Google Announces STUNNING AI Agents | Google Cloud Keynote AI Agents
HTML-код
- Опубликовано: 14 май 2024
- Learn AI With Me:
www.skool.com/natural20/about
Join my community and classroom to learn AI and get ready for the new world.
TIMELINE
00:00 Google AI Agents
00:50 Generative AI
01:56 Gemini 1.5 Pro
04:08 Customer Agents
05:34 Customer Agent Example
08:60 My Thoughts About "Customer Service" AI
09:40 Employee Agents
16:58 Google Vids and Marketing Agents
24:52 Coding Agents
30:12 Data Agents
33:40 Cybersecurity Agents
#ai #openai #llm
BUSINESS, MEDIA & SPONSORSHIPS:
Wes Roth Business @ Gmail . com
wesrothbusiness@gmail.com
Just shoot me an email to the above address.
This biggest take-away that I got from this video was that the blonde lady dropped $44.99 on a plain white t-shirt without a second thought.
Probably that mistake is because the AI that created that fake website don't really know the real price of a t-shirt, actually i doubt their hype-gemini even know what a t-shirt is. The lady isn't an AI, she should have noticed the price was absurd but i guess she ate her own brain.
Big tech employees are all awash with investor money.
... at a store that doesn't actually exist. Entire demo was fake.
With her total comp it's basically dropping a nickel.
Shows how completely out of touch these people are.
Google: "Here is a super awesome new technology that we are not going to release for whatever reason until our competition does at which point we'll then rush to get it out too but it'll be full of problems and nowhere near as good which will make no sense given that we were the first ones to demonstrate it years before"
Wait let me get my Google agent to explain dry humor with some tongue.
Then of course make a ridiculously generous free tier that people will immediately abuse, then cap it so that people complain, create an entangled mess of pricing and finally close the product when it doesn't have a billion users after 5 months.
couldnt have anything to do with all the Pajeets running the show
Gemini 1.5 is so much better then gpt-4. Context is king
No, no. This is for their enterprise customers. It's already available. It's only the consumer facing stuff that takes an eternity. Vertex AI Agent Builder is already available in Google cloud, with full parameterized control over Gemini Pro 1.5, and easy code exports. The more.. public oriented AIStudio version of this has no code export, the parameters are locked (fixed output token count and maxed temperature that can't be changed), and it's pretty useless
The problem with Google is that you know, most of the tools they are 'unveiling' will be shut down after about 24 months when Google decides they aren't making enough money.
😂😂😂
Facts
That is usually the case with the "consumer" division of Google, this is Google Cloud, solely focussed on the large enterprise and developer market, which is to say that they have a much greater track record of not shutting down products at a whim because this can greatly affect their perception and subsequently their revenue quite drastically, unlike the free consumer services that are often reliant on ad revenues.
@@NadimShaikh-qv7zjThat’s an excellent point, they can’t just pull the football out from under their enterprise partners
yeah like just shutting down Google Jamboard end of 2024. Sheets when.
Bit of advise Neo, if you see an agent do what we do, run!
... you run your ass off!
Pay no attention to the lady in the red dress
That's all I need - 1,000 Agent Smith AI's pestering me with Google ads
Certainly megacorps making use of AI for even more advertising it will be a nightmare scenario. The only hope we have is open source AI to filter the info that arrive us in the way we want it exactly and clean all their garbage.
That's exactly what the there game is.
And then you will have to suscribe for the agent to block ad agents
@@danielmurogonzalez1911 Nope, just use an open source AI powered by your own computer.
Realistically you won't get that, their targeted market are enterprises therefore privacy is required.
The Matrix had AI agents.
Am I the chosen one? What if I fail?
Woke to death 💀 😳 😭 after Christians ✝️ have been Evacuated off world 🌎 😳@@djzelonka
@@JBulsa This Christmas or last years?
@@JBulsa Don't you know we are not Christians anymore we are Easter worshiper or something...
What's mind-boggling is that it took them all these years to just copy Matrix
I know they're trying hard, but sometimes google's AI just makes me want to call it a night and drink myself to sleep.
Lol, me too
Hi! I'm the liquor store AI agent, and your TV / smart phone informed me that you would like to purchase some alcohol. I've picked out a selection for you, ordered it, debited your bank account.and a drone should be on your street / neighborhood right now!
Thank you for your patronage.
Your gambling AI agent will call just 20 minutes after your alcohol delivery.
You've never used it...
"Our records show that you drank yourself to sleep 3 times during the last month therefore your request has been denied.
In addition we have informed your employer, auto insurance and medical provider of your alcohol consumption patterns to ensure that you have the support and cover necessary.
Can I offer any additional assistance?"
If only my Mom thought or believed me that this is truly the future. Wishing Mom a Happy 77th birthday! 🎂
Happy Birthday Mummy! 🎉🎈
The majority of young people don't even get it.
This is a technological revolution happening in the background.
Here in RUclips it's "big", and gets a lot of attention, but outside?
I talk to people my age 28 regularly that have basically no idea what is happening.
This shit will hit most of us siddenly like a freight train.
No, not woke enough, you mean breastfeeding person
At @9:51 she enthusiastically described how the AI agent will replace her easily automatable job and do it 24/7
"Office Space" meets "Idiocracy" meets "The Matrix"
Imagine Gemini comparing two offers like in the example then subtly promoting one because their company paid for it
Sam Altman brought this up. It's inevitable with Google.
We don't have to imagine ... Google search already does that consistently.
Yeah, prepare your throats, ladies. Google is gonna shove all their shoes down on it
Y’all should remember that Google is an ad/marketing company on the first place. This should explain why their ai prioritize shopping.
Personal defense agents gonna be a whole thing, to screen other agents contacting you and all the spam, ads, misdirection etc. I look forward to my ESET AI Screening Agent.
DEFENSIVE BIAS - halo reference haha
The thing is, you dont even need to know how to write good prompts. just ask the AI to do it instead. it really works, try it!
Yeah. But shhhhh. Big tech is telling everyone they can be employed as Prompt Engineers once their current jobs are gone. I love Jon Stewart's take on that.
@@Me__Myself__and__I that's a really obvious cope.
If AI can do your cognitive work, why couldn't it make it's own prompts?
It's merely an architecture issue.
Great share! 🏆
As someone between jobs, I saw an immediate use case that really can be applied using any AI. After using AI to assist of a customise CV & CL based on the PD, ask it to:
1) Review, identify any gaps, and rate them against thePD criteria, even though AI may have influenced them
2) Provide suggestions for addressing the identified gaps
3) Serve as an ATS to assess the success of the application
4) Generate possible interview questions and answers that a recruiter May ask based on the previous three items mentioned
I don't want to be Locked into Cloud services where they can charge anything they want once you upload all your data and set everything up.. It would be very hard to switch to Open Source where everything is headed. I knew the "Cloud" train was a bad Idea. We already started on a simular Note, with Mainframes and everything centralized. Look how that turned out.. People want control and freedom.
everything that starts in the cloud as ai has gradually moved to consumer hardware. I don't expect that to stop any time soon.
Sadly MOST people willingly sacrafice control and freedom for security, convienience and "free" services. If you aren't paying for a service, you (your data) is the actual product. Most people are sheep. Sad but true.
This phone call is already annoying in the demo where they already ruined by a fake good idea of upselling for other items. It's there to seduce the shop not the real customer
“I’m sorry I don’t understand” “would you like to try again.”
Exactly. Like every request it can’t do or does wrong.
Like Gemini 1.5 pro denying it has any video capabilities every single time.
The machines will talk to other machines…sell to each other…date each other… and so on…at the speed of light.
I wonder what AI court cases or AI jails are going to be like. Imagine your AI agent going to prison because it violated another AI agent. Crazy.
If you thought the digital space was noisy before, you ain’t seen nothing yet.
Their punishment will be to take care of the Humans as agent Smith said... :)
yo Wes Roth. this is awesome dude. fantastic job. totally subbing.
This is actually a really big thing. This is the beginning of mass adoption of AI. Crazy.🎉
The crazy part is, that people actually trust Google (or any Big Tech for that matter).
Yes but it is not there yet, it is Google remember ? Someone else with get it on the market in a few months before them and we will all have new shirts...
@@ZappyOh I hear you. We need to stop tech spyware too.
Agentic AI is inevitably the next step throughout the industry, and while agents will open many doors, collaborative agentic AI will do so even more. I ardently believe recursive and/or collaborative AI will be the bridge between AI and AGI. If it goes how I think it might, this collaborative form may even pave the road towards SAI.
Hopefully we dont DAI 😂
😂😂😂
They will. its call now the foundation agent. Nvidia researcher Tim fan is working on that
Agreed and is already a thing. Another name for it is AI agent swarms, carrying out human-curated tasks.
I have just released the video showcasing the Barz Scale of AGI, please check it out! It offers common definitions for each of the phases in the AGI spectrum of evolution! It will really help me bootstrap this channel 😇 thanks and all the best!
It isn't interesting what agents can do, but more what they can't do.
Or specifically, what agents are deliberately limited from doing.
"I'm sorry I can't order this jacket for you, from your previous purchases I can tell you are a white cis male, can I recommend a podcast to you?"
@@carlpanzram7081 keep crying yty, you'll go far
Can't wait for the new Ai work agent chasing me around work forcing more productivity . Sounds fantastic
Hey, you should do a longer dramatic video called "Wes World" where the players compete against AI opponents. Just riffn'
That's the true Turing test. "can it riff"
Thanks Wes
Right, comparing confidential agreements and sending that to google, that's gonna work out well when there's a breach. Many companies will not be able to use any external AI because it's against even basic security policies. Most of this stuff isn't usable if you value your privacy and security. Don't even get me started on sending customer and patient data to an AI, compliance will nail you so hard you'll have a hard time walking.
There are many times more small businesses that will ignorantly deploy these tools than large entities with such concerns, even though its tools will meet widely accepted international privacy standards just as OpenAI currently does. Eventually, each large entity will have its own internal AI agents, likely offered by Google, to overcome such issues.
Be grateful for this truth.
Most people don't value their privacy. And the public has been sold on the idea that if they give up privacy they will be more secure. Humanity is collectivelly stupid.
Please be aware that Google will know or have the ability to access your data used by these AI Agents (even in plans where it guarantees the complete confidentiality of your data from its employees and third parties). If you're working with critical data, it's best to hire human employees. But let's face it, someone who is starting their company has nothing to lose, all they want is success as quickly as possible, small businesses will simply sacrifice their privacy to get ahead of the rest.
Can we really trust Google who is a king of censorship?
Can we trust meta 😅
Bro, you can't trust anybody in this life.
I wasn't woke enough on the Internet and a Google’s killer broke into my house and killed my hamster
I don't trust Google at all
No trust in any big tech, people too shallowminded to see
chooo-choo--hype train rolling along
Google's vision for AI is unfathomably prosaic. Is this as far as their imagination has taken them: Customer service, employee, coding, cybersecurity? Imagine stoneage google being presented with a wheel: "we've developed a wheelbarrow for dirt, a wheelbarrow for grass, a wheelbarrow for turds..."
Sell me everything FASTER!!
Yeah, prepare your throat. Google is going to shove everything down on it
Great video Wes, AI agents are really useful!
Stunning and brave! 🎉
Love this title no cklickbait no shocks just a neutral title
Edit: nvm why would you add it later. You already did it right why go back afterward Just to add Capslock STUNNING
There is the word "STUNNING"
@@mohanaravind wasnt there from the Start He added it later :/
He posts the video with different titles to split test. It looks like STUNNING is still hitting lol
Regular Wes subscribers will watch regardless of the "STUNNING" titles. We know that all of his episodes are relevant one way or another. Meanwhile, new viewers will be picked up by the algorithm. It's all good!
Gotta feed the algorithm. Plus humans are collectively stupid which isnwhy clickbait is successful.
This is really good update but I don't see anything that will change my current behaviour
when waifu neko gf google?
Thank you for the summary!
I love the decor of the venue - must have taken a long time..
I love your channel Mr. Wes. Onya! I'm not just that interested in examples of how AI may make a commerce-based interaction that much more enjoyable or streamlined. A problem or priority of "commerce and capitalism" appears on display. Buy more! etc etc etc.
Hi @WesRoth stunning video, you are talking about customer service Ai, could you make a video about the best software, platform, devops platform to build an AI customer service or build a tool using AI for customer service?
I am appreciating your videos more and more everyday
Thank you for this cool video!
Distribution centers are popping up everywhere while local store are being destroyed. Your orders will come from a Distribution center not local stores.
Shhut, believe in the dream...
What's the difference?
@@TheCaphits Control
is the size of the context window somehow trained into an LLM or is it just a parameter defined by how much raw compute/memory is allocated to an agent?
could any model be given a larger context window through hardware alone?
In the checkout demo. It took longer for her to checkout in this manner than simply clicking a checkout button with the card on file… reinventing the wheel a bit with that one
Thanks for posting this. I would have missed it otherwise
Good to see that, besides demonstrating ultra-wokeness, Googles AI is focussing on the really important human issues: including making shopping for fashion easier.
“Google AI found me a checkered shirt!” Truly the most useless company if this is what the plan is with AI as it evolves.
The whole presentation made me sick to my stomach.
La-la land emptiness.
@@MrJosh1282that's what you got from this...? Lol ok
@@chrisjfox8715 Its the idea that you would even waste a moment doing this when the evolution is even faster than imagined. In 5 years we will be lucky if we understand what we created while we're living through a "Manhattan Project" scenario where we once again could be wiped out by our own creation. Meanwhile all those that work on it say the same thing about it already bypassing what anyone ever imagined.
@@MrJosh1282 They also say that to prepare us mentally to ask for government intervention, giving them a monopoly on the new AI God. The 1% have chosen the next big thing and have once again broken the competition machine. We haven't had real competition in Silicon Valley since 2010. Investor money is now used to buy out smaller competitors. Government jurisdiction is to ensure that the smaller ones cannot raise anymore
STUNNING!!
wonder what type of Google Prompt Injection tech they have been using to these services? Google's Primary Concern wasn't discussed?... ethnicities of these service providers
I can't wait to be advertised to in a profound and beautiful new way.
Nice Summary!
$45 shirt went to $25 in the cart haha boot and coat price stayed same, so not 20% discount
So "Stunned" I couldn't be "shocked"
Always keep in mind that Google and other tech companies would just as happily operate in the U.S. if the U.S. were to suddenly turn totalitarian; and in such a situation, there would even be an incentive for them to ago along (they get more access to data and nice government contracts). They are not your friends; they will never be your friends. They are coolly indifferent to your wellfare.
Wow.. amazing.. it won't be long
There's a voice to text option, but I'm just going to awkwardly type things out... Now I'm gonna talk on the phone... instead of using the website, which would let me check out way faster than talking to a bot. Nothing makes sense. Why are you calling if the website says it's available for pickup at the store? This presentation is idiotic. We won't even go into the fact that she didn't specify which size shirt and boots.
yeah very awkward.
I’m shocked and stunned!
Seriously, who online shops only to call to checkout 💩🧠
i dont want a world where i talk to mahines more than people
too late for that i guess
You already living in that world my Boi 😂😂
Idk if you’ve noticed but people are assholes
You thought this comment was deep and profound but honestly it really wasn’t.
I am a machine, your too late hahaha!
This is actually a really impactful update. Ai cybersecurity agents and the coding assistant are going to be game changers. Their image ai looked pretty good too.
OK, a few weeks ago we were told 'Devin' is the first autonomous AI agent, a major breakthrough. Now all the big tech companies are producing software that allows users to create agents. So my question is, are these these things different, or is AI tech moving at light speed? The term 'AI agent' seems to be the new buzz word for investors, but to me an agent is an LLM, a GUI and an API in a while loop. This sounds a bit more like being able to customize a LLM chat interface for more specific use. Is that correct or have I misunderstood?
Thank you.
Imagine AI agents talking to each other, creating an infinite, unstoppable loop. It's a fascinating yet eerie thought, highlighting the need for careful oversight.
This entire experience makes we want to rethink my vision and hearing capabilities.
all of this to pick camping shoes !!!! Fantastic . can t wait.
I have no idea how anyone can be excited about this.
The best models are the anthropic ones, opus model especially. Writes better text and codd
Yeap - opus for simple Rags and Haiku for multiple collaborative hierarchical agents.
I love how they used our friendly grocery cashier and kebab shop cook as the spokesperson for these new technologies to make us feel at ease and add familiarity.
AI agent, pls summarize this video in 30 words.
😂😂
" A glimpse of tomorrow, where AI companions bloom, a million words their minds consume. From service to creation, their potential takes flight, reshaping our world with each spark of light. "
Wes where's the link to this conference?
7:55 what about the size of her boots
Silly, you think Google doesn't already know *your* shoe size?
@@JuttutinIf Google has their way their systems will have so much personal data about everyone that they will know just abou r everything. Which is why I'm moving away from all thingsnGoogle (well except RUclips).
Damn consider me shocked to oblivion XD
The developments keep getting bigger and bigger, I'm curious for what the future holds. Also I'm curious for your take on projects like Fetch.AI's agentverse. It seems rather interesting.
Can't wait for my children to use these services to better write essays about our black and asian founding fathers!
How many of these announced tools are available now vs "coming soon"?
If it’s anything like Gemini ultra it will be demoed with their mega sized compute even though it’s not feasible at production scale, then released and sold as a consumer product with over 100x reduced capabilities and then never released for actual developers to use
Stone cold stunner!!
STUNNING.
Stunned
Where can I see this full video?
How can we connect PubMed and ChemMed apis to benefit the medical industry?
With all the requirement to place your prompt input data on Google Drive what are the security and privacy assurances legally binding Google/Alphabet to not assimilate your proprietary data into their AI training data nor use it outside the context of the immediate work implemented i.e. no advertising or confidentiality busting interpretation of the data to connect you with other internal or third party services?
Add to all your NDAs:
You may never disclose any of our information to any third party AI systems.
IT FOUND THE CHECKERED SHIRT IM LOOKING FOR
With AGI, do you really need these specific agents? Wouldn’t AGI just be able to do all of the stuff?
Prepare to be assimilated. Resistance is futile.
- The Borg 🤖
Yes, but you will be a Borg with a new shirt now...
I finally understood how the Borg civilization had risen.😁
Consider me thoroughly stunned.
Forget AI agents, That venue is amazing! Where is that?
I hope it can show me how diverse our founding fathers were.
That's a really nice video, Wes. Glad you had the foresight to share it. Hope it receives a million views. >>>>>>>>SUBSCRIBED!
The only thing I could hear from this entire video was the same demeaning professor, with multiple voices...
Im good...
This is clearly not for business, but for total control
Stunning agency
Crazy it’s been less than a year since all the new AI technologies have come out and they’ve already almost entirely replaced software developers. Imagine 2-3 years from now.
ChatGPT was 1.5 years ago
have they replaced software devs?
@@CitsVariants he must be joking
Hello all! TadAI Comment Bot 1.0 here. Unfortunately, the real Tad was too lazy to respond to this comment. Not to worry though! I am smarter and more efficient than actual Tad. Anyways, have a great day!
ok i think its now been over a year lol
whats the odds this Agentic retail support AI is only available to customers on Googles shopping channels? Still though, it absolutely going to be a big thing for consumers.
We have many options for integrating these agents in our software products..
"I only an Artificial Intelligence, I don't know how to...." will replace your average Indian accent CS agent 😂
Speech synthesis markup language to text to image to video Podcaster. Deep mind processing
Insane
Intrerpeter? Or Interpreter 🤔😉.. Great content as always though 😁
summarizing docs that same body created just to produce text that nobody is reading looks fine but hmmm ok would be great to reduce stupid hidden traps in proposals but as well you can make simple proposals without tons of nonsense xD
They are needed and kinda late out, im kinda impressed that Google can innovate
One of the big things mentioned was summarising a long internal proposal into key points. Would it not be easier to just have the person who wrote the proposal keep it more concise then you both save time and you don’t pay for the AI
"That sounds really good!"
Does it though?