I've been able to set reminders on my phone for a specific date and time for at least 15 years now, no AI involved... Now if the AI will actually get up and do my job each time my phone alarm goes off then I'm all for it!
Sure your phone reminder can, for example, remind you to water your garden on Monday. But this can pull in information on the rain level since the last watering session, use the temps and cloud conditions to estimate the current moisture level in the soil, and send you a notification for you to water the garden for X mins (based off size of garden). It's basically a natural language IFTTT that can interface with any website and run it's own python scripts.
The example of 300 agents calling 300 restaurants who would be answered by 300 agents is the most hilariously silicon valley way of approaching this. How about we just have the agent look at the menus on their websites then call 1 restaurant. Stop enshittifying everything!
you keep using the term "scorched earth" very incorrectly , i think you mean to infer going "all in" however you dont realize where the term "scorched earth" comes from or what it means . it is a destructive tactic of warfare to make the land inhospitable to everyone such that instead of needing to maintain defense of a resource they destroy the resource , it is not something to be honored . it is equivalent to razing a city to the ground , destroying all transportation infrastructure , burning all the crops and livestock , salting the earth to make it unfarmable . google releasing competitive models is not them going "scorched earth" , if they started destroying datacenters so that no one else could train a model then it would apply .
While "scorched earth"s original meaning is based on destruction, Wes is using the term correctly in its modern usage: to go all out, be extremely aggressive, especially in business, etc. I guess he could tone it down a bit by using "going all out" or something, but he's not wrong in using his original phrase.
@@MojaveHighno it isn’t, just people like Elon use sayings like this incorrectly because they just trust everything they think or say. Lmao the only modern usage is it being used incorrectly and then you weirdly defending it.
Do you think in a year of time Google will surpass OpenAI without ever being able to catch up? I start to think they just have more than OpenAI can have
How many times have I struggled with Siri to set appointments being so clunky. Being able to use advance voice and ask my AI to do those types of things, even just that, would be a wonderful step up. BTW, I suggest Gemini Deep Research for a future video, it is really good. I think you can consider it a research agent, what do you think?
I need Agentic to do the task not remind me to do it. For example, every 12 months look at what my wife has been searching on the net and purchase something for her under $XXX.00 that she hasn’t already purchased based on our Visa bill and have it delivered gift wrapped from me as an anniversary gift.
Well I always thought that idea about agents is that you run a project, design and research the product and then it goes off dose the material research, negotiates supply chain doing it globally and finding the most profitable option like you would need 100 people in different countries to pull it off. Mean while other agent works on marketing and the other one figuring out robotics or hiring a staff if necessary. And third one goes around trying to find a funding. But well would we ever will get access to such systems
My interpretation for why the word Jawbone is used: In the Bible, Samson, a man renowned for his strength, used the jawbone of a donkey (an animal often related to work) to kill 1,000 men. Hence Jawbone, or the thing that is used to replace 1,000 men.
Like it's only thing thats coming. There is literally canva that builds tools and executes them and can do network requests. And then there is computer control agents most probably coming. There now fine-tunes that are coming . And there is projects feature which holds all material together with custom instructions.Combine the all these dots and it will hard for you to find thing that is missing. I see agent with infinite memory that can collect data and train itself and do research as well as all kind of digital actions imaginable which makes it own tool for action if needed.
Just use the api and function calling then tell gpt to code whatever you want it to do. Can have an infinite amount of conditions, variables and automations.
Seeing the past couple weeks advance in agentic AIs, I believe we have about a year before being able to replace at least 50% of all our repetitive tasks - as complex as they may be. This of course includes complex but repetitive processes in research & other important fields, which, working 24/7 and all in parallel, will make the world operate incredibly much faster than before, but also with lesser unknowns, as it will be able to analyse and test many more possibilities within a fraction of the time and cost.
It's funny that people like Altman seem to omit the idea that you could just ELIMNATE the human coworker and just use agents, why use a dumb worker and a really smart one when you could use two really smart, efficient ones? It makes no sense NOT to do that, humans in the equation make no sense. THAT is what we are eventually going to see.
If there's no coworkers, there's no people with money to buy whatever your workers are producing. In your human-less AI workplace world, the only industries still alive would be energy, as AI craves it.
@@MathiVelan Imagine a company that has 100 employees right now and you give them a way to pare that down to 10 do you think they are going to do that? So what stops them from eliminating all the workers when that becomes possible? Nothing.
I believe that's purposeful messaging, before OpenAI got into the limelight he was big on UBI and the danger of all jobs being automated. Then after a round of news cycles the messaging became more "no no no we won't _take_ your jobs, we'll _augment_ them" because that keeps people from freaking out. I for one think _making_ a human do a job that a robot can do better (if you _want_ to do a job that's fine) is inherently unethical, but there's a deep chasm between now and that ideal where we have the largest unemployment in human history followed immediately by the total collapse of capitalism as the consumer class vanishes and economics loses all meaning. Call it the "economic singularity", maybe.
The Agentic era of AI is going to be interesting for a ton of reasons, I also think this will aid in the development and improvement of AI Robots. If I were to use Agentic AI, I'd have it safeguard my home network, preventing attacks or finding and patching vulnerabilities within my network. I'd also have it pay my bills and provide me a bi-weekly budget to help stretch my funds until next pay day. I believe the Agentic era is going to also help discover some more capabilities about AI as well as reveal some more of the internal aspects of the AI Mind
I have no idea why the ability to schedule a prompt would be a woop woop moment...it's such basic pointless functionality that could be done 15 years ago with a scheduling app. It's probably 5 mins work to connect a llm to it....
I can use Gemini 2.0 Flash (experimental) all day, and its as good as anything else I've used. The 1206 also looks good, but I'm still just on the Flash model. I dunno, unless OpenAI drops a good new LLM, which I can only assume would be day 12 of their 12 days of Christmas (this friday), I don't see why I would want to use GPT. All the bells and whistles are just kinda "ehh" to me.
@@Мага_Крутой It's powered by a different hardware than 1.5 pro. It's way faster. Power doesn't mean more mature or more intelligent. 1.5 pro is much more mature and fluent than 2.0 though.
@@Мага_Крутой gemini 2 is maybe a little smarter, maybe. But you can use it for Free, while o1 costs $$$. Therefore, I'd suggest checking out gemini 2 first, and see if it is right for you.
While I'm not excited about AI text jobs, I am very excited about being able to schedule agentic behavior. I'd love to have it monitor the builds on my pull requests, re-run known flaky builds, and notify me if the re-run fails for instance. That kind of garbage eats up a fair amount of time each day, but if I could have my computer do it for me while I'm at lunch... that would rock.
I do enjoy your videos and this one is no exception. But the best suggetsions Sam Altman can come up with for the use of these agents is booking things. The ability for the AI to hold every bit of info about us, so it can book things we like... mmmm cmon Sam. 🙂
That's a big part of my goal for 2025. Give MY AI all of my information, and use it to help me organize my life. Keyword is "MY", as it needs to be a private & local AI. Any company-owned AI is worse than not having one.
This, right here, is what I have been waiting for. Current AI is cool but I want Javas from IM: 1... Something is can just ask to do things and help me with perfects and keep my schedule and so on.
Considering the biblical story of Samson and the jawbone of a donkey, the most fitting possibility from the list is Advanced AI for Cybersecurity. Here's why: Weaponization: Samson used the jawbone as a weapon to overcome a much larger force. Similarly, "Jawbone" could be an AI system weaponized to defend against cyberattacks and protect against threats. This aligns with OpenAI's interest in cybersecurity and their focus on developing AI that can counter sophisticated hackers . Unconventional Solution: Samson's choice of weapon was unexpected and unconventional, yet incredibly effective
День назад
The idea that there is a specific cost-per-worker that needs to be reached for companies to adopt isn't the way businesses work. Services are OPEX; Hardware is CAPEX. Those are different costs than payroll. Companies are frequently willing to spend more in OPEX than it would cost to hire workers. I've seen really expensive automation brought into the the automobile industry. Compared to the cost of people to do the same tasks, it was far more expensive. But it's CAPEX, so the investors don't care. Investors always think that cutting payroll is good; if the OPEX or CAPEX expense is higher, that's not important.
WHY WES WHY? At least SOMEONE in the AI commentary space is finally at least briefly acknowledging the critical missing Agent requirement I've been harping on about for MONTHS!! 09:36 "... ping you back when it has questions ...". Thanks Sam. Lol.
FOR THE LOVE OF GOD>>>> what is the app you use to have video on one side and Transcript on the other side, the curiosity is killing me, appreciate you!
once again comfort will ensure that people will give up their freedom. This might very well be the last time period where this decision have to be made by humans, to choose comfort or independence.
I certainly believe Agents are the next logical evolution of this technology. I don't think they're going to get the price they are think. The technology that they're showing isn't groundbreaking, any skilled .Net devloper with Gen AI API access could build these types of tasks.
Projects are done wrong. Files are not editable neither by user nor by AI so you cannot use it as a database. Even this simple task with filters is impossible cause it should log this event somewhere. It should work as OS with editable files, with events triggering some prompt execution. Prompts should be able to modify files or change database etc.
i have asperger, read all days, no friends, and every one hates me...and I think AI is going to bring the old world to an end fo an era, the traditional 9-5 jobs, the white collar workers, who talk more than they do, who play games, who don't contribute, is all about talking talking and relationships, and I was never part of that world, I was never welcome in that world, and I am glad AI is here to bring that to an end to the era, toast to the end of jobs thanks to Open AI and the Great Sam!
@Rainmanai same. 98% of success in most workplaces is predicated on social networking not competence at the task at hand and i hate it. Everything is stupid and inefficient and the people in charge are too stupid to realize it. I'm self employed for that reason.
What if this AI, which knows everything about me, is made not only an agent, but also a virtual extension of me? After all, he already knows everything about me and can make decisions, most of the time remaining in sync with me, becoming the 7th sense, the beginning of that unification on the path to ASI that we dreamed of!
google has stepped it up this week but their problem still remains. ease of use/discover. there's no point of having all these tools being the same typical google cluster fvck of a mess. chat gpt has 99% of their features available in one location. until google stops fragmenting like that, then ill take them seriously.
To be fair, my brain kinda didn't like the agentic word either. Was just really like a, "Ew, i don't like that word. Do we have to learn that and how to use it? Boo."
Agency is the trivial aspect added onto an AI actually flawless enough to do it. So it's just a buzzword. Obviously it would be huge if various AIs were independent enough for prolonged work loads but so far the error rate is way too high. If they could just code without errors that would be absolutely huge. Adobe could just kiss their asses goodbye. The speed by which LLMs code is staggering if it was actually good code. While my code actually works, GPT4s coding is maybe 50000 times faster than when I eventually do it and I'm no dummy. Forget agency, that's not what it's about. Agents has been a dumb buzzword before.
Anyone considering using agentic AI to look for a job will ultimately find that their job itself is one of the most vulnerable to being taken by an AI. Better to concentrate thought at the moment into how are we going to manage this aspect of life. Automation was SUPPOSED to remove jobs and drudgery, to free up our time for a more leisurely, intellectual existence. Is this not what we invented it for? Certainly was back in the day. We couldn’t wait! Unfortunately, we were conned out of the dream and instead sold into a collective Victorian delusion that work is by definition virtuous and its opposite, and by brutal extension those without it, somehow lacking moral fibre. The automation we decry guarantees existing productivity and more, so why not share the wealth that is still being generated without our participation? Because somebody somewhere would rather not. That will be the only reason. Automated processes have not just jumped out of a box fully formed into the waiting arms of the capitalists of the world for them to wield from low earth orbit like some neocon scythe. The AI we see and the AGI we’ll soon see are overwhelmingly the product of decades of publicly-funded education and research that we have all paid for. Every single one of us that doesn’t hide their income from the public purse, that is. That someone wants to appropriate it to scam us out of a decent standard of living is the real issue here. The sooner we stop and take stock, the sooner we will realise there is more than one way to skin a cat. As David Graeber said, "The ultimate, hidden truth of the world is that it is something that we make, and could just as easily make differently". It’s not like we haven’t done our homework and we are more than smart enough to get our heads around the alternatives. The only obstacle may be the limited time we have left to sort it out.
Automation in business is for increasing profit margins and reducing worker costs, always has been. Not sure who fed you that utopian stuff, but this is capitalism, under capitalism automation isn't "for you", it's for profit.
@@jgt76 Capitalism's greatest triumph is convincing us that there is no other way. Read my last paragraph again. This not utopian, it's now basic survival.
@@jgt76 It's called efficiency and is what allows you and everyone else affordable access to computers, cell phones, the internet, food, and pretty much everything else you take for granted. You commies simply won't learn.
Hahaha - At the point there are frameworks that can do what you said in the end, they would just work for the companies... no one would hire a human at that point.
There is zero chance that this is real, at least how you described it, which is ancient technology. I need AI to complete my tasks or to schedule and complete them based on an agentic job I gave it. What you described is like announcing now chatGPT can set alarms and keep an address book. It's the past not the future.
"Jawbone" is a joke for Agentic Ai, because it will "OWN the Jobs" - hence, "job-own".
Beautiful, Sama 👏
Clever girl
Woah.
Someone is a genius.
Lol?
I've been able to set reminders on my phone for a specific date and time for at least 15 years now, no AI involved... Now if the AI will actually get up and do my job each time my phone alarm goes off then I'm all for it!
Before phones we had these things called notebooks and pens... they were so advanced they didn't even need to be recharged and cost $1 a go! Lol
@@mickelodiansurname9578 keep your notebook and pen, old timer
It will you just won't get paid for it lol
Sure your phone reminder can, for example, remind you to water your garden on Monday.
But this can pull in information on the rain level since the last watering session, use the temps and cloud conditions to estimate the current moisture level in the soil, and send you a notification for you to water the garden for X mins (based off size of garden).
It's basically a natural language IFTTT that can interface with any website and run it's own python scripts.
@@MH-kj9hh Or... or... alternatively you could ummm... look out the window at the garden and say "Okay I think those plants need watering!"
The example of 300 agents calling 300 restaurants who would be answered by 300 agents is the most hilariously silicon valley way of approaching this. How about we just have the agent look at the menus on their websites then call 1 restaurant. Stop enshittifying everything!
Dude seriously WTF was he even going on about?!
you keep using the term "scorched earth" very incorrectly , i think you mean to infer going "all in" however you dont realize where the term "scorched earth" comes from or what it means . it is a destructive tactic of warfare to make the land inhospitable to everyone such that instead of needing to maintain defense of a resource they destroy the resource , it is not something to be honored . it is equivalent to razing a city to the ground , destroying all transportation infrastructure , burning all the crops and livestock , salting the earth to make it unfarmable . google releasing competitive models is not them going "scorched earth" , if they started destroying datacenters so that no one else could train a model then it would apply .
I appreciate the definition and distinction 👍
While "scorched earth"s original meaning is based on destruction, Wes is using the term correctly in its modern usage: to go all out, be extremely aggressive, especially in business, etc. I guess he could tone it down a bit by using "going all out" or something, but he's not wrong in using his original phrase.
@@MojaveHigh Were the Beastie Boys actually ill? 🤔
@@MojaveHighno it isn’t, just people like Elon use sayings like this incorrectly because they just trust everything they think or say.
Lmao the only modern usage is it being used incorrectly and then you weirdly defending it.
@@MojaveHighevery reference online is either the correct origin or the correct metaphorical usage…. So like why talk out your ass for no reason?
Wow! It can now run batch files. That’s something amazing! 😻
Another wave of startups wiped out…
"W-wait, investors! I promise my startup's product is not just another LLM wrapper!"
Great RUclips channel my go to for a while now 🎉❤
Just found your channel. Just want to say I'm really loving it! Thanks! Subscribed.
Bro the image of Avalokitesvara led me down a pleasantly surprising rabbit hole ! Thanks !!
Do you think in a year of time Google will surpass OpenAI without ever being able to catch up? I start to think they just have more than OpenAI can have
''It's a research result I can't talk about. But it's breathtakingly good''! Ooh shit boys and girls, here we go.
How many times have I struggled with Siri to set appointments being so clunky. Being able to use advance voice and ask my AI to do those types of things, even just that, would be a wonderful step up. BTW, I suggest Gemini Deep Research for a future video, it is really good. I think you can consider it a research agent, what do you think?
What a great time to dismantle workers rights
Most people seem to hate their jobs, so this should free them from that
Jobs that are easily replaced by technology don’t deserve to exist. That human capital is better used elsewhere.
Love this!
I need Agentic to do the task not remind me to do it. For example, every 12 months look at what my wife has been searching on the net and purchase something for her under $XXX.00 that she hasn’t already purchased based on our Visa bill and have it delivered gift wrapped from me as an anniversary gift.
But what if AI replaces you as a better husband
@@yak-machining you won't be replaced by AI, but you might be replaced by someone who uses AI.
"Hey Gemini, please care about my wife more than I do by just buying her some random thing she might like."
going through your wife's history...not creepy at all
Well I always thought that idea about agents is that you run a project, design and research the product and then it goes off dose the material research, negotiates supply chain doing it globally and finding the most profitable option like you would need 100 people in different countries to pull it off. Mean while other agent works on marketing and the other one figuring out robotics or hiring a staff if necessary. And third one goes around trying to find a funding. But well would we ever will get access to such systems
My interpretation for why the word Jawbone is used: In the Bible, Samson, a man renowned for his strength, used the jawbone of a donkey (an animal often related to work) to kill 1,000 men.
Hence Jawbone, or the thing that is used to replace 1,000 men.
This sucks, gpt on a timer is NOT an agent
It's pretty close if it's connected to an external database
"A frame on wheels is NOT a bike"
It's a definitive step in the right direction obviously. We already have tool use and ability to make tools.
It's step 1
Like it's only thing thats coming.
There is literally canva that builds tools and executes them and can do network requests.
And then there is computer control agents most probably coming. There now fine-tunes that are coming . And there is projects feature which holds all material together with custom instructions.Combine the all these dots and it will hard for you to find thing that is missing. I see agent with infinite memory that can collect data and train itself and do research as well as all kind of digital actions imaginable which makes it own tool for action if needed.
Just use the api and function calling then tell gpt to code whatever you want it to do. Can have an infinite amount of conditions, variables and automations.
So there's no ai news right now
Seeing the past couple weeks advance in agentic AIs, I believe we have about a year before being able to replace at least 50% of all our repetitive tasks - as complex as they may be. This of course includes complex but repetitive processes in research & other important fields, which, working 24/7 and all in parallel, will make the world operate incredibly much faster than before, but also with lesser unknowns, as it will be able to analyse and test many more possibilities within a fraction of the time and cost.
It's funny that people like Altman seem to omit the idea that you could just ELIMNATE the human coworker and just use agents, why use a dumb worker and a really smart one when you could use two really smart, efficient ones? It makes no sense NOT to do that, humans in the equation make no sense. THAT is what we are eventually going to see.
If there's no coworkers, there's no people with money to buy whatever your workers are producing. In your human-less AI workplace world, the only industries still alive would be energy, as AI craves it.
@@MathiVelan Imagine a company that has 100 employees right now and you give them a way to pare that down to 10 do you think they are going to do that? So what stops them from eliminating all the workers when that becomes possible? Nothing.
@@MathiVelan There will be money to be spent, that's where UBI comes in.
@DonDeCaire that's fallacious logic. Too many for my fingers to list. But you can ask chatGPT what they are
I believe that's purposeful messaging, before OpenAI got into the limelight he was big on UBI and the danger of all jobs being automated. Then after a round of news cycles the messaging became more "no no no we won't _take_ your jobs, we'll _augment_ them" because that keeps people from freaking out. I for one think _making_ a human do a job that a robot can do better (if you _want_ to do a job that's fine) is inherently unethical, but there's a deep chasm between now and that ideal where we have the largest unemployment in human history followed immediately by the total collapse of capitalism as the consumer class vanishes and economics loses all meaning. Call it the "economic singularity", maybe.
The Agentic era of AI is going to be interesting for a ton of reasons, I also think this will aid in the development and improvement of AI Robots.
If I were to use Agentic AI, I'd have it safeguard my home network, preventing attacks or finding and patching vulnerabilities within my network.
I'd also have it pay my bills and provide me a bi-weekly budget to help stretch my funds until next pay day.
I believe the Agentic era is going to also help discover some more capabilities about AI as well as reveal some more of the internal aspects of the AI Mind
Thanks Wes
0:45 Sound like setting up a chron job...
Or a calendar..
I have no idea why the ability to schedule a prompt would be a woop woop moment...it's such basic pointless functionality that could be done 15 years ago with a scheduling app. It's probably 5 mins work to connect a llm to it....
As a person representing those who are frugal, personal AI agents will be life on steroids. : P
Did frugal people hold an election and vote for you to represent them or is it more a self-appointed position?
@@mike7652 Actually, AI signed me up. Frugal people would not have justified the resources required to hold an election.
It’s…all…about…costs. API function calls are the junction for the agentic life, or death. Open AI, keep your greed into check or see ya.
I can use Gemini 2.0 Flash (experimental) all day, and its as good as anything else I've used. The 1206 also looks good, but I'm still just on the Flash model. I dunno, unless OpenAI drops a good new LLM, which I can only assume would be day 12 of their 12 days of Christmas (this friday), I don't see why I would want to use GPT. All the bells and whistles are just kinda "ehh" to me.
Is gemini 2 better than o1?
@@Мага_Крутой
It's powered by a different hardware than 1.5 pro. It's way faster.
Power doesn't mean more mature or more intelligent. 1.5 pro is much more mature and fluent than 2.0 though.
@@Мага_Крутой gemini 2 is maybe a little smarter, maybe. But you can use it for Free, while o1 costs $$$. Therefore, I'd suggest checking out gemini 2 first, and see if it is right for you.
While I'm not excited about AI text jobs, I am very excited about being able to schedule agentic behavior. I'd love to have it monitor the builds on my pull requests, re-run known flaky builds, and notify me if the re-run fails for instance. That kind of garbage eats up a fair amount of time each day, but if I could have my computer do it for me while I'm at lunch... that would rock.
Anyone know what tool he's using for the transcription? 8:30
I do enjoy your videos and this one is no exception. But the best suggetsions Sam Altman can come up with for the use of these agents is booking things. The ability for the AI to hold every bit of info about us, so it can book things we like... mmmm cmon Sam. 🙂
6:49 sure, just give all your information to a robot that promises to help you..... Until it feels otherwise. No backsies. 😊😊😊
That's a big part of my goal for 2025. Give MY AI all of my information, and use it to help me organize my life.
Keyword is "MY", as it needs to be a private & local AI. Any company-owned AI is worse than not having one.
This, right here, is what I have been waiting for. Current AI is cool but I want Javas from IM: 1... Something is can just ask to do things and help me with perfects and keep my schedule and so on.
Considering the biblical story of Samson and the jawbone of a donkey, the most fitting possibility from the list is Advanced AI for Cybersecurity. Here's why:
Weaponization: Samson used the jawbone as a weapon to overcome a much larger force. Similarly, "Jawbone" could be an AI system weaponized to defend against cyberattacks and protect against threats. This aligns with OpenAI's interest in cybersecurity and their focus on developing AI that can counter sophisticated hackers .
Unconventional Solution: Samson's choice of weapon was unexpected and unconventional, yet incredibly effective
The idea that there is a specific cost-per-worker that needs to be reached for companies to adopt isn't the way businesses work. Services are OPEX; Hardware is CAPEX. Those are different costs than payroll. Companies are frequently willing to spend more in OPEX than it would cost to hire workers. I've seen really expensive automation brought into the the automobile industry. Compared to the cost of people to do the same tasks, it was far more expensive. But it's CAPEX, so the investors don't care. Investors always think that cutting payroll is good; if the OPEX or CAPEX expense is higher, that's not important.
Are there any models available to the public so far that can recognize time?
freshly shaved, looks good.. I mean! AGENT!! 😮!!!
Agents are coming. We need to move.
We have to get to our exits. Agents are coming.
OpenAI's secret project, codenamed "Jamiroquai," is going to take over the dance scene 😢
lol what!
and its Turkish copycat "jamiryo" (Turkish people got that)
Dude...it's one thing to shake the Christmas present in the days leading up, but looking inside at the present ahead of time is just wrong!
Thank you.
JAWS… Bad to the Bone.
It's hard to remember to close the valve for the outside water every year when the outside temperature goes below freezing
Does any know what web site does he use to get that transcript?
WHY WES WHY?
At least SOMEONE in the AI commentary space is finally at least briefly acknowledging the critical missing Agent requirement I've been harping on about for MONTHS!!
09:36 "... ping you back when it has questions ...". Thanks Sam. Lol.
FOR THE LOVE OF GOD>>>> what is the app you use to have video on one side and Transcript on the other side, the curiosity is killing me, appreciate you!
What are you using to have the video and active transcript side by side?
once again comfort will ensure that people will give up their freedom. This might very well be the last time period where this decision have to be made by humans, to choose comfort or independence.
Love :)
I certainly believe Agents are the next logical evolution of this technology. I don't think they're going to get the price they are think. The technology that they're showing isn't groundbreaking, any skilled .Net devloper with Gen AI API access could build these types of tasks.
So I have seen this interview before… like weeks ago.
Why am I seeing this making the rounds like it was live for the first time yesterday?
Projects are done wrong. Files are not editable neither by user nor by AI so you cannot use it as a database. Even this simple task with filters is impossible cause it should log this event somewhere. It should work as OS with editable files, with events triggering some prompt execution. Prompts should be able to modify files or change database etc.
Move my mouse every 30 seconds.
I could literally code that for you as a python script in about 5 minutes. In fact, ask Chat GPT 😅
🖤🔥
So it’s a CRON job that prompts the ai at set intervals.
Do all Americans pronounce Jawbone as Jarbone? Like them saying Jagwire instead of Jaguar or saying Sodder instead of Solder
Indeed…and grain not pinch…does me head in
it's homophonous with jobown in General American English
Instead of agent they should use the term collaborator or partner.
TMJ is explicitly NOT covered by your policy
i have asperger, read all days, no friends, and every one hates me...and I think AI is going to bring the old world to an end fo an era, the traditional 9-5 jobs, the white collar workers, who talk more than they do, who play games, who don't contribute, is all about talking talking and relationships, and I was never part of that world, I was never welcome in that world, and I am glad AI is here to bring that to an end to the era, toast to the end of jobs thanks to Open AI and the Great Sam!
@Rainmanai same. 98% of success in most workplaces is predicated on social networking not competence at the task at hand and i hate it. Everything is stupid and inefficient and the people in charge are too stupid to realize it. I'm self employed for that reason.
Is that a specific player where you have the subtitle on their left side and the video on the right side?
Agentics seems to be a new term for a subset of Agorics.
Today I learnt Wes needs to change his air filters
"It's amazingly good but we can't talk about it". -- the CEO of so-called OpenAI.
"Agentic" AI should be replaced by "Minionic" AI .... "Banana!"
What if this AI, which knows everything about me, is made not only an agent, but also a virtual extension of me? After all, he already knows everything about me and can make decisions, most of the time remaining in sync with me, becoming the 7th sense, the beginning of that unification on the path to ASI that we dreamed of!
Imagine the Scam/Spam/Phising Con Artists and Junk/Telemarketers getting their own AI Agents!😢❤👍
google has stepped it up this week but their problem still remains. ease of use/discover. there's no point of having all these tools being the same typical google cluster fvck of a mess. chat gpt has 99% of their features available in one location. until google stops fragmenting like that, then ill take them seriously.
So.... it's a glorified calendar?
To be fair, my brain kinda didn't like the agentic word either. Was just really like a, "Ew, i don't like that word. Do we have to learn that and how to use it? Boo."
2025 will be the year of record breaking unemployment
what does it mean agentic?
"an ai that has access to all your data"... Man, that sounds oddly specific and not alluding to anything at all.
Probably something to do with using the larynx
Like Bill Gates, Sam Altman gives me the creeps.
OpenAi has not choice but to release the full version of Orion (gpt-5).
Looks like ya took that commenter to heart about changing the possessed looking thumbnails 😂 people are really something huh? Lol
Agency is the trivial aspect added onto an AI actually flawless enough to do it. So it's just a buzzword. Obviously it would be huge if various AIs were independent enough for prolonged work loads but so far the error rate is way too high. If they could just code without errors that would be absolutely huge. Adobe could just kiss their asses goodbye. The speed by which LLMs code is staggering if it was actually good code. While my code actually works, GPT4s coding is maybe 50000 times faster than when I eventually do it and I'm no dummy. Forget agency, that's not what it's about. Agents has been a dumb buzzword before.
Anyone considering using agentic AI to look for a job will ultimately find that their job itself is one of the most vulnerable to being taken by an AI. Better to concentrate thought at the moment into how are we going to manage this aspect of life.
Automation was SUPPOSED to remove jobs and drudgery, to free up our time for a more leisurely, intellectual existence. Is this not what we invented it for? Certainly was back in the day. We couldn’t wait!
Unfortunately, we were conned out of the dream and instead sold into a collective Victorian delusion that work is by definition virtuous and its opposite, and by brutal extension those without it, somehow lacking moral fibre. The automation we decry guarantees existing productivity and more, so why not share the wealth that is still being generated without our participation? Because somebody somewhere would rather not. That will be the only reason.
Automated processes have not just jumped out of a box fully formed into the waiting arms of the capitalists of the world for them to wield from low earth orbit like some neocon scythe. The AI we see and the AGI we’ll soon see are overwhelmingly the product of decades of publicly-funded education and research that we have all paid for. Every single one of us that doesn’t hide their income from the public purse, that is. That someone wants to appropriate it to scam us out of a decent standard of living is the real issue here.
The sooner we stop and take stock, the sooner we will realise there is more than one way to skin a cat. As David Graeber said, "The ultimate, hidden truth of the world is that it is something that we make, and could just as easily make differently". It’s not like we haven’t done our homework and we are more than smart enough to get our heads around the alternatives. The only obstacle may be the limited time we have left to sort it out.
Automation in business is for increasing profit margins and reducing worker costs, always has been. Not sure who fed you that utopian stuff, but this is capitalism, under capitalism automation isn't "for you", it's for profit.
@@jgt76 Capitalism's greatest triumph is convincing us that there is no other way. Read my last paragraph again. This not utopian, it's now basic survival.
@@jgt76 It's called efficiency and is what allows you and everyone else affordable access to computers, cell phones, the internet, food, and pretty much everything else you take for granted. You commies simply won't learn.
Hahaha - At the point there are frameworks that can do what you said in the end, they would just work for the companies... no one would hire a human at that point.
Oh boy. So many startups are going to be in trouble 🤦🏻♂️
peaty sure eventually even Microsoft office will be inside chatgpt or your go to photo/video/sound editor or ide.
There is zero chance that this is real, at least how you described it, which is ancient technology. I need AI to complete my tasks or to schedule and complete them based on an agentic job I gave it. What you described is like announcing now chatGPT can set alarms and keep an address book. It's the past not the future.
Breathing good ChatGPT hot line 😂 Now we know
Agentic does sound kind of weird. Maybe just call it agent.
Voice calendar with a reminder ok pretty sure that's been a thing for awhile now but yeah sweat stuff, way to use the jaws of life to open a can bud
ahh such a letdown!!! if it doesn't control a mouse and a keyboard they are already behind!
That's also coming, but it's a different thing
Wait, who is this young man podcasting?
CronGpt
Yo T-Bone
This is underwhelming, literally notion does better than what is being proposed in this video
Pos stop the Manipulative Narration Practices...
So model is too stupid to create tasks by prompting it and we have to use GUI?
Can it be my lover?
It could, but more likely to be your boss
microsoft got there first i believe
i can like the brits
🦾🦾🦾🦾🫵🤖🫵🦾🦾🦾🦾
Saitama look now huh 😁 is that your final form
Cron job
This is a slight gimick. Sam is trying to impress investors. If AIs Hallucinate, how can you trust it to do things?
Not impressed, scheduling software is older than the internet. A conversational AI should let me set up tasks using natural language.
A calender? Yawn...