini with photos of their hands, or of drawings or other objects. In other words, the voice in the demo was reading out human-made prompts they’d made to Gemini, and showing them still images. That’s quite different from what Google seemed to be suggesting: that a person could have a smooth voice conversation with Gemini as it watched and responded in real time to the world around it."

Wow Google you must be desperate...

[romanatorx]

A shame that it is not 100% real and google admitted to editing it to make it look better... Fake advertising?

[SoloPirate]

Tasteful touch at the end with the constellation drawing. So far Gemini is living up to the hype. Looking forward to using it come 2024.

[greatbritishmale]

They’ve edited the video guys to make it look better. The AI was not responding to the live actions of this guy, it was responding to still images and text. Very strange to act like its AI is capable of this.

[Inter-Dimensions_Studios]

I have always thought Google has the best chance to take generative A.I. to a super level.