My Takeaways: Supervised AI: A system designed to give the right answer based on labeled data. Unsupervised AI: A system designed to group things based on similarities or patterns. Reinforcement AI: A system designed to learn by doing, like playing a video game and receiving feedback. Generative AI: A system designed to create plausible fakes, often by combining different systems. The conversation around AI has evolved from researchers focusing on theoretical capabilities to builders creating at-scale products, and now to users. The recent excitement around Transformers, which has been around for six years, is not an AI revolution but a design revolution. The shift to user experience design is significant and should be acknowledged. The user experience approach of the last decade was seamless correctness, delivering a good experience without making users worry about what's happening under the hood. The current approach is giving users a raw material for solving individual problems, encouraging creativity. This change marks a shift in design thinking, from solving theoretical problems to solving individual problems. Users are now using existing AI tools to solve their own problems in a creative way. This individual approach is different from building Enterprise-scale AI systems. The focus has shifted from correctness to usefulness for the individual. This change is huge and should be embraced. Automation can save time and effort, but it also raises questions about job displacement and income inequality. AI has the potential to democratize access to productivity tools and resources, but it also risks widening the digital divide. As individuals, we can use AI to our advantage, but we also need to be mindful of the social and ethical implications of its use. The development and deployment of AI requires collaboration and coordination across different sectors and stakeholders, including policymakers, industry leaders, and researchers. It's important to identify and understand the problems facing society. We need to take responsibility for our actions and their impact on society. Collaboration and cooperation between individuals, organizations, and governments are essential to solving society's problems. It's crucial to have a long-term perspective and focus on sustainable solutions. Addressing social, economic, and environmental issues requires a multi-disciplinary approach. It's crucial to involve all stakeholders in the problem-solving process, including marginalized communities and future generations. The solutions should consider the diversity of perspectives and interests, but also prioritize the common good. Technology can be a powerful tool to address societal problems, but it should be used responsibly and ethically. Solving society's problems requires a systemic change, including changes in values, norms, and institutions.
I've been a fan since I watched your Statistical Thinking playlist. I still remember watching those videos on repeat, and I was so inspired that I decided to make a presentation at my company with the same name. Your stage presence is incredible. You are beautiful, and you have an inner confidence that radiates outward. You walk onto the stage like a supermodel. Your energy makes me feel good, and I'm looking forward to seeing more of your creative work about AI and Stats in the future
I feel like this misses the fundamental issue that many people are having with AI in it's current state. On one hand, yes, you personally were able to do many more things out of your normal range, faster, but you also eliminated the use of a copywriter, editor and graphic designer/illustrator at the same time for a result that was less nuanced and rife with errors. We'll overlook the fact that the backbone of this technology is literally built off the backs of years of hard work and effort by creatives; but it's the commodification of creativity by people who don't really understand its purpose that is leaving a sour taste in peoples mouths.
Fun fact: I have always made 100% of my content myself for years, no editors or copywriters or illustrators. It was never about the money - having other folks work on my presentations would have always taken too much time since I'm pretty efficient at it already thanks to learning how to do it during grad school and I accept any errors, which would have been mine either way. But now I get some extra time to spend at the gym.
@@kozyrkov kuka robotic from Jerman has succeeded in producing robots capable of making textiles. but the price of this robot is still more expensive than the workforce in Bangladesh. Bangladesh's textile industry employs 45% of the workforce. and generates 80% of exports for the country. when the price of robots is cheaper than these workers. we will first see a country go out of business.(bankrupt)
Sorry, angry post deleted.@@kozyrkov, can you give us some insight in the kind of jobs that AI cannot handle or might create, so people know in what direction to upskill? Most people need a job to survive and there needs to be some perspective or light at the end of the tunnel. Without hope, many people will give up and/or vote for anti-system politicians in addition to financial hardship, poverty, deaths of despair and an unemployment rate akin to the Great Depression (speaking of anti-system politicians...). For now, everyone's plan seems to be to become a plumber, which I don't think is what you had in mind.
encouraging to hear Google's chief decision scientist re-framing this as a design revolution for individual user experience and calling on leaders to step up to prevent the unequal distribution of powerful technologies
It is difficult to see reasonable points about AI. There is so much noise around, from either people that satanizes or idolize it, and off course, the sci-fi lovers. Your talk was so refreshing! A coherent narrative around AI, that try to make sense of the changes across time, and places the importance of using technology to solve REAL human problems. 💓
Hi Cassie! I was there at Web Summit Rio seeing your Talk and it was amazing! For me your talk was one of the best talks from the whole event! Congratulations!
Many years ago I was a software manager in a medium sized tech firm. I had a director who was quite a practical guy, and quite talented too. I remember describing a problem that we had with our programmers, and how I thought that presenting improvement courses would help with the problem. His response? "Motherhood. You're describing motherhood. That doesn't work in industry, the employees will always do the minimum, you will waste your breath and your time." He was right. And sadly, what Cassie is asking for here is motherhood. Laudable, but unfortunately extremely unlikely. Individuals will draw whatever benefit they can, regardless of the impact on others.
I really love how you can hear, see and feel the excitement of Cassie when she Talks about the current development and shows her Expertise in the Talk without being over complicated. 😃
as an artist myself, this video crystalizes what i find kind of boring about AI/LLMs: a huge part of being an artist is the act of creating. going through the process of discovery. when i pick up my guitar and fiddle around, i go through an entire journey of playing melodies, riffs, chord progressions, etc. and shaping that into a song. then playing that with other musicians adds an entirely new dimension to that (not to mention improvising live in front of an audience). i don't really get any of that with AI. i just type prompts into a bot: "give me a tab for a song like 'enter sandman'" or whatever and it does it for you. honestly, where's the fun in that? but, of course, silicon valley is about reducing everything to Efficiency and Productivity. the act of creation? an inefficient waste of time. must increase Content™ productivity for the algorithm. grim shit!
Cassie, great to see you in the centre stage of websummit! For me, it was the best presentation of websummit Rio. Congratulations on your amazing capacity of making a complicated topic into a delightful talk
I agree we must keep up with new technology. This doesn’t not equate let them make us unemployed. It’s easy to apply these softwares to automate many jobs (which would probably be dishumanized and maybe qualitatively worse, but performed at super speed and very cheaply). Previous technological revolutions won’t compare to AI. They slowly replaced very few jobs and actually immediately created many new ones and with higher income. Now it’s easy to see most knowledge-based jobs severely impacted or replaced, with very uncertain (and probably unpleasant) new jobs. People have been given real problems with very vague, naïve solution in sight. We’re dealing with people jobs, not leisure time! One will say. All this in the name of creativity and enhanced productivity: how many ordinary people in the world have asked for them? Very, very few I guess.
What an amazing talk! I completely agree that in the medium term, inequality of access is likely the biggest issue we will face. "Haves" and "Have Nots" of AI is a very scary concept that spans geography, age, and cultural categories
I think the question people aren’t asking is how will this affect commerce, because that is where we get jobs. Will businesses accelerate productivity and earn more, or will they do like Twitter and get leaner? Although getting lean is a much more difficult job than we’d think.
The point reminds me of just how important and unrecognized Alpha Fold is from Demis Hasabis of Google Deepmind. 3D Modeling of a single protein took a PHD 5 years. Alpha fold predicative modeling has logged 200 million protein predictions representing billions of PHD hours of work which as Cassie mentioned, is making what was impossible possible. More importantly, it was and is open source, meaning that the medical community is now free to tinker with the results allowing for a far greater scope of innovation in medical science.
im a programmer, and i train and develop my own models. im glad to see people learning more about ai and why its taking off so much right now. i would consider automated instructions an umbrella term for all types of models, automated instructions may not exactly fit for some but same with artificial intelligence. theres some issues we need to work on, the first we should focus on it letting ai be open and public to all, no elon musk trying to stop development, no governments trying to stop models from being closed source for the public. theres been a growing surge in developers running ai locally, on your own hardware instead of massive datacenters and we are slowly finding ways to make open source local ai solutions for all devices, for everyone, for free. ai should be public!
After looking at your bard example, I beleive, it's only a matter of time before AI finds its way into meeting tools and, at the end of a meeting, highlights the minutes of the meeting and even makes recommendations. Wouldn't that be an interesting use case?
Google is getting salty and it's understandable. They released the first paper on the transformer, attention network, an revolutionary LLM. But now OpenAI screwed with their modified weapon. 😂😂. Its Funny
Interesting that despite using AI she didn't manage to generate spoken content that responds to the question in the title :) : "whose jobs does AI automate?" . But very good speech nonetheless. I loved the heartfelt moralistic conclusion especially. It's completely understandable that sensitive issues should be only glossed over and not spoken about straightforwardly.
Google may be left in the dust here, Microsoft has just embedded chatGPT into its office productivity suite, does that give Microsoft end users an advantage over Google suite users because ChatGPT is so much better?
I love this part - "Just try it" so true! Instead of explaining away concepts of language models, I simply told my friends/fam- try it. So funny to see their mouth drop. 😮 All tech evangelists should embrace this design revolution. Come to think of it, gospel evangelists should say this too! Just try Jesus.😅
OK, I'm going to be honest. This video just sort of popped up as are many AI related videos these days. When this woman strutted out on the stage like a wanna-be super model, I thought it was a joke. I thought the background was a green screen and she was pretending to be giving a talk to a large audience. But, I hung in there. And, by the end, I was giving her a standing ovation. As a software developer, I've been using AI daily now -- to help me build AI-powered tools. I've always seen it as an instrument to increase my productivity. I've always known it's not "intelligence", although I will admit to having told it "Thank you." on more than one occasion. And, I've never heard anyone else say this, but I totally agree: I hope this technology is used to help solve our societal problems. Bravo. Well done.
Good talk but disappointed that she didn't answer the question in the title. Machine-learning has not displaced software-developers. First of all it is not applicable everywhere as her mentioned 6,5 hour course points out - in the course she forgets to point out that machine learning is unnecessarily complicated or worse not just when you can look up the answer, BUT ALSO if it's necessary to get a deterministic answer (by an algorithm) which is almost always the case in software - and also because applied machine-learning needs a lot of classical software-development around it for collecting data, training the model and using the generated model (video suggestions can be done much better by machine learning than handcrafted algorithms, and finding cat photos is impossible without machine-learning, but these models don't work on their own). EDIT: She knows all this but her audience does not. What I expected from the title was a video about her views on what jobs will get automated by the new "personalized" AI.
As an aspiring computer scientist, I found this video very informative. What career advice do you have for aspiring computer scientists? Plus How can I still be useful to society? So my pockets don't starve to death.
Honestly, we should probably start thinking about changing the economic model in which we live. As long as you need a job to afford existing, there will be a conflict between people and automation. Not just CS but layers, doctors, and other types of specialists will be out of a job in the blink of an eye.
@@mattgray666 If you really want a complete solution, it's called Central planning. UBI just ensures you stay alive enough to be a good little consumer fire the owner class.
kuka robotic from Jerman has succeeded in producing robots capable of making textiles. but the price of this robot is still more expensive than the workforce in Bangladesh. Bangladesh's textile industry employs 45% of the workforce. and generates 80% of exports for the country. when the price of robots is cheaper than these workers. we will first see a country go out of business.(bankrupt)
so how do we survive in the future or near future? How do we feed our families cause some rich people decided its better for them to make money from people using AI?
Sounds like she is mixing Automated Instructions with Artificial Intelligence 🤣 obviously people weren't too exited about AI until now because ordinary people couldn't use it directly for anything.
Cassie has no idea for what's coming, this whole presentation was hocus pocus and very underwhelming without anything substantial to what really is happening in the field.
@@nickwilson7241oh yeah like the other guy that works on AI at Meta and talks non sense on Twitter all the time. Also if you think that “working” at Google is some kind of validation and not to be skeptical at all you are in for a rude awakening. Same as the other guy who worked at Google and said the AI was sentient. Now more to the objective facts, first her presentation, absolute low quality, nothing substantial, no argument backed by any data or evidence. Second the message was at most just a surface level opinion with no backed proof. Third, was she part of any AI paper on Google, go do your homework boy. What about the transformers models did she have any input? Fourth, she doesn’t even have a PhD on any AI related field. At best she is just a data scientist. Now before you follow like a sheep have the decency of having a brain and follow through on anything an “expert” says. Cheers
I think that there should be another step in your scheme between the input(data) and the instructions (machine’s language), which is the process that is done by the AI algorithms. These technology make it is easier to communicate with machines by using ways which human use to learn and apply knowledge. Supervised- get the correct answer. This is something that involves critical thinking, evaluation and analysis. “Finds the common patterns in the data and turn those into instructions” - humans make connections between the things they are exposed too and learn how to use/apply them. Reinforcement - Learn by Trial and error. In other words, this systems design to function in away that is similar to the way that humans learn - recognizing patterns and making connections, making decisions baes on evaluation (in a very general way), upgrading our skills and developing ourself by learning from our failures and success. Computers now can learn and apply knowledge (information,data) in a way that is similar to humans, in a much faster way and in a larger scale. Furthermore, These technology is being developed and upgraded in a faster rate than the ability of humans to process it. Yes, it can accelerate our productivity, but it also can out succeed us.
Cassie has an incredible way of making the complex simple regarding Automated Instructions (AI).
She doesn't need you hitting on her bro
My Takeaways:
Supervised AI: A system designed to give the right answer based on labeled data.
Unsupervised AI: A system designed to group things based on similarities or patterns.
Reinforcement AI: A system designed to learn by doing, like playing a video game and receiving feedback.
Generative AI: A system designed to create plausible fakes, often by combining different systems.
The conversation around AI has evolved from researchers focusing on theoretical capabilities to builders creating at-scale products, and now to users.
The recent excitement around Transformers, which has been around for six years, is not an AI revolution but a design revolution.
The shift to user experience design is significant and should be acknowledged.
The user experience approach of the last decade was seamless correctness, delivering a good experience without making users worry about what's happening under the hood.
The current approach is giving users a raw material for solving individual problems, encouraging creativity.
This change marks a shift in design thinking, from solving theoretical problems to solving individual problems.
Users are now using existing AI tools to solve their own problems in a creative way.
This individual approach is different from building Enterprise-scale AI systems.
The focus has shifted from correctness to usefulness for the individual.
This change is huge and should be embraced.
Automation can save time and effort, but it also raises questions about job displacement and income inequality.
AI has the potential to democratize access to productivity tools and resources, but it also risks widening the digital divide.
As individuals, we can use AI to our advantage, but we also need to be mindful of the social and ethical implications of its use.
The development and deployment of AI requires collaboration and coordination across different sectors and stakeholders, including policymakers, industry leaders, and researchers.
It's important to identify and understand the problems facing society.
We need to take responsibility for our actions and their impact on society.
Collaboration and cooperation between individuals, organizations, and governments are essential to solving society's problems.
It's crucial to have a long-term perspective and focus on sustainable solutions.
Addressing social, economic, and environmental issues requires a multi-disciplinary approach.
It's crucial to involve all stakeholders in the problem-solving process, including marginalized communities and future generations.
The solutions should consider the diversity of perspectives and interests, but also prioritize the common good.
Technology can be a powerful tool to address societal problems, but it should be used responsibly and ethically.
Solving society's problems requires a systemic change, including changes in values, norms, and institutions.
I've been a fan since I watched your Statistical Thinking playlist. I still remember watching those videos on repeat, and I was so inspired that I decided to make a presentation at my company with the same name. Your stage presence is incredible. You are beautiful, and you have an inner confidence that radiates outward. You walk onto the stage like a supermodel. Your energy makes me feel good, and I'm looking forward to seeing more of your creative work about AI and Stats in the future
I feel like this misses the fundamental issue that many people are having with AI in it's current state. On one hand, yes, you personally were able to do many more things out of your normal range, faster, but you also eliminated the use of a copywriter, editor and graphic designer/illustrator at the same time for a result that was less nuanced and rife with errors. We'll overlook the fact that the backbone of this technology is literally built off the backs of years of hard work and effort by creatives; but it's the commodification of creativity by people who don't really understand its purpose that is leaving a sour taste in peoples mouths.
Fun fact: I have always made 100% of my content myself for years, no editors or copywriters or illustrators. It was never about the money - having other folks work on my presentations would have always taken too much time since I'm pretty efficient at it already thanks to learning how to do it during grad school and I accept any errors, which would have been mine either way. But now I get some extra time to spend at the gym.
@@kozyrkov It's not AI that replaces humans. but 1 human who uses AI replaces 10 humans who do not use AI
@@kozyrkov all repetitive work using computers can be automated within 2 years by ai.
@@kozyrkov kuka robotic from Jerman has succeeded in producing robots capable of making textiles.
but the price of this robot is still more expensive than the workforce in Bangladesh.
Bangladesh's textile industry employs 45% of the workforce. and generates 80% of exports for the country. when the price of robots is cheaper than these workers. we will first see a country go out of business.(bankrupt)
Sorry, angry post deleted.@@kozyrkov, can you give us some insight in the kind of jobs that AI cannot handle or might create, so people know in what direction to upskill? Most people need a job to survive and there needs to be some perspective or light at the end of the tunnel. Without hope, many people will give up and/or vote for anti-system politicians in addition to financial hardship, poverty, deaths of despair and an unemployment rate akin to the Great Depression (speaking of anti-system politicians...). For now, everyone's plan seems to be to become a plumber, which I don't think is what you had in mind.
encouraging to hear Google's chief decision scientist re-framing this as a design revolution for individual user experience and calling on leaders to step up to prevent the unequal distribution of powerful technologies
we'll always have the GOP to tell us every turd laid by Wall Street is a rose
It is difficult to see reasonable points about AI. There is so much noise around, from either people that satanizes or idolize it, and off course, the sci-fi lovers. Your talk was so refreshing! A coherent narrative around AI, that try to make sense of the changes across time, and places the importance of using technology to solve REAL human problems. 💓
Hi Cassie! I was there at Web Summit Rio seeing your Talk and it was amazing! For me your talk was one of the best talks from the whole event! Congratulations!
Many years ago I was a software manager in a medium sized tech firm. I had a director who was quite a practical guy, and quite talented too. I remember describing a problem that we had with our programmers, and how I thought that presenting improvement courses would help with the problem.
His response?
"Motherhood. You're describing motherhood. That doesn't work in industry, the employees will always do the minimum, you will waste your breath and your time."
He was right. And sadly, what Cassie is asking for here is motherhood. Laudable, but unfortunately extremely unlikely. Individuals will draw whatever benefit they can, regardless of the impact on others.
I really love how you can hear, see and feel the excitement of Cassie when she Talks about the current development and shows her Expertise in the Talk without being over complicated. 😃
I can't believe you were here in Brazil, how amazing, I hope you enjoyed Rio
as an artist myself, this video crystalizes what i find kind of boring about AI/LLMs: a huge part of being an artist is the act of creating. going through the process of discovery. when i pick up my guitar and fiddle around, i go through an entire journey of playing melodies, riffs, chord progressions, etc. and shaping that into a song. then playing that with other musicians adds an entirely new dimension to that (not to mention improvising live in front of an audience). i don't really get any of that with AI. i just type prompts into a bot: "give me a tab for a song like 'enter sandman'" or whatever and it does it for you. honestly, where's the fun in that? but, of course, silicon valley is about reducing everything to Efficiency and Productivity. the act of creation? an inefficient waste of time. must increase Content™ productivity for the algorithm. grim shit!
Cassie, great to see you in the centre stage of websummit! For me, it was the best presentation of websummit Rio. Congratulations on your amazing capacity of making a complicated topic into a delightful talk
I agree we must keep up with new technology. This doesn’t not equate let them make us unemployed.
It’s easy to apply these softwares to automate many jobs (which would probably be dishumanized and maybe qualitatively worse, but performed at super speed and very cheaply). Previous technological revolutions won’t compare to AI. They slowly replaced very few jobs and actually immediately created many new ones and with higher income. Now it’s easy to see most knowledge-based jobs severely impacted or replaced, with very uncertain (and probably unpleasant) new jobs. People have been given real problems with very vague, naïve solution in sight. We’re dealing with people jobs, not leisure time! One will say. All this in the name of creativity and enhanced productivity: how many ordinary people in the world have asked for them? Very, very few I guess.
What an amazing talk! I completely agree that in the medium term, inequality of access is likely the biggest issue we will face. "Haves" and "Have Nots" of AI is a very scary concept that spans geography, age, and cultural categories
Thank you soo much Cassie Ma'am. Absolutely Loved your talk!
Let's Solve for Society🙏
Can you please make a video(s) where you apply these 12 steps of ML in a real case study/dataset or maybe a blog ? Thanks Cassie
So beautifully stated. Thank you for the talk !
15:57 and that is exactly the problem, hun.
Wow you looked so fantastic on stage, Cassie!
I think this talk elegantly points out the paradigm shift we all currently experiencing.
I think the question people aren’t asking is how will this affect commerce, because that is where we get jobs. Will businesses accelerate productivity and earn more, or will they do like Twitter and get leaner? Although getting lean is a much more difficult job than we’d think.
Well, someone is still needed to build the NNet that allows the "recipe" to learn by examples isn't it ?
Soon we'll have AI that helps build other AI.b
The point reminds me of just how important and unrecognized Alpha Fold is from Demis Hasabis of Google Deepmind. 3D Modeling of a single protein took a PHD 5 years. Alpha fold predicative modeling has logged 200 million protein predictions representing billions of PHD hours of work which as Cassie mentioned, is making what was impossible possible. More importantly, it was and is open source, meaning that the medical community is now free to tinker with the results allowing for a far greater scope of innovation in medical science.
Great talk and reflexions. Thank you
im a programmer, and i train and develop my own models. im glad to see people learning more about ai and why its taking off so much right now. i would consider automated instructions an umbrella term for all types of models, automated instructions may not exactly fit for some but same with artificial intelligence. theres some issues we need to work on, the first we should focus on it letting ai be open and public to all, no elon musk trying to stop development, no governments trying to stop models from being closed source for the public. theres been a growing surge in developers running ai locally, on your own hardware instead of massive datacenters and we are slowly finding ways to make open source local ai solutions for all devices, for everyone, for free. ai should be public!
That was a really good video....I am going to share.
Good point: to solving society's problems, not to developing society's problems.👍
Insightful * Enjoyable = Cassie video
After looking at your bard example, I beleive, it's only a matter of time before AI finds its way into meeting tools and, at the end of a meeting, highlights the minutes of the meeting and even makes recommendations. Wouldn't that be an interesting use case?
Microsoft recently announced a tool that does just that: ruclips.net/video/rLC2frnUasw/видео.html
That is already a thing.
The AI is called Fireflies.
@@A_Box And if you are after something enterprise grade, teams premium will do it all inside the Microsoft 365 environment, including video.
Yup, there are a few of those apps around already.
Microsoft Teams is bringing it
I wish I was there 👏
Google is getting salty and it's understandable. They released the first paper on the transformer, attention network, an revolutionary LLM. But now OpenAI screwed with their modified weapon. 😂😂. Its Funny
19 minutes of talking and promoting Bard without even answering the question in the title.
Interesting that despite using AI she didn't manage to generate spoken content that responds to the question in the title :) : "whose jobs does AI automate?" . But very good speech nonetheless. I loved the heartfelt moralistic conclusion especially. It's completely understandable that sensitive issues should be only glossed over and not spoken about straightforwardly.
If I do not report for work tomorrow, how will AI automate my job when I am 10,000 miles away from US mainland?
Google may be left in the dust here, Microsoft has just embedded chatGPT into its office productivity suite, does that give Microsoft end users an advantage over Google suite users because ChatGPT is so much better?
I am sorry but I don’t think that the name of this presentation does represent the content shared
I love this part - "Just try it" so true!
Instead of explaining away concepts of language models, I simply told my friends/fam- try it. So funny to see their mouth drop. 😮 All tech evangelists should embrace this design revolution. Come to think of it, gospel evangelists should say this too! Just try Jesus.😅
Google had a "kodak" moment. They lost AI search to Microsoft and OpenAI.
@@alpha0xide9 Oh damnnn. Did I really write OpenCV?? 😀 Sorry bruh
Nice
OK, I'm going to be honest. This video just sort of popped up as are many AI related videos these days. When this woman strutted out on the stage like a wanna-be super model, I thought it was a joke. I thought the background was a green screen and she was pretending to be giving a talk to a large audience. But, I hung in there. And, by the end, I was giving her a standing ovation. As a software developer, I've been using AI daily now -- to help me build AI-powered tools. I've always seen it as an instrument to increase my productivity. I've always known it's not "intelligence", although I will admit to having told it "Thank you." on more than one occasion. And, I've never heard anyone else say this, but I totally agree: I hope this technology is used to help solve our societal problems. Bravo. Well done.
Good talk but disappointed that she didn't answer the question in the title. Machine-learning has not displaced software-developers. First of all it is not applicable everywhere as her mentioned 6,5 hour course points out - in the course she forgets to point out that machine learning is unnecessarily complicated or worse not just when you can look up the answer, BUT ALSO if it's necessary to get a deterministic answer (by an algorithm) which is almost always the case in software - and also because applied machine-learning needs a lot of classical software-development around it for collecting data, training the model and using the generated model (video suggestions can be done much better by machine learning than handcrafted algorithms, and finding cat photos is impossible without machine-learning, but these models don't work on their own).
EDIT: She knows all this but her audience does not.
What I expected from the title was a video about her views on what jobs will get automated by the new "personalized" AI.
As an aspiring computer scientist, I found this video very informative. What career advice do you have for aspiring computer scientists?
Plus How can I still be useful to society? So my pockets don't starve to death.
Honestly, we should probably start thinking about changing the economic model in which we live. As long as you need a job to afford existing, there will be a conflict between people and automation. Not just CS but layers, doctors, and other types of specialists will be out of a job in the blink of an eye.
@@A_Box The solution already on the discussion table is Universal Basic Income. Perhaps the future will not be about existing, but being relevant?
@@mattgray666 If you really want a complete solution, it's called Central planning. UBI just ensures you stay alive enough to be a good little consumer fire the owner class.
Every job
for some time , the way she walked , the way she inroduced though she was an AI model
all repetitive work using computers can be automated within 2 years by ai.
kuka robotic from Jerman has succeeded in producing robots capable of making textiles.
but the price of this robot is still more expensive than the workforce in Bangladesh.
Bangladesh's textile industry employs 45% of the workforce. and generates 80% of exports for the country. when the price of robots is cheaper than these workers. we will first see a country go out of business.(bankrupt)
q) whose job does AI automate?
ans) yes. google made GPT first
so how do we survive in the future or near future? How do we feed our families cause some rich people decided its better for them to make money from people using AI?
Sounds like she is mixing Automated Instructions with Artificial Intelligence 🤣
obviously people weren't too exited about AI until now because ordinary people couldn't use it directly for anything.
Need a green card ❤
so many words.........and no answer to the question.........waste of time.
Play it again from 2:50-2:59.
Please reprocess audio, this is horribly volatile.
AI automated the sound technician
@@ltx4096😂
Dora's job was called as a comptist - in India for instance.
Worst crowd ever.
Cassie has no idea for what's coming, this whole presentation was hocus pocus and very underwhelming without anything substantial to what really is happening in the field.
She works in AI at Google lmao
What are your industry credentials?
@@nickwilson7241oh yeah like the other guy that works on AI at Meta and talks non sense on Twitter all the time. Also if you think that “working” at Google is some kind of validation and not to be skeptical at all you are in for a rude awakening. Same as the other guy who worked at Google and said the AI was sentient.
Now more to the objective facts, first her presentation, absolute low quality, nothing substantial, no argument backed by any data or evidence. Second the message was at most just a surface level opinion with no backed proof.
Third, was she part of any AI paper on Google, go do your homework boy. What about the transformers models did she have any input?
Fourth, she doesn’t even have a PhD on any AI related field. At best she is just a data scientist.
Now before you follow like a sheep have the decency of having a brain and follow through on anything an “expert” says.
Cheers
strong independent woman
Pure Google hype. Didn't answer the question of the presentation at all. Basically "but we invented it first, love google"
I think that there should be another step in your scheme between the input(data) and the instructions (machine’s language), which is the process that is done by the AI algorithms. These technology make it is easier to communicate with machines by using ways which human use to learn and apply knowledge.
Supervised- get the correct answer.
This is something that involves critical thinking, evaluation and analysis.
“Finds the common patterns in the data and turn those into instructions” - humans make connections between the things they are exposed too and learn how to use/apply them.
Reinforcement - Learn by Trial and error.
In other words, this systems design to function in away that is similar to the way that humans learn - recognizing patterns and making connections, making decisions baes on evaluation (in a very general way), upgrading our skills and developing ourself by learning from our failures and success.
Computers now can learn and apply knowledge (information,data) in a way that is similar to humans, in a much faster way and in a larger scale. Furthermore, These technology is being developed and upgraded in a faster rate than the ability of humans to process it.
Yes, it can accelerate our productivity, but it also can out succeed us.
Insightful * Enjoyable = Cassie video