They probably have your credit card details so what kinds of things are you scared of them getting? It's probably already on the internet (ssn, creditcards, address, ph, ex girlfriend talking about your pepe of facebook, credit score, your digital scent (mac/key) across reddit and any other forums and websites, cookies, etc).
We need more use cases. Right now, I have the impression that this agent can only summarise and forward emails. Needless to say, I would not trust such an agent to work with a potential client. You have no idea how many clients send disastrous first emails, but after some reasonable conversation, they turn out to be fantastic new partners. So, right now, such an agent is not helping but instead wasting time.
"Agents" is becoming an even more blurred term than AGI. Adding RAG or fine-tuning to a model via a dedicated UI, and using a lot of those models within the same ecosystem, is definitely going to be the type of "agent" people are going to talk about for the next ~year.
We're in the age of the dinosaurs trying to shove AI into all of their products. Meanwhile the next generation of apps are being developed by startups all over the world that will be 100x better than what MS, Google, etc have offered us so far.
@@uw10isplayaagents can perform actions without being prompted outside of their initial instruction. Previously you would upload the document and give it a prompt. Agents do this automatically.
This looks mind-blowing... However, the set of agents and triggers and the end of the video looks ripe for embarrassing mistakes. Fine if the communicating emails are all inside the company, but terrible the moment someone wants to apply the concept to relations with 3rd-party partners. There should be a way to sandbox/rehearse these scenarios, before deploying the system for public use.
I believe these agents don't really have the capability to do the work like people who actually know the office like I do. Most analysts spend their day doing all kinds of tasks that they can't even explain how they get done. They pull information from different sources, not just emails. And today, there's already a lot of know-how to automate things with RPA, Excel macros, Power BI, and so on. So, I don't think much will change, and for this specific use, the hype feels empty. That's my take
Translating corporate speak, "humans are inefficient, and we can do this like a million times better." Not sure what it means for our future the fact that they're 100% right. Other than I imagine a lot of lay-offs.
So emails get summarised and delivered to a responsible person. This chap will have another agent taking similar action on incoming emails. And so we have a chain of agents forwarding emails/collaborating. Who does the job?
The way gpt still makes mistake understanding a well written input, don't really hype me considering that normal CS experience is flooded with terrible grammar or mispell (mostly due to speech to text). Yeah, no... Thank you. I'd rather keep checkint the email myself rather than following wrong instruction.
This isn’t agents , tf are these people talking about , I’m beginning to think these ai stuff has no real business use case it solves , maybe its too error prone for businesses to use
I agree, they dedicated a whole presentation just to show how their so called "agent" organizes emails, this is ridiculous. I want to see how agents become my digital copies and do parts of my job in parallel with me or even the whole projects with my supervision only.
So most people said this same thing one year ago. Today ever major business are dumping shit tons of money to have there own models. The use case saving companies millions is in the transportation industry of goods. Next these AI are exceptionally good at research and creating new marketing content. So do they save companies money yes that a fact. Are they AI of the sci-fi future no they are all of 2-3 years old in the wild. It took about 20 years from the computers creation to massive adoption and then another ten to really make these things super usable. The internet was like this but slightly faster about ten years where it took over. AI is no different it has to mature but AI is has an exponential growth curve. So give it another five years year after year it will start taking over things. They’re literally bringing nuclear facilities back online to power these things. Nothing happens over night.
Copilot is where MS is invested why would he not say the short comings of the product. He’s not talking about the dumb answers that copilot actually provides in many situations
MS has a big hold on corporate enterprise data , intelligence etc so building their own ai agent on top of those conversations doesn’t excite me much. Are there non corporate examples of ai agent though where the data , business intelligence field is wide open. For eg real estate investors ?? Can ai agents help investors by removing the middleman aka realtor completely ?
This (and all AI) will help the Age of Abundance appear much sooner if we put our objections aside and just focus on it making the whole world better. It's all about intent in the present moment of your existence (where the path to all possibilities starts from).
All this sounds very fancy however I doubt large corporations will allow Microsoft to literally read and make decisions on their most sensitive information
Augmentation of everything else that has been billed. Tapestry is the perfect word for it because that’s about how much it will be used for. What happened to Google will not happen to Microsoft it just happens to be the book the seals are of the same origin
Nah, IMO it's just the same number of people now doing more things. I think the use of AI will also drive the demand for more business activity like this. SO supply goes up a lot (because AI can do it) but demand yeah also goes up more (the need to do this all) and there will still be human workers overseeing it all - a big part because the tech will create so much more of it. In my mind the big question will be: is it more profitable (incrementally) to keep the size workforce there is now and add AI on top of it as opposed to eliminating people and riding on the savings. I'd think it would be the former, but who knows I guess and we will see.
offer a carrot for free or smal fee. gather data and make it mandatory on windows. Once you have to use it for your pc to work, start to ramp up fees and plans. Microsoft in a nutshell These arent really agents. seems like its just prompts on top of prompts, the only difference is the prompts are on some webpage and generic. Also they didnt show any agents working togetaher at same time. this is just a personal asistant if your coworker that knows these things is on vacation hahaaha
The end of humanity is coming sooner than everyone can expect, initially I thought we have ~5 years but recent AI developments arou d the globe tell me we have only 2 more years
Nonviable.. Why have humans spend their time creating automations that can be automated or require futher automation? A multi-agent system (MAS) is composed of multiple interacting intelligent agents - autonomous entities that can sense, learn models of their environment, make decisions and act upon them. Microsoft is still working at the singal level agent.. in other words a Co-Pilot assistant with too many wistles and bells !
I'm not believing anyone can deliver an AI-related feature if OpenAI can't. Get your shit together (MS & google), release a model that matches OpenAI (or dare I say, excel it), then we'll listen to you. Your voice chat is absolute trash. I don't know how even after buying OAI MS is still ages behind... how do you fuck up this bad!? Oh, I know, looks like they not only didn't fire whoever was managing Skype during COVID but probably made them in charge of handling the AI development.
"All we want is all your data."
Well, you're not getting it from me at least.
They probably have your credit card details so what kinds of things are you scared of them getting? It's probably already on the internet (ssn, creditcards, address, ph, ex girlfriend talking about your pepe of facebook, credit score, your digital scent (mac/key) across reddit and any other forums and websites, cookies, etc).
@@Seriouslydave nice logic, good luck with life
@@anubisai 🤣
@@tunestaryes they are
7:49 The precision at this moment is just flawless. Major props! 👏
We need more use cases. Right now, I have the impression that this agent can only summarise and forward emails. Needless to say, I would not trust such an agent to work with a potential client. You have no idea how many clients send disastrous first emails, but after some reasonable conversation, they turn out to be fantastic new partners. So, right now, such an agent is not helping but instead wasting time.
Microsoft’s version of Zapier within its ecosystem
Yeah, try to build something with the Co-Pilot Studio and let's talk then. It is the most frustrating, bad working piece of crap I've ever used in AI.
I can't wait to start finding all the security flaws and data leaks in these "agents".
this does not sound like agents, this sounds like using an ai that you can add extra post training data to that it can pull data from
"Agents" is becoming an even more blurred term than AGI. Adding RAG or fine-tuning to a model via a dedicated UI, and using a lot of those models within the same ecosystem, is definitely going to be the type of "agent" people are going to talk about for the next ~year.
We're in the age of the dinosaurs trying to shove AI into all of their products. Meanwhile the next generation of apps are being developed by startups all over the world that will be 100x better than what MS, Google, etc have offered us so far.
@@uw10isplayaagents can perform actions without being prompted outside of their initial instruction.
Previously you would upload the document and give it a prompt. Agents do this automatically.
Agents are to ai what css is to websites
Isnt agents something we use in reinforced learning ? Why do they keep recycling these words to make it even more confusing.@@Seriouslydave
This looks mind-blowing...
However, the set of agents and triggers and the end of the video looks ripe for embarrassing mistakes. Fine if the communicating emails are all inside the company, but terrible the moment someone wants to apply the concept to relations with 3rd-party partners. There should be a way to sandbox/rehearse these scenarios, before deploying the system for public use.
Microsoft always finds the way to make everything boring.
Real business, where the real money is made, is boring.
maybe a marketing strategy to look different at all price
Is there a way to have an agent to check the addresses that is on Zillow and put it on a word document?
Two consecutive videos that have ended abruptly!
Thought it was just me 😂 I think he was rushing to get them out.
@@matrepharaoh8260what else is new…
Probably because he uses AI for the videos and doesn't check the videos before uploading
@@joselarenas likely
You make it till the end😢
Advanced voice mode just got out to eu btw
that's late.. im guessing eu regulatory nonsense
Really? In Wich country?
Really? In Wich country?
I believe these agents don't really have the capability to do the work like people who actually know the office like I do. Most analysts spend their day doing all kinds of tasks that they can't even explain how they get done. They pull information from different sources, not just emails. And today, there's already a lot of know-how to automate things with RPA, Excel macros, Power BI, and so on. So, I don't think much will change, and for this specific use, the hype feels empty. That's my take
agreed. these agents "really" look like normal software features that are just able to understand the text/data they are passing around a lil better
They relaunched GPT‘s
😄 Yep!
ALL We desire isbfull control over each other
Translating corporate speak, "humans are inefficient, and we can do this like a million times better." Not sure what it means for our future the fact that they're 100% right. Other than I imagine a lot of lay-offs.
What an inspiring event! Can't wait for the next one to show Taskade’s autonomous AI agents and agentic workflows in action! :)
Good one! Can AI agent develop, test and deploy a web application if I provide UI and business requirements?
This is huge, it seems that o1 really gets to work under the hood... more layoffs incoming
so maby we need more policies to transfer wealth from the poor to the rich, to excelerate this process even further
@@App-Generator-PRO sure, I'm sure the Fed will bring the 5% interest rate back very soon
So emails get summarised and delivered to a responsible person. This chap will have another agent taking similar action on incoming emails. And so we have a chain of agents forwarding emails/collaborating. Who does the job?
The way gpt still makes mistake understanding a well written input, don't really hype me considering that normal CS experience is flooded with terrible grammar or mispell (mostly due to speech to text).
Yeah, no... Thank you. I'd rather keep checkint the email myself rather than following wrong instruction.
Very useful thank you. What tool u used to create this video by thr way.
Admin support. Bye bye. No new jobs created. AI doing its' job for its' overlords beautifully.
When did this keynote take place?
This isn’t agents , tf are these people talking about , I’m beginning to think these ai stuff has no real business use case it solves , maybe its too error prone for businesses to use
This changes the game completely for processing referrals for my mental health clinic. Hubspot wanted to charge me $800/month for what Copilot can do.
yes for the moment its more a combianison of human + AI. The AI increase productivity but the human should watch if the AI task is correct
I agree, they dedicated a whole presentation just to show how their so called "agent" organizes emails, this is ridiculous. I want to see how agents become my digital copies and do parts of my job in parallel with me or even the whole projects with my supervision only.
@@pandereodium it will come in 2 or 3 years
So most people said this same thing one year ago. Today ever major business are dumping shit tons of money to have there own models. The use case saving companies millions is in the transportation industry of goods. Next these AI are exceptionally good at research and creating new marketing content. So do they save companies money yes that a fact. Are they AI of the sci-fi future no they are all of 2-3 years old in the wild. It took about 20 years from the computers creation to massive adoption and then another ten to really make these things super usable. The internet was like this but slightly faster about ten years where it took over. AI is no different it has to mature but AI is has an exponential growth curve. So give it another five years year after year it will start taking over things. They’re literally bringing nuclear facilities back online to power these things. Nothing happens over night.
I think Microsoft stole the term copilot from Zapier. They were using the term in 2022 and Microsoft Copilot came out a year later.
Copilot is where MS is invested why would he not say the short comings of the product. He’s not talking about the dumb answers that copilot actually provides in many situations
Apple has pages. How did they get the naming of that without copyrighted infringement?
MS has a big hold on corporate enterprise data , intelligence etc so building their own ai agent on top of those conversations doesn’t excite me much. Are there non corporate examples of ai agent though where the data , business intelligence field is wide open. For eg real estate investors ?? Can ai agents help investors by removing the middleman aka realtor completely ?
that's assistants, essentially, not agents ..
This (and all AI) will help the Age of Abundance appear much sooner if we put our objections aside and just focus on it making the whole world better. It's all about intent in the present moment of your existence (where the path to all possibilities starts from).
🤔
New line of attack for corporate espionage, just ask their AI agent to summarize the proprietary information it has access to
Wouldn't be that easy.
All this sounds very fancy however I doubt large corporations will allow Microsoft to literally read and make decisions on their most sensitive information
It looks very compute and prompt size heavy.
Augmentation of everything else that has been billed. Tapestry is the perfect word for it because that’s about how much it will be used for. What happened to Google will not happen to Microsoft it just happens to be the book the seals are of the same origin
Welcome to software 4.0 - The Age of Botification(ua-cam.com/video/HMejvWDMYrE/v-deo.html)
These tools are awesome
This is good for scammers not for complex company's systems.
Say goodbye to your admin job
100%
Nah, IMO it's just the same number of people now doing more things. I think the use of AI will also drive the demand for more business activity like this. SO supply goes up a lot (because AI can do it) but demand yeah also goes up more (the need to do this all) and there will still be human workers overseeing it all - a big part because the tech will create so much more of it.
In my mind the big question will be: is it more profitable (incrementally) to keep the size workforce there is now and add AI on top of it as opposed to eliminating people and riding on the savings. I'd think it would be the former, but who knows I guess and we will see.
For real
Amazing 🎉
Zapier or n8n do the same today
We see that AI creates admin overhead.
Microsoft’s AgentForce (SalesForce)
Why do I need a work or school email to use this?! cmoooon
They can wake me up when its included i the office 365 standard subscription
Internet computer protocol AI can make any software you want in seconds
How does an economy work when no one has a job?
AI has limits; it is not human.
Why does this sound like they’re running palantir underneath
soo... it's a custom GPT then. Just like how I can upload docs to a GTP and ask it questions about the docs. This is SO lame.
8:42
Good for Microsoft… 🤖🖖🤖👍
So making everybody unemployed is exciting to these companies
offer a carrot for free or smal fee. gather data and make it mandatory on windows. Once you have to use it for your pc to work, start to ramp up fees and plans. Microsoft in a nutshell
These arent really agents. seems like its just prompts on top of prompts, the only difference is the prompts are on some webpage and generic. Also they didnt show any agents working togetaher at same time. this is just a personal asistant if your coworker that knows these things is on vacation hahaaha
This is typical microsoft. Overly complicating things and the setup requires too much work and change management even though they say it is easy.
Useless tool real agents just do the work for you. Ai needs to be more in 3d and films
The end of humanity is coming sooner than everyone can expect, initially I thought we have ~5 years but recent AI developments arou d the globe tell me we have only 2 more years
RAG is everywhere
Not very different from notebook lm
So they built a custom gpt 😂
a bot watched this and released it to github next day
This big corrupted company will end humanity one day
He is only trying to sell. Go from sales team to tech team and you get to the reality of the offer.
Nonviable.. Why have humans spend their time creating automations that can be automated or require futher automation?
A multi-agent system (MAS) is composed of multiple interacting intelligent agents - autonomous entities that can sense, learn models of their environment, make decisions and act upon them. Microsoft is still working at the singal level agent.. in other words a Co-Pilot assistant with too many wistles and bells !
I'm not believing anyone can deliver an AI-related feature if OpenAI can't. Get your shit together (MS & google), release a model that matches OpenAI (or dare I say, excel it), then we'll listen to you. Your voice chat is absolute trash. I don't know how even after buying OAI MS is still ages behind... how do you fuck up this bad!? Oh, I know, looks like they not only didn't fire whoever was managing Skype during COVID but probably made them in charge of handling the AI development.
agent 007
Stop copying diary of a ceo thumbnails. Its lame
❤️🍓
Don't trust them.
ua-cam.com/users/shortsbhClALUYNlk?si=B4XbOymRz5II3qO5
A second brain
First !
Zzzz
Agent, copilot,new blah blah blah.
So its only good for businesses? Garbage