This competition is about putting the consumers out of work. The rich and powerful who can afford GPU's and robot armies will rule all resources of earth. They won't need human consumers, workers or voters. They will be self reliant with robots. Childish people talk about Universal Base Income, which means free equal money to all, or equal stakes on all resources of earth and all produces of robots. If history is any lesson, how likely is that.
@@TruthCorps It is very good! It still has bugs (it's experimental), but it can see your computer screen or your webcam and the voice is very natural like ChatGPT's AVM. Fast, pretty smart, lots of other features.
Guys!!! We all have to remember gemini has a history of a half baked abilities from Google. Let see it our selves first before we get excited. Action always speaks louder than words.
Gemini is built into android and replaces google assistant. All other AI agents will replace googles main business model if they don't make gemini work. It's one thing to have a 3rd party android app AI agent. Its another thing to have that AI agent integrated into android that can get live results from the internet, for free. Not only can it compete with openAI's userbase as a webapp, your phone will be the AI one day.
The great google paradox. Deepmind has some of the greatest advancements in the field, consistently. Gemini has some of the baddest fumbles in the product space.
As an italian I can sat that I've never heard a quality so high of the Text-to-speech in italian. Seriously... It would be hard to understand that is AI made and not a recording from an italian voice actor. Astonishing!
Yeah, we're "solving" things one by one, which is how they refer to reaching essentially perfection in various skills / tasks. We're very much on the cusp of solving all worldwide languages and translation...
Wasn't this the same company who showed us an AI assistant 5 years ago, that would schedule appointments, call your barber for you etc? I believe it once it drops. Google is known for over exaggeration.
🇧🇷🇧🇷🇧🇷🇧🇷👏🏻 Astra is excellent absolutely top tier! Honestly, it’s the kind of innovation OpenAI has been holding back from releasing to the public, but Google went ahead and did it. Congratulations, Google!
AI sure is advancing so much faster. Especially during these last few months in 2024. What do others think will be the big advancements in AI during 2025?
I'd really love it if the agent could help me create my own game! That would be so cool, it can already to a degree, helping me with pointers or even righting out the code I need but when I tried it last it did make a lot of mistakes but I can imagine at some point it will just... do it all for you
Have no idea what that means in actual practical terms, other than just wild scare mongering. The reality is that in about 10 years there won't be the need for humans to "do" anything to have food, clothing, housing and healthcare, since we're headed for the End of Human Labor there or thereafter. You could literally make your same statement about 100 technologies today: "can't operate on you without an X-Ray", "can't get to work without cars", "can't work from home without VPN", etc. etc.
Ugg vocal fry is the worst. It's prevalent among one certain side, and those from areas such as California. Considering that Google, and many other large tech companies, are based out of there. It's not surprising to hear vocal fry in AI audio voice models.
This is so dystopian. Not only that the ai voices now come with vocal fry. But also the fact that this is a way to outsource your agency to ai - and that means that you lock in to the platform of one specific giant corporation. This corporation will shape not just how you interact with the world, no it will in fact *shape your world*. Will you subscribe to the Google life? Or will you prefer a life spent in META?
we outsource everything to ai and slowly die out as population decreases and all that's left of us will be these agentic systems that play out our lives forever
if people are still jaded about google....this should quiet them down. This is truly amazing...no hint of vaporware... just speed and agi capabilities. The people still talking about "stochastic parrots" or griping about hallucinations are truly going to be steamrolled by all the people that actually follow these videos.
AI models are one thing. To implement these into a useful environment is another story. Google is here clearly the leader of the pack. Considering the vast array of APIs to which you can connect Gemini Google is creating huge opportunities for creative developers.
So my test for creativity is poetry. Older GPT couldn’t write without couplets but now can create excellent poems in different structures. So far it can’t resist couplets and when writing a song it plagiarized jingle bells.
i personally used gemini 2.0 flash live stream thing and it was pretty bad. yes it could analyze my screen pretty well but the problem is it had 0 memory. so for example i wanted it to guide me through scripting and i told it what i want to achieve then a response later it says like 'what now' in different words. but what do you mean what now. like he forgot the entire mission
The Mariner demos are somewhat useless in terms of demonstrating something that AIs are better than humans. The only value was truly showing automation of actions… nonetheless the AI could have done all of those in the background
@@TheAiGrid isnt it peculiar that in the first example of the sculpture from every possibly random sculptures in the world they pick one by Eva ROTHSCHILD?!
this is what u called improvement dont know whats wrong with open Ai i am so upset i just stop talking about open ai this is the type of improvement i wanted from open ai i think they really hit the wall
big ben is not a landmark -it is a bell the big ben clock tower is a land mark -as big ben is not visible it is in the tower and is a BELL -the hour bell -so not a landmark -so the progammer got it wrong -??
Is this all actually available or will it be available any time soon ? Google are famous for showing things that never see the light of day or that they quickly kill. They're great at research but crap at marketing or following things through to real world completion. They like playing like kids but aren't real adults
The problem with all these models that are by big Tech is there great until you actually wanna do something powerful with them. It’s like having a chainsaw that doesn’t work when you don’t use it on a tree, but you need to use the chainsaw maybe two cut up a car your friend is stuck in, but the chainsaw doesn’t allow you to do it. That’s what these these models are like.Anytime you try to do something that would give you something incredible and out of the box it tells you that’s wrong that’s politically incorrect and so even though the models are very advanced what you’re getting is a water down piss poor fucking British politically correct crap that does shit.
Once upon a time, I would have chosen Google before any other company - especially those aligned with Microsoft. But ever since the Google AI-fake-video with the original Gemini, my only question nowadays is: how much is gonna turn out to be fake this time? I guess, probably about 50% at least.
Google is back for real this time! Sure, they probably half-faked that outdoor stuff but their latest LLM beats OpenAI's "thinking" models in math with much faster response and this truly multimodal stuff seems to be much better than the stupid DallE -3
Before I watch this video: if Google didn't actually "shocked the entire industry" with Gemini 2, I will ask RUclips to NOT recommend this channel. We are tired of garbage videos presented as fantastic. Gemini 2 is probably the advancement anyone expected. I wish I'm wrong although statistically I'm on the right assumption.
I've tried to use the unpaid version of Gemini. It doesn't follow instructions very well. I bet the translator is half baked as well. Also it isn't true AI, when it doesn't understand its input and output.
Dont use gemini for a simple reason: can not create a simple calendar entry or convert lets say 100chf to euro and always say it cant get realtime access to conversion rates
Didn't OpenAI already show these demos like 6 months ago, and say it'd be out within a month? I'm still waiting for my video chat buddy that can narrate ducks in a pond to a blind person or hail a cab for said blind person.
This is the time when AI started to become dangerous. It takes humanity centuries to take back control after they gave it to AI for common day tasks. Letting machines do their thinking and making choices for them, becoming automatons to their base instincts, not realising their true potential as spiritual beings in a material projection.
After reading this I nearly went onto some spiritual guru's youtube channel and warned them how AI powered robots will hose them down with bleach, but then I realised I wouldn't gain anything from doing that. I mean, I'm obviously not into that stuff so why would I waste my time watching their videos, unless I derived some kind of flaccid pleasure in wafting my opinions in their direction, but who would do that?
@@Dose_0x0 You too can gain superpowers by dry fasting for 40 days or 960 hours. Watch as you become aware that the observer within affects reality and reveals that causation dually by natural phenomena and conscious thought. Give to those with less and watch how you unexpectedly receive from those with more. As if some kind of hidden forces and laws of both thought and energy exist. Just as one cannot see a magnetic field with the naked eye, and air is often invisible, you might find that the inner eye is the master of the realm of the unseen. Now, about that bleach…
I don't want my video game to be overused nostalgia. These humans will relentlessly recycle pale imitations, that do not have the contextual essence of the original experience. Those same humans have been trapping entire genres in overreliance, while preventing genuine innovation.
Only Microsoft can compete with Google. Having good AI models is one thing. The other thing is to let them do useful things. Neither OpenAI nor X have an environment where you can implement AI to do useful things. Anthropic can do it with Amazon, that is also working. How the relationship between Microsoft and OpenAI will work out we have to see. The only frontier model developer who has it all under one roof is Google. That gives Google many massive advantages although we are seeing only the first baby steps yet ...
Finally we have Agents that could clear our search history with efficiency.
Sorry, I can’t do that Dave😂
Try incognito mode first.
Top Comment for sure 😂
...and immediately log all of it, storing it and sharing it with the Google overlords and anyone who might want it.
You wish 😅
Competition is great for us consumers
Customers? Unemployed do you mean?
This competition is about putting the consumers out of work. The rich and powerful who can afford GPU's and robot armies will rule all resources of earth. They won't need human consumers, workers or voters. They will be self reliant with robots. Childish people talk about Universal Base Income, which means free equal money to all, or equal stakes on all resources of earth and all produces of robots. If history is any lesson, how likely is that.
@@ajarivas72 lol
Not in case of AI lol
This is the first demo I watched from google that really nails it.
The voice interaction has improved greatly it seems.
I was not impressed compared to chat gpt
2.0 Flash looks incredible. Can't wait to get home and play with AI Studio.
Any results? I did not have a good experience with Gemini 2
@@TruthCorps It is very good! It still has bugs (it's experimental), but it can see your computer screen or your webcam and the voice is very natural like ChatGPT's AVM. Fast, pretty smart, lots of other features.
Guys!!! We all have to remember gemini has a history of a half baked abilities from Google. Let see it our selves first before we get excited. Action always speaks louder than words.
Gemini is built into android and replaces google assistant. All other AI agents will replace googles main business model if they don't make gemini work. It's one thing to have a 3rd party android app AI agent. Its another thing to have that AI agent integrated into android that can get live results from the internet, for free. Not only can it compete with openAI's userbase as a webapp, your phone will be the AI one day.
The great google paradox. Deepmind has some of the greatest advancements in the field, consistently. Gemini has some of the baddest fumbles in the product space.
Most of this is available to either public or developers right now
So we should all try to remember how terrible the initial automobiles were before buying a car today?
this is already out for testing
Seeing how horrible Gemini 1 was I’m not convinced this ad isn’t purely science fiction
the ad is not too far from the real model though
As an italian I can sat that I've never heard a quality so high of the Text-to-speech in italian. Seriously... It would be hard to understand that is AI made and not a recording from an italian voice actor. Astonishing!
Yeah, we're "solving" things one by one, which is how they refer to reaching essentially perfection in various skills / tasks. We're very much on the cusp of solving all worldwide languages and translation...
Wasn't this the same company who showed us an AI assistant 5 years ago, that would schedule appointments, call your barber for you etc?
I believe it once it drops. Google is known for over exaggeration.
I mean they did deliver on that. Call Screening on the pixels? Remember?
@NadimShaikh-qv7zj What does screening spam calls have to do with anything OP mentioned?
tbf the google assistant can actually do this its on select business though they have to opt in
Al Agents will be more practical and pleasant than a marriage partner.
🇧🇷🇧🇷🇧🇷🇧🇷👏🏻 Astra is excellent absolutely top tier! Honestly, it’s the kind of innovation OpenAI has been holding back from releasing to the public, but Google went ahead and did it. Congratulations, Google!
How can you be happy about this? So dangerous, inevitable but foolish.
Google paying you to write this or..?
AI sure is advancing so much faster. Especially during these last few months in 2024.
What do others think will be the big advancements in AI during 2025?
I'd really love it if the agent could help me create my own game! That would be so cool, it can already to a degree, helping me with pointers or even righting out the code I need but when I tried it last it did make a lot of mistakes but I can imagine at some point it will just... do it all for you
Generating video titles with no click-baity language!
@@svenhoekhow will bro go viral
Generating high quality videos for less than $200 a month, with sound so I can make an anime show out of the script I have written
@@blutacckk can’t wait to make a video game with ai
10 years from now, nobody will be able to do anything without asking AI, this is scary.. They want you dependant
Have no idea what that means in actual practical terms, other than just wild scare mongering. The reality is that in about 10 years there won't be the need for humans to "do" anything to have food, clothing, housing and healthcare, since we're headed for the End of Human Labor there or thereafter.
You could literally make your same statement about 100 technologies today: "can't operate on you without an X-Ray", "can't get to work without cars", "can't work from home without VPN", etc. etc.
Her begins.
I'm SHOCKED !
google doesnt disappoint
Wow. This new "door code memory" feature finally caught up with the technology of a Sharpie and my arm.
Guess you need to learn to print very very small. It has unlimited memory.
Think of Google training such AI's on their new in house quantum chip.
Unbeatable AI greatness on way towards Google!
Great !! Its like having mom do all my homework 😊😊
“Do not mistake mimics for Agents “
Assume the agent is standing right behind and in front of you without being compelled to say anything
Wow, this is impressive! Google really stepped up their AI game.!!
Ok, now I want all this as part of my Home Assistant. And first, the native audio !
FFS even the AIs are using vocal fry now. Make it stop.
Ugg vocal fry is the worst. It's prevalent among one certain side, and those from areas such as California.
Considering that Google, and many other large tech companies, are based out of there. It's not surprising to hear vocal fry in AI audio voice models.
😂
It is so fucking irritating.
This is going to revolutionize traveling internationally. What language barrier?
This is so dystopian. Not only that the ai voices now come with vocal fry. But also the fact that this is a way to outsource your agency to ai - and that means that you lock in to the platform of one specific giant corporation. This corporation will shape not just how you interact with the world, no it will in fact *shape your world*. Will you subscribe to the Google life? Or will you prefer a life spent in META?
we outsource everything to ai and slowly die out as population decreases and all that's left of us will be these agentic systems that play out our lives forever
Yeah because no one today can switch from iPhone to Android, simply impossible.
Stop remembering and thinking we already killing out attention span... Future is bright trust me.💫
This is next level.
And great timing. Gemini 2.0 can be in the OpenAI 12 day countdown.
“Agentic”, not “Generic”…
if people are still jaded about google....this should quiet them down. This is truly amazing...no hint of vaporware... just speed and agi capabilities. The people still talking about "stochastic parrots" or griping about hallucinations are truly going to be steamrolled by all the people that actually follow these videos.
That's impressive - I can see myself using that in my work and personal life.
i'd like my AI to be local, with no spyware please waiter?
No brakes on the surveillance capitalism train
Now that is shocking! Hopefully they'll release early Q1 2025
Awesome use of the collected data from your users and a step above by spending your money without hesitation. Be careful what you wish for.
Opening the car door was pretty cool.
Except that Gemini put a frame around the driver side window that wasn't there previously.
looks like im about to jump ship... im going team google yall
good at talking will see when it launches
21:12 I can proudly say, my responses were used in training that, Google made me do image captcha almost everywhere
With guardrails it’s just a search bar with voice but without guardrails it’s god
“I say 'your civilization' because as soon as we started thinking for you it really became our civilization"
when did it say that?
The fact that google releases almost nothing usable (or is so late to release stuff), this doesnt hit me, even though its so great and perfect.
There is only one more step to complete here and when it happens you know the world ends - removing the need to speak to it
AI models are one thing. To implement these into a useful environment is another story. Google is here clearly the leader of the pack. Considering the vast array of APIs to which you can connect Gemini Google is creating huge opportunities for creative developers.
it has blown my mind . i just build a multilayered code for a first person shooter
So my test for creativity is poetry. Older GPT couldn’t write without couplets but now can create excellent poems in different structures. So far it can’t resist couplets and when writing a song it plagiarized jingle bells.
Like a bad episode of Black Mirror.
Ill probably buy myself a new googlephone etc.
Lmao I'm dying over here! 😂
Same here ...
@@HaraldEngels and Google glass too etc.
I have found the experimental 2.0 Flash version .. but is there also a non-flash 2.0 version too?
Not yet, I don't think.
i personally used gemini 2.0 flash live stream thing and it was pretty bad. yes it could analyze my screen pretty well but the problem is it had 0 memory. so for example i wanted it to guide me through scripting and i told it what i want to achieve then a response later it says like 'what now' in different words. but what do you mean what now. like he forgot the entire mission
Points phone at wife - "Hey Gemini, how do i get her to stop being so annoyed all the time?" ... Gemini crashes.
an ai that helps us once we are about to be isekaid
When she said "it's okay I know interruptions are part of the demo" 💀
Remove the snapshot feature, the ideal condition for VI users would be audio cue the cam which responds without taking a photo
Thanks, and how do we access Gemini 2.0?
Every thumbnail nowadays "shocked the industry!!"
If that voice gets better there going to be unrecognizable…that’s crazy
5 seconds and they can copy anyone's voice now.
This is going to END careers lmao.
and begin some more
The Mariner demos are somewhat useless in terms of demonstrating something that AIs are better than humans. The only value was truly showing automation of actions… nonetheless the AI could have done all of those in the background
Bro I wanna advise you if you could make your lectures into a story then it can be interesting and rest are good. Love from India. Where you from?
The uk!
@@TheAiGrid isnt it peculiar that in the first example of the sculpture from every possibly random sculptures in the world they pick one by Eva ROTHSCHILD?!
Skeptical until I got my hands on it. Google had history of killing off projects after announcing them.
When does it help me get my work done faster so I can have a better work-life balance?
oh don't worry, your work-life balance will soon shift quite dramatically
Who needs work-life balance when you can skip straight to no life after being replaced by AI :)
Never going to happen
@@drinkingrobot2252 short-sighted if you think that..
THE SHOCKER IS BACK!!!
this is what u called improvement dont know whats wrong with open Ai i am so upset i just stop talking about open ai this is the type of improvement i wanted from open ai i think they really hit the wall
@ 7:24 How come at the end of each one of that girls sentences she drops her voice down low it sounds weird… she sounds like an AI.
@15:33 They really need to vary the 'i can do that' quick cos that's gonna get very annoying ,very quick.
I assume they named the AI agent "marijuana" in reference to previous versions of Gemini acting SUPER high.
Dude, the cheese isn't sticking to the pizza, bro. You should like... you like... you should use glue, man.
Brah... did you know, like... did you hear... I just heard from a guy, sharks are like, older than the MOON.
Do you eat rocks? My dude, I'm NOT even joking right now. You HAVE to eat at least one rock every day.
Don't listen to the establishment about your health, fam. Boost your immune system by eating boogers.
Stop even sweating that appendix, homie. You LITERALLY just gotta boil up some mint leaves for that.
Feel like GPT is already capable of such feats
2025 just became more exciting! At this rate we may have AGI before 2030!
big ben is not a landmark -it is a bell the big ben clock tower is a land mark -as big ben is not visible it is in the tower and is a BELL -the hour bell -so not a landmark -so the progammer got it wrong -??
Yeah but why do tec companies have to present it with the same grafics and voice that would be used on the beggining of an apocaliptic movie
Is this all actually available or will it be available any time soon ? Google are famous for showing things that never see the light of day or that they quickly kill. They're great at research but crap at marketing or following things through to real world completion. They like playing like kids but aren't real adults
HEY EVERYBODY! Look! They're bring back the awesome glasses nobody wanted the first time!
Time to check out Meta glasses, rated top AI hardware out today.
Interesting stuff. Mariner/ chrome extension is not open to public yet hey? 🧐 2.0 flash sounds like it could be good for voice agents 🎤🤖
The problem with all these models that are by big Tech is there great until you actually wanna do something powerful with them. It’s like having a chainsaw that doesn’t work when you don’t use it on a tree, but you need to use the chainsaw maybe two cut up a car your friend is stuck in, but the chainsaw doesn’t allow you to do it. That’s what these these models are like.Anytime you try to do something that would give you something incredible and out of the box it tells you that’s wrong that’s politically incorrect and so even though the models are very advanced what you’re getting is a water down piss poor fucking British politically correct crap that does shit.
fr thats scary. On the other side.. its FINALLY something new again
Well if it works better than gemini 1.5 ;)
Once upon a time, I would have chosen Google before any other company - especially those aligned with Microsoft. But ever since the Google AI-fake-video with the original Gemini, my only question nowadays is: how much is gonna turn out to be fake this time? I guess, probably about 50% at least.
Everything shocks the AI industry huh 😂
We are even more cooked
I lost it during Marianna: Web Shopping. Why can’t they find a better presenter ? I feel like he is reading from a script
22:48 😅
Rip human tour guides 😢
Where is GPT-5? 😂😂
4.5 on friday the 20th, just a hunch
Do you have to use a pixel phone.?
Well - there goes thinking.
The Trump vote is evidence there's already none to leave remaining.
Google is back for real this time! Sure, they probably half-faked that outdoor stuff but their latest LLM beats OpenAI's "thinking" models in math with much faster response and this truly multimodal stuff seems to be much better than the stupid DallE -3
this is literally what im doing with my programs xD; these companies are so slow xD
What is the context length?
Before I watch this video: if Google didn't actually "shocked the entire industry" with Gemini 2, I will ask RUclips to NOT recommend this channel. We are tired of garbage videos presented as fantastic.
Gemini 2 is probably the advancement anyone expected.
I wish I'm wrong although statistically I'm on the right assumption.
I've tried to use the unpaid version of Gemini. It doesn't follow instructions very well. I bet the translator is half baked as well. Also it isn't true AI, when it doesn't understand its input and output.
UGH!! I want to build a AI agent now.
it seems hard af not even mentioning the gpu power you need. I doubt people will be able to run local a.i for the foreseeable future.
Cool wanna try that 😮
African American Alan Turing would be proud.
how is this different from AGI?
AGI is expert level at everything, literally EVERYTHING. Nothing about that demo'd here.
Dont use gemini for a simple reason: can not create a simple calendar entry or convert lets say 100chf to euro and always say it cant get realtime access to conversion rates
Jebus effin christ, that's new.
Didn't OpenAI already show these demos like 6 months ago, and say it'd be out within a month? I'm still waiting for my video chat buddy that can narrate ducks in a pond to a blind person or hail a cab for said blind person.
Its live i just tested it
@thymenwestrum7011 where? I went to OpenAi and can't find any video chatting functions.
This is the time when AI started to become dangerous. It takes humanity centuries to take back control after they gave it to AI for common day tasks. Letting machines do their thinking and making choices for them, becoming automatons to their base instincts, not realising their true potential as spiritual beings in a material projection.
Butlerian jihad when?
After reading this I nearly went onto some spiritual guru's youtube channel and warned them how AI powered robots will hose them down with bleach, but then I realised I wouldn't gain anything from doing that. I mean, I'm obviously not into that stuff so why would I waste my time watching their videos, unless I derived some kind of flaccid pleasure in wafting my opinions in their direction, but who would do that?
@@Hiroprotagonist253 The spice must flow.
@@Dose_0x0 You too can gain superpowers by dry fasting for 40 days or 960 hours. Watch as you become aware that the observer within affects reality and reveals that causation dually by natural phenomena and conscious thought. Give to those with less and watch how you unexpectedly receive from those with more. As if some kind of hidden forces and laws of both thought and energy exist. Just as one cannot see a magnetic field with the naked eye, and air is often invisible, you might find that the inner eye is the master of the realm of the unseen. Now, about that bleach…
I don't want my video game to be overused nostalgia. These humans will relentlessly recycle pale imitations, that do not have the contextual essence of the original experience.
Those same humans have been trapping entire genres in overreliance, while preventing genuine innovation.
It’s Eva Rothschild name drop for me
The google ui is so bad tho. I just dont get why and how can such a big company get such a basic thing wrong.
why does the voice sound fried
Can't wait for grok 3.0 to blow all of these out of the water
LOL. Ok Elon.
Only Microsoft can compete with Google. Having good AI models is one thing. The other thing is to let them do useful things. Neither OpenAI nor X have an environment where you can implement AI to do useful things. Anthropic can do it with Amazon, that is also working. How the relationship between Microsoft and OpenAI will work out we have to see. The only frontier model developer who has it all under one roof is Google. That gives Google many massive advantages although we are seeing only the first baby steps yet ...
My next phone: Pixel 10 Pro
Wasnt gemini the one who tried to create a language