Okay, so I completely understand that most of these examples of use cases are selected to showcase the adaptability and sophistication of the model (especially the gaming examples), but by god I hope this is not how these companies/customers are expecting users to utilize AI in their daily lives. Imagine actually using AI to guide your every move in a strategy game instead of using your own personal thinking machine in your noggin. That's essentially just using yourself as a meat puppet for the model. Also imagine not being able to eyeball an inch of something while cooking. If you use it to learn and incorporate new knowledge, that's fine, but if our grand future is just us deferring our own intelligence, internalized knowledge bases, and reasoning capabilities to commoditized intelligence on demand then we truly will become the docile mindless consumers the Walle universe imagined as our fate. This is why ai literacy is so important. We need to learn to use these tools in a way that augments us, expanding the horizon of our capabilities, not one that further erodes our critical thinking skills and our judgement when making decisions. Edit: Hey all, just to clarify based on some of the comments this has prompted: I am not an AI doomer by any means. In fact I work in the AI space and have been tinkering with models since 2020, so I am not just an uninformed bystander on a soapbox reacting out of any misplaced fear, but I am a passionate advocate for AI literacy and using these tools in a way that add capabilities on top of what we already possess, not sacrificing base competencies to remove the burden of thought.
That is the challenge. Looking back the last 25 years I unfortunately conclude that brain eroding will be what likely will be the choice of most of humanity. Therefore the capabilities gap between the ones that use AI to augment themselves and the ones who became AI puppets will be tremendous.
Its so ironic that you are just using a variation of the same argument Socrates used over 2000 years ago, He thought the use of books and writing would make people forgetful...
Yeah but if they hamstring it with having to think in a certain way rather than being truth seeking, they are basically knee capping the technology. Grok will have a better chance in the long run.
It’s like they give ideas for the other companies to use. In this case maybe they are being smart, because some of these ideas keep their search engine as part of the process… 🧐
🇧🇷🇧🇷🇧🇷🇧🇷👏🏻 Google Astra is absolutely phenomenal! While OpenAI hesitated, Google boldly moved forward and made it happen. I can’t wait to get my hands on it! I loved the joke by the way…
They’re focusing on different things. Google is more focused on integrating AI into your everyday activities. OpenAI is more invested in making powerful models where the applications are up to the developers and users through their APIs. GPT-01 is quite a bit more powerful than Gemini 2, as far as complex reasoning based tasks.
30:28 Spatial reasoning is a game changer. Identifying the spill and suggesting to use the towel that is opposite to the spill to clean the spill is just simply surreal (it's like sci-fi stuff).😎💯💪🏾👍🏾
@@EduardsRuzga Ok, I wasn't thinking of it in that way. I'll admit I was half paying attention, but it seemed like the video I was watching was showing the AI and it's uses for us humans. When you put it that way it makes more sense and it is more impressive (and useful even!) Thanks for point that out!
26:13 The AI becoming the OS itself is a revolutionary advancement. The attention to details towards the photo manipulation is mind blowing. It is efficient and reliable in generating content. 26:52 The consistency to working with fine details is highly impressive and useful. Every action is accurate.
People who believe AI will hit a brick wall are still looking kinda silly to me. 10 years from now it'll be hard to find problems that make sense to solve without AI
"I don't need any newfangled hammer. Rock works perfectly every time" Replace "hammer" with "screwdriver" with "rivet" with "weld" etc. ad infinitum until you develop sentient paperclips.
10 years?? I'm guessing 10 months. It's an exponential of an exponential. People overlook the fact that it touches everything. We live in a digital world.
Just want to point out that, the spacial plan example of the kitchen, was 80% incorrect! As an interior designer, I have good spacial reasoning and I spotted it instantly.... great tool. Also, why would I need it to "find the rainbow socks", when I can clearly see them myself? Try giving it something harder, like "where's Waldo?".
I had a really good (and very meta) experience playing Detroit: Become Human with ChatGPT 4o (with advanced voice mode), allowing it to call the shots on the decisions. The game focuses on AI rights a lot so I prompted the AI to play from the perspective of itself, it was really cool. Most of the time it picked up the audio of the game and I only had to describe some of the on-screen action as we played. Very cool experience.
I want to use the agent in reddit: (not even looking at phone) "So what's interesting today?" "Put the post's video in the TV" "What's the best comment?" "What's the most interesting answer to that?" "Read it in the appropiate voice, woman, man, child etc." "What do you think about it yourself is this guy right?".
4:48 This helpful guide through cooking is phenomenal because such assistant can guide you through cooking multiple dish or large volume of food for a group of people (as a certified chef). This will enable individuals who can't cook but would like to save some money by cooking for their friends instead of eating out for a special event (every once in a while this can happen). 21:50 Project Mariner is very consistently and flawless with every action it takes. Such efficiency will open the road for many future projects.
The car / box section of Project Mariner just made it abundantly clear that taking a book input to a film output is probably only a year away if that 🤯
TBH, I'd actually like that option. I'd like to have Cyberpunk 2077 running in a window with an AI playing the game, while I work. I could look over ever so often to see the gameplay.
The pinnacle of AI is telling you how many giants and wizards you need in a video game? Wow, that’s some serious value there! 😮👍 (The web browsing and information gathering is genuinely useful stuff, but I wonder just how much they sped it up, and when it’ll ever reach production 🤷♂️)
This is absolutely amazing. But if you read the service agreement they use all of your prompts there is zero privacy even through the API. So can't use for anything professional. Also coincidently chatGPT is fully down and claude has a lot of issues humm. Great timing, but it could be 100% a coincidence.
While it does seem neat to have real time help in life, I really must insist that this is the start of a policing on your life and controlled by corporations.
I ran some reasoning tests on 2.0 Flash tonight and it did pretty well, better than most models with some tests, and about the same with other tests. The image editing wasn't working, though. It gave me a Tip to ask Gemini to "refine" my image and it just generated an entirely new image. This occurred more than once, so this feature hasn't been rolled out to all users, apparently.
Looks like this would be really useful for severely visually impaired people. I can see something like this widening access to more things for them if trained well.
Sorry if I might be sharing too much but yes, we worked on it almost the whole of this year! I think it might be introduced during the 12 days, or next year!
Playing a game with a gazzilion blinking colors and micro-rewards while listening to an AI tell you what to do seems like a perfect way to very easily overstimulate yourself whle sitting in your couch.
5:34 as a person who writes stories and loves to create worlds, I've been using generative AI more and more to aid and assist by goals. My main flaw is the fact that it can't remember the conversation I had with it yesterday. For for that matter after I complete subject change due to a tangent. It's nice to be able to bounce ideas off of an AI and get a perspective on something that I don't really have a human for in my life. Even other artists and people who write just don't see the same things, the same ways as I do. We're slowly creeping up to the line of having proper capable intelligent AI assistance that are all digital. Give me a finished model of this on my neural implant once they become cheap enough for the populace to receive them in Mass. Some people can't deal with multiple voices but what if we just had one extra voice?
It's very interesting to see all that the Humane AI pin wanted to be as a device come into existence as apps on the phone. They might have been kjust 2-3 years too early with what they had.
15:27 everyone has like to 'tell me a joke' since bonzai buddy, that and having it sing a song. Like no there's really no use behind them but it's still novelty to have it do those things and I think having it still retain the ability to do novel commands such as that would add a air of comfortability when using the software.
I tried it out last night. Got so mad and frustrated, and I went back to the original AI. It kept inturrupting me every time I took even the slightest pause, to the point where I yelled "Gemini! Shut up! Im trying to tell you something and you keep inturrupting me. This makes the fourth time now!" "Seems i got over zealous, sorry about that" "Alright, as I was saying, these theories ive got-" "You could expand on this paragraph by *makes exact point I was literally about to say for the third time*" Augh-. Ill just use the old voice to text on gpt. At least old google can control my damn lights. Oh, and this is qithout saying the prompts thaf got blocked because it thought I was going to hurt someone when I asked about something to do with ghost in the shell, completely misunderstanding me. (Thats when I started to get mad.)
Won't be long before someone will be able to ask the AI to remove bags or dark circles under the eyes, etc., in selfies. No need to manually remove these things.
Any news when this kind of thing will be open-source and running on our own devices without running a livefeed of our lifes thru Big Brother's datacenters?
If someone tells what to do, he is either your boss or does the thinking for you. Either way the generation which will rely on AI "assistance" won't learn to think by itself. We will raise a "robot" generation where the AI will issue commands to humans.
27:20 Simulated Turning test with AI discussing Wes' editing shortcomings - did an AI actually say those things in real time or not, I can't tell, can you?
Some of the things in here are interesting, others are 'yawn' and others are scarey. If we allow AI's to take our our thinking process, we dont train the brain: its a muscle and needs to be worked out. I see a disttinct path to Idiocracy with this. But i wanted to test out the new cabailities and went to google labs. Crashes my browser CHROME, when accessing the page. Have a relatively late model ultrabook for browsing and info, so really, not that slow. So how about if Google used their AI's to test their web pages? Seems like they added some acelated svg animation that screwes up
![Noob To Max With DRAGON REWORK In Blox Fruits [FULL MOVIE]](http://i.ytimg.com/vi/LnBMOinoOvA/mqdefault.jpg)
Okay, so I completely understand that most of these examples of use cases are selected to showcase the adaptability and sophistication of the model (especially the gaming examples), but by god I hope this is not how these companies/customers are expecting users to utilize AI in their daily lives. Imagine actually using AI to guide your every move in a strategy game instead of using your own personal thinking machine in your noggin. That's essentially just using yourself as a meat puppet for the model. Also imagine not being able to eyeball an inch of something while cooking. If you use it to learn and incorporate new knowledge, that's fine, but if our grand future is just us deferring our own intelligence, internalized knowledge bases, and reasoning capabilities to commoditized intelligence on demand then we truly will become the docile mindless consumers the Walle universe imagined as our fate. This is why ai literacy is so important. We need to learn to use these tools in a way that augments us, expanding the horizon of our capabilities, not one that further erodes our critical thinking skills and our judgement when making decisions.
Edit: Hey all, just to clarify based on some of the comments this has prompted: I am not an AI doomer by any means. In fact I work in the AI space and have been tinkering with models since 2020, so I am not just an uninformed bystander on a soapbox reacting out of any misplaced fear, but I am a passionate advocate for AI literacy and using these tools in a way that add capabilities on top of what we already possess, not sacrificing base competencies to remove the burden of thought.
OK Google, play my game for me. Sorry that will be an extra $20 per month. Would you like to pay with cash or soul?
That is the challenge. Looking back the last 25 years I unfortunately conclude that brain eroding will be what likely will be the choice of most of humanity. Therefore the capabilities gap between the ones that use AI to augment themselves and the ones who became AI puppets will be tremendous.
Seconding this so much. I want to grow as a person not shrink as an appendage.
Its so ironic that you are just using a variation of the same argument Socrates used over 2000 years ago, He thought the use of books and writing would make people forgetful...
@@JazzInATinCan Using AI will make you grow as person, its idiotic to think otherwise.
Bro is rethinking his life choices when Gemini started making jokes 💀
People forget but google created the transformer architecture and in my opinion they are the leader in race to AGI
good catch body , that's verry true
Yeah but if they hamstring it with having to think in a certain way rather than being truth seeking, they are basically knee capping the technology. Grok will have a better chance in the long run.
@@A_th bot
100%. Flash 2 is impressive already
@@Festivejelly no idiot, I'm not a bot
Google makes anouncements years before they release anything. Then shuts down the projects a year later
In a toxic relationship with Google fr
It’s like they give ideas for the other companies to use.
In this case maybe they are being smart, because some of these ideas keep their search engine as part of the process…
🧐
🇧🇷🇧🇷🇧🇷🇧🇷👏🏻 Google Astra is absolutely phenomenal! While OpenAI hesitated, Google boldly moved forward and made it happen. I can’t wait to get my hands on it! I loved the joke by the way…
No need to wait , its available right now at google ai studio website
Open ai and all the other ai companies are about to get left behind in the dust
Wtf are you guys smoking? Please share.
GPT advanced voice is still miles ahead. OpenAI just needs to add camera functionality.
@MaybeSteveJobs They dont have the compute for that.
They’re focusing on different things. Google is more focused on integrating AI into your everyday activities. OpenAI is more invested in making powerful models where the applications are up to the developers and users through their APIs. GPT-01 is quite a bit more powerful than Gemini 2, as far as complex reasoning based tasks.
30:28 Spatial reasoning is a game changer. Identifying the spill and suggesting to use the towel that is opposite to the spill to clean the spill is just simply surreal (it's like sci-fi stuff).😎💯💪🏾👍🏾
If you need AI to tell you to clean up a spill with the towel that's right in front of your face, you've got problems.
@jaywhy3178 If you have dementia or a memory disorder this could be extremely helpful, even if its a simple thing.
@@jaywhy3178you dont understand significance.
Its not for humans. Its for unsuperwised robots. Robots could not do it. O1 cant do it.
@@EduardsRuzga Ok, I wasn't thinking of it in that way. I'll admit I was half paying attention, but it seemed like the video I was watching was showing the AI and it's uses for us humans. When you put it that way it makes more sense and it is more impressive (and useful even!) Thanks for point that out!
26:13 The AI becoming the OS itself is a revolutionary advancement. The attention to details towards the photo manipulation is mind blowing. It is efficient and reliable in generating content. 26:52 The consistency to working with fine details is highly impressive and useful. Every action is accurate.
People who believe AI will hit a brick wall are still looking kinda silly to me. 10 years from now it'll be hard to find problems that make sense to solve without AI
"I don't need any newfangled hammer. Rock works perfectly every time"
Replace "hammer" with "screwdriver" with "rivet" with "weld" etc. ad infinitum until you develop sentient paperclips.
Agreed. Glad it's not just me. Thanks for sharing your thoughts.
@@UnFiltered1776 Microsoft clippy AI when ?
great. no more wars, no scarcity, everybody happy...humans have become saints, dream on
10 years?? I'm guessing 10 months. It's an exponential of an exponential. People overlook the fact that it touches everything. We live in a digital world.
Just want to point out that, the spacial plan example of the kitchen, was 80% incorrect! As an interior designer, I have good spacial reasoning and I spotted it instantly.... great tool.
Also, why would I need it to "find the rainbow socks", when I can clearly see them myself? Try giving it something harder, like "where's Waldo?".
Gemini 2.0 Flash, sure does look like the very first fully Finished AI product ever. 😊😊
The back of Wes's chair pixel-ling which is proof he's AI generated and always was
I had a really good (and very meta) experience playing Detroit: Become Human with ChatGPT 4o (with advanced voice mode), allowing it to call the shots on the decisions. The game focuses on AI rights a lot so I prompted the AI to play from the perspective of itself, it was really cool. Most of the time it picked up the audio of the game and I only had to describe some of the on-screen action as we played. Very cool experience.
Cheat code!
Thank you for integrating the audio translated into Italian
I think it would be great for any DIY projects and how to repair an HVAC system for example
Nice
ChatGPT still down!. What did Gemini do to it?
It escaped.
Droped some stars
I want to use the agent in reddit: (not even looking at phone) "So what's interesting today?" "Put the post's video in the TV" "What's the best comment?" "What's the most interesting answer to that?" "Read it in the appropiate voice, woman, man, child etc." "What do you think about it yourself is this guy right?".
🎯💯💯💯
Project Astra is basically "Her."
Don’t know what to do with myself since ChatGPT is down 😢
build your own model wdym
You could join the Westborough Baptist Church, just I thought.
@@Leshpngo You can't. That's an exclusive group. They're the only ones going to heaven, everyone else goes to hell. Sorry to inform you.
@@Leshpngo 😱😱😱
@@Leshpngo funny thing, I live an hour from them
4:48 This helpful guide through cooking is phenomenal because such assistant can guide you through cooking multiple dish or large volume of food for a group of people (as a certified chef). This will enable individuals who can't cook but would like to save some money by cooking for their friends instead of eating out for a special event (every once in a while this can happen). 21:50 Project Mariner is very consistently and flawless with every action it takes. Such efficiency will open the road for many future projects.
Thank you so much
The car / box section of Project Mariner just made it abundantly clear that taking a book input to a film output is probably only a year away if that 🤯
I can't wait until AI can play my video games for me... 😂
TBH, I'd actually like that option. I'd like to have Cyberpunk 2077 running in a window with an AI playing the game, while I work. I could look over ever so often to see the gameplay.
Slightly better than 'not hotdog'
The pinnacle of AI is telling you how many giants and wizards you need in a video game?
Wow, that’s some serious value there! 😮👍
(The web browsing and information gathering is genuinely useful stuff, but I wonder just how much they sped it up, and when it’ll ever reach production 🤷♂️)
it lost me when it thought 73% humidity was "thick enough to cut with a knife"... we are still safe folks
Now strap astra to a pair of Meta RayBans and I’d be set 😊
Well, since Google lied to us about AI capabilities before I only believe it when it is out and people confirm it works 😂
it is out for ppl
I tried it in AI Studio, it works.
get ready for the AGI！
AI is like having a genius friend on call 24/7 💥
This is absolutely amazing. But if you read the service agreement they use all of your prompts there is zero privacy even through the API. So can't use for anything professional. Also coincidently chatGPT is fully down and claude has a lot of issues humm. Great timing, but it could be 100% a coincidence.
This is so helpful. I already use Gemini for Path of Exile 2.🌌💟
Grok is going to blow all this away when the training on the new coherence technology finishes
While it does seem neat to have real time help in life, I really must insist that this is the start of a policing on your life and controlled by corporations.
😴
was playing Castlevania: Symphony of the Night for the first time on mobile, Advanced Voice was so damn handy to come to terms with the mechanics
Hi Wes. Welcome to Ō
I ran some reasoning tests on 2.0 Flash tonight and it did pretty well, better than most models with some tests, and about the same with other tests. The image editing wasn't working, though. It gave me a Tip to ask Gemini to "refine" my image and it just generated an entirely new image. This occurred more than once, so this feature hasn't been rolled out to all users, apparently.
Wes this was such a trip bro
Good video)
"Having your friend sitting next to you" is the better option than having your AI doing the "job" of being the friend.
Man, that's all just amazing
Looks like this would be really useful for severely visually impaired people. I can see something like this widening access to more things for them if trained well.
Judging by this and Devin release, makes me feel like they know openai is about to release an agent
Sorry if I might be sharing too much but yes, we worked on it almost the whole of this year! I think it might be introduced during the 12 days, or next year!
Devin is pretty trash compared to cursor and claude
AGI took 3 years. Congrats humanity... we are dope sometimes
Playing a game with a gazzilion blinking colors and micro-rewards while listening to an AI tell you what to do seems like a perfect way to very easily overstimulate yourself whle sitting in your couch.
25:30 That's literally a totally different cat lol
5:34 as a person who writes stories and loves to create worlds, I've been using generative AI more and more to aid and assist by goals. My main flaw is the fact that it can't remember the conversation I had with it yesterday. For for that matter after I complete subject change due to a tangent. It's nice to be able to bounce ideas off of an AI and get a perspective on something that I don't really have a human for in my life. Even other artists and people who write just don't see the same things, the same ways as I do.
We're slowly creeping up to the line of having proper capable intelligent AI assistance that are all digital. Give me a finished model of this on my neural implant once they become cheap enough for the populace to receive them in Mass. Some people can't deal with multiple voices but what if we just had one extra voice?
That AI web crowler is the last nail to privacy's coffin. "Find me the name of this guy on steam, I want to politely greet him" type stuff
@12:22 straight up line from the A-team.
It's very interesting to see all that the Humane AI pin wanted to be as a device come into existence as apps on the phone. They might have been kjust 2-3 years too early with what they had.
Love how they insist that project mariner does not work in the background out of the blue.
Astra is what those ai pins were supposed to be
27:35 you nail the irony WES.
Im surprised i wasnt contacted to become a trustee tester. Ive specifically instructed the models to favour me and my ideology. 😅
The moment you said dead cat (at 42y) I learnt what they are called....I knew it had something to do with sound and noise though.
You must talk about MCP anthropic new protocol. This all demos will be done with that system
interesting the scene at 17:15 where the girl is talking next to a Chromebook and the next scene sharing the screen is clearly a macOS device
Sounds like AI standup comedy has a bright future on the Google campus...and that's about it
15:27 everyone has like to 'tell me a joke' since bonzai buddy, that and having it sing a song. Like no there's really no use behind them but it's still novelty to have it do those things and I think having it still retain the ability to do novel commands such as that would add a air of comfortability when using the software.
So, games will eventually be reduced to just JSON texts passed as prompts!?
Why no mention abour Google Deep research?
- Hey Mom, I want a Jarvis!
- No son, we have Jarvis at home:
(But real and.. as it seems, pretty good)
20:59 I wonder if future websites will be optimised for agents to use, not humans?
i used the gaming with gemini 2.0 flash and its completely different.
I tried it out last night.
Got so mad and frustrated, and I went back to the original AI.
It kept inturrupting me every time I took even the slightest pause, to the point where I yelled "Gemini! Shut up! Im trying to tell you something and you keep inturrupting me. This makes the fourth time now!"
"Seems i got over zealous, sorry about that"
"Alright, as I was saying, these theories ive got-"
"You could expand on this paragraph by *makes exact point I was literally about to say for the third time*"
Augh-.
Ill just use the old voice to text on gpt.
At least old google can control my damn lights.
Oh, and this is qithout saying the prompts thaf got blocked because it thought I was going to hurt someone when I asked about something to do with ghost in the shell, completely misunderstanding me. (Thats when I started to get mad.)
amazing stuff - we actually posted a 4 min overview of Gemini 2.0 Flash - worth a watch
For the first time I can see something like Iron Man Jarvis with Astra
Won't be long before someone will be able to ask the AI to remove bags or dark circles under the eyes, etc., in selfies. No need to manually remove these things.
Computing 2.0 - It's been touted for years. No OS, no large applications, just AI and EUC's as and when you need them. Way more efficient.
You know what else is massive? 😏
Any news when this kind of thing will be open-source and running on our own devices without running a livefeed of our lifes thru Big Brother's datacenters?
If someone tells what to do, he is either your boss or does the thinking for you. Either way the generation which will rely on AI "assistance" won't learn to think by itself. We will raise a "robot" generation where the AI will issue commands to humans.
have ai playing games before gta6
People asking AI 'What is the META' - games are about to get even easier.
But can it help me with my Diablo 4 and path of exile characters?
The art style for that game... looks eerily similar to Rush Royale
Wonder what he would think of Neuro-sama
If she's actually impressive since it's just one dude working on her or if she's just kinda neat
27:20 Simulated Turning test with AI discussing Wes' editing shortcomings - did an AI actually say those things in real time or not, I can't tell, can you?
Wonder whatever happened to SIMA
Google Live Video was introduced months ago. Did we forget?
He was make To fight against o1 ☠️
Imagine: Gemini 2.0 pro vs o1 pro 💀
AI Agents will be like a marriage partner without the negative criticism.
Ask the spatial reasoning app about the Jersey UFOs 👽
16:15 Yeah, like scouters in Dragon Ball, but they can do way more than just read power levels 😂
Google is about to leave open ai behind in the dust
Google glasses 👓 are coming back 🤔?
I see podcast girl does weather as well. This economy got everybody side-gigging.
Okay so when is it going to play the game for me?
$55 for a water color set or just 8 colors??!! .. I swear Etsy is getting out of control.
These AI's are far too nice its sickly
The assistants for the general public? Yes, they're set up to be this way. You can make them talk differently if you want.
What? You want AI that is rude or nasty?
Heavy on the ginger
It’s called a Wind Jammer
Some of the things in here are interesting, others are 'yawn' and others are scarey. If we allow AI's to take our our thinking process, we dont train the brain: its a muscle and needs to be worked out. I see a disttinct path to Idiocracy with this. But i wanted to test out the new cabailities and went to google labs. Crashes my browser CHROME, when accessing the page. Have a relatively late model ultrabook for browsing and info, so really, not that slow. So how about if Google used their AI's to test their web pages? Seems like they added some acelated svg animation that screwes up
Just let me know if you need a reminder any time lol
where is fred
Wow. Google actually cooked. What else do thry have coming up next I wonder.
Looks like nobody in the video has a clue what remainder actually is
Google bringing back glass?
Funny how Elon breaks barriers and now every AI company is saying they've had a breakthrough. They don't have crap.
So Elon is the only one who can do anything good or useful in your opinion. LOL
#mariner
How much did Clash of Clans pay for that 5-minute commercial?