iPhone Moment For AI Has Happened | Arvind Mathur | AWS | Simulted Reality | AIM TV

00:26:31
https://www.youtube.com/watch?v=3cEAwxayUkc

Zusammenfassung

TLDRThe episode explores the advancements in generative AI and its transformative potential in various industries. The discussion compares generative AI's rise to the advent of smartphones, emphasizing its consumerization before corporatization. Leaders are advised to support employees in leveraging AI and dispel fears regarding job security. The conversation addresses misconceptions related to data privacy and AI costs. Significant opportunities for AI in education and healthcare are also highlighted, predicting substantial improvements in personalized learning and patient care over the next decade.

Mitbringsel

  • 📱 Generative AI is compared to the iPhone moment, highlighting its rapid consumer acceptance.
  • ⚙️ Organizations need to integrate AI thoughtfully into their operations for maximum benefit.
  • 💼 Leaders should empower employees to leverage AI for growth rather than fear job loss.
  • 🔍 Misconceptions about AI include concerns over data privacy and cost-effectiveness.
  • 📈 AI has the potential to revolutionize education by offering personalized learning experiences.
  • 🏥 In healthcare, AI can lead to earlier disease detection and improved patient outcomes.

Zeitleiste

  • 00:00:00 - 00:05:00

    The emergence of generative AI is compared to the iPhone moment, highlighting how it has rapidly become integral in both personal and business contexts. The speaker emphasizes the shift from corporate to consumer use, with a call for organizations to adopt this technology.

  • 00:05:00 - 00:10:00

    Generative AI, akin to the internal combustion engine, needs a supportive environment including proper use cases and implementation strategies to be effective. Leaders are encouraged to think of the technology as just one part of a larger 'vehicle' that must be designed for specific purposes.

  • 00:10:00 - 00:15:00

    Corporate leaders are urged to empower employees to embrace generative AI as a tool that enhances career growth rather than a threat to job security. By leveraging existing experiences with AI, employees can find new opportunities within their roles, transforming traditional jobs into new positions within the evolving tech landscape.

  • 00:15:00 - 00:20:00

    Concerns about job displacement are common, especially among junior employees. However, the speaker argues that while generative AI may reduce the number of developers needed, the overall demand for software will significantly increase, presenting new job opportunities in the long run.

  • 00:20:00 - 00:26:31

    Myths about AI, including fears of data theft and high costs, often stem from misunderstandings about the technology's design and application. Leaders should recognize AI as a tool that, when used correctly, enhances efficiency rather than posing inherent risks.

Mehr anzeigen

Mind Map

Video-Fragen und Antworten

  • What significant changes have occurred in AI recently?

    Generative AI has become popular across various demographics, making it an essential topic in businesses, similar to the iPhone's impact.

  • How can organizations effectively implement generative AI?

    Organizations should focus on understanding the purpose of AI and tailor its use to specific business needs, while also considering the supporting structures required.

  • What common misconceptions about AI should industry leaders be aware of?

    Common misconceptions include the belief that AI steals data and that it is too expensive.

  • How can corporate leaders empower employees with generative AI?

    Leaders should encourage curiosity and experimentation with AI, highlighting its potential benefits for career growth.

  • Will AI lead to job displacement?

    While there may be fewer jobs needed per unit of code, the overall demand for software will likely increase.

Weitere Video-Zusammenfassungen anzeigen

Erhalten Sie sofortigen Zugang zu kostenlosen YouTube-Videozusammenfassungen, die von AI unterstützt werden!
Untertitel
en
Automatisches Blättern:
  • 00:00:00
    with generative ai ai has been
  • 00:00:04
    consumerized before it's been certied I
  • 00:00:06
    call it the iPhone moment for AI has
  • 00:00:09
    happened how and in which ways can
  • 00:00:11
    organizations go about implementing it
  • 00:00:14
    what has happened with generative AI or
  • 00:00:16
    llms in general is I almost call like
  • 00:00:18
    the invention of something as
  • 00:00:20
    significant as the internal combustion
  • 00:00:21
    engines how can corporate leaders go
  • 00:00:23
    about empowering their employees in the
  • 00:00:25
    best possible way that they are using
  • 00:00:27
    generative AI tools for their work
  • 00:00:29
    generative is just another in a line of
  • 00:00:33
    disruptive technologies that shake up
  • 00:00:35
    how our personalized and our business
  • 00:00:38
    lives
  • 00:00:38
    work as an industry leader yourself have
  • 00:00:41
    you ever heard this from Junior
  • 00:00:43
    employees colleagues that sir is our job
  • 00:00:45
    in danger because of
  • 00:00:48
    AI what are some of the common AI
  • 00:00:51
    misconceptions that industry leaders
  • 00:00:53
    should be aware of
  • 00:00:56
    [Music]
  • 00:01:07
    arvind welcome to this episode of
  • 00:01:09
    simulated Reality by a media house how
  • 00:01:11
    are you how have you been I'm doing
  • 00:01:13
    great exciting to be here in India
  • 00:01:15
    lovely to have you here I want to
  • 00:01:17
    straight away jump to the very first
  • 00:01:20
    question something that I myself have
  • 00:01:21
    had in my mind for quite some time now
  • 00:01:24
    artificial intelligence is not brand new
  • 00:01:26
    it's been there so much so that in the
  • 00:01:28
    '90s Steven Spielberg made a movie
  • 00:01:29
    called artificial intelligence what on
  • 00:01:31
    Earth happened in the last 2 three years
  • 00:01:34
    with AI and with generative AI that it
  • 00:01:36
    is in every which way the talk of the
  • 00:01:38
    top that is such an amazing question I
  • 00:01:40
    think we're all trying to figure this
  • 00:01:42
    out the I think what's really happened
  • 00:01:45
    is that with generative
  • 00:01:48
    ai ai has been
  • 00:01:51
    consumerized before it's been
  • 00:01:53
    corporatized and I explain what that
  • 00:01:54
    means I call it the iPhone moment for AI
  • 00:01:57
    has happened why because if you if you
  • 00:02:00
    go back 10 15 years before smartphones
  • 00:02:03
    were a thing a lot of us in corporate it
  • 00:02:06
    were trying to explain to Business
  • 00:02:09
    Leaders that there's an opportunity to
  • 00:02:10
    create consumer applications and engage
  • 00:02:12
    people Etc it wasn't quite cutting
  • 00:02:15
    through but when the iPhone happened
  • 00:02:17
    suddenly they were using it their kids
  • 00:02:18
    were using it their Grandmom were using
  • 00:02:20
    it they all had this gut feel oh my God
  • 00:02:22
    this is going to change my business and
  • 00:02:25
    they invested in it and built a lot of
  • 00:02:27
    apps and look how the world has changed
  • 00:02:29
    because of that
  • 00:02:30
    and I think same thing was has been
  • 00:02:32
    happening for the last 10 plus years
  • 00:02:34
    before the chat GPD moment happened that
  • 00:02:37
    we were all talking about big data and
  • 00:02:39
    this Ai and that and it was some
  • 00:02:42
    progress was happening but when chat gbd
  • 00:02:45
    dropped again it it consumerized it
  • 00:02:48
    everyone was using it the Business
  • 00:02:50
    Leaders the kids the grand moms and
  • 00:02:53
    again they got this strong sense of in
  • 00:02:57
    their gut that this is going to change
  • 00:02:58
    my business it's going to create
  • 00:03:00
    efficiency it's going to create consumer
  • 00:03:02
    experience improvements it's going to
  • 00:03:04
    create new business models and therefore
  • 00:03:06
    they just want to they want to move
  • 00:03:08
    forward with this everyone gets it you
  • 00:03:10
    don't have to explain it and there is
  • 00:03:13
    therefore this pull and demand for it in
  • 00:03:16
    businesses as well as in our personal
  • 00:03:18
    lives which is just unprecedented this
  • 00:03:20
    is one of those 15 once in a 15 years
  • 00:03:22
    things you mentioned the massive demand
  • 00:03:24
    for it in businesses as well Rising by
  • 00:03:27
    the Year by the month by the fortnite to
  • 00:03:29
    be honest
  • 00:03:30
    how and in which ways can organizations
  • 00:03:33
    go about implementing it in their
  • 00:03:35
    day-to-day culture in their work a day
  • 00:03:37
    culture in the best way
  • 00:03:39
    possible i' I'd like to use a analogy in
  • 00:03:42
    this case to explain I think how we
  • 00:03:44
    should think about about uh generative
  • 00:03:48
    AI what has happened with generative AI
  • 00:03:51
    or llms in general is I almost call like
  • 00:03:53
    the invention of something as
  • 00:03:55
    significant as the internal combustion
  • 00:03:57
    engine so an engine has been in ented
  • 00:04:00
    right and now you don't sit on an engine
  • 00:04:03
    and go somewhere right you need you need
  • 00:04:06
    an engine to have wheels and
  • 00:04:08
    transmission and gearbox and accelerator
  • 00:04:11
    and brakes and Lane assist and rear VI
  • 00:04:14
    mirror all of those
  • 00:04:15
    things so to your question I think what
  • 00:04:19
    what is important to understand is that
  • 00:04:21
    this fabulous new engine has been
  • 00:04:24
    created and the big tech companies are
  • 00:04:27
    investing a lot of money to build better
  • 00:04:29
    and and better and more powerful engines
  • 00:04:31
    but the engine by itself is not
  • 00:04:34
    sufficient and that's why we also see a
  • 00:04:36
    lot of these stories about the kind of
  • 00:04:38
    things going out of control for folks
  • 00:04:40
    because those are examples of people
  • 00:04:42
    just sitting on an engine and trying to
  • 00:04:43
    go somewhere it doesn't quite work that
  • 00:04:45
    way you need the rest of the of the
  • 00:04:48
    environment the chassis and the and the
  • 00:04:51
    and the control systems that create a
  • 00:04:54
    vehicle to to go somewhere another
  • 00:04:57
    aspect of the same analogy is you got to
  • 00:04:59
    be really clear about what are you
  • 00:05:00
    trying to do are you going to commute to
  • 00:05:02
    work or are you taking kids to the game
  • 00:05:05
    or are you trying to transport you know
  • 00:05:08
    50 people uh somewhere the engine that
  • 00:05:11
    you use the kind of chasc that you use
  • 00:05:14
    all of those things have to be fit for
  • 00:05:17
    purpose again one of the challenges we
  • 00:05:19
    see is people are trying to use the same
  • 00:05:23
    kind of engine to do EV all kinds of
  • 00:05:27
    purposes without fully thinking through
  • 00:05:29
    what
  • 00:05:30
    is a use case or the application needs
  • 00:05:33
    uh and therefore what what we end up
  • 00:05:35
    talking a lot with folks is is think
  • 00:05:38
    about the purpose first what are you
  • 00:05:40
    trying to do what engine with what cost
  • 00:05:44
    to Performance to weight ratios are best
  • 00:05:48
    for your situation and and what is the
  • 00:05:50
    rest of the environment that's needed to
  • 00:05:52
    make it work for your situation and
  • 00:05:54
    there'll be all of those there'll be a
  • 00:05:57
    million different kinds of vehicles that
  • 00:05:59
    that can we run with an internal
  • 00:06:00
    combustion engine and all of those have
  • 00:06:03
    to be created over time so so I that's
  • 00:06:06
    the piece which I think we've got to
  • 00:06:08
    figure out Beyond just the excitement
  • 00:06:11
    about the llm itself it's all of the
  • 00:06:12
    other stuff that puts it to use you've
  • 00:06:16
    been in a leading position for multiple
  • 00:06:17
    years now as a leader yourself I want to
  • 00:06:20
    ask you how can corporate leaders go
  • 00:06:22
    about empowering their employees in the
  • 00:06:25
    best possible way that they are using
  • 00:06:27
    generative AI tools for their work but
  • 00:06:29
    without the fear of job displacement see
  • 00:06:32
    the way I think of this and way I
  • 00:06:33
    encourage everyone to think about this
  • 00:06:35
    is generative AI is just another Inner
  • 00:06:40
    Line of disruptive technologies that
  • 00:06:43
    shake up how our personaliz and our
  • 00:06:45
    business lives work right and this is
  • 00:06:47
    not the first one it's not going to be
  • 00:06:48
    the last one there have been quite a few
  • 00:06:51
    the important thing really is to is to
  • 00:06:53
    figure out a way
  • 00:06:55
    to get people to recognize that this
  • 00:07:00
    if they figure out how to use it well
  • 00:07:02
    it's a huge Advantage for them this is
  • 00:07:05
    something that will drive their careers
  • 00:07:07
    forward and create new opportunities for
  • 00:07:09
    them uh but if if
  • 00:07:13
    they but also that this will also create
  • 00:07:17
    disruptions right so in an environment
  • 00:07:20
    like that which one do you want to be do
  • 00:07:22
    you want to be the person who takes
  • 00:07:24
    advantage of something like this and
  • 00:07:27
    uses this as a propellent for their for
  • 00:07:29
    the careers uh and opportunities for
  • 00:07:32
    themselves or not so therefore it's
  • 00:07:33
    important to be curious about this to
  • 00:07:37
    learn about it to experiment with it to
  • 00:07:39
    see what works what doesn't work uh and
  • 00:07:42
    kind of put yourself in a position where
  • 00:07:44
    your previous experience pregenerative
  • 00:07:47
    AI which is the kind of the new thing
  • 00:07:49
    now uh but your experiences before that
  • 00:07:52
    came into the scene becomes why someone
  • 00:07:55
    wants you to be part of the process to
  • 00:07:58
    leverage gentiv in that situation and
  • 00:08:00
    I'll take an example for that one of the
  • 00:08:01
    previous companies I work with uh we
  • 00:08:04
    were it was an insurance company and we
  • 00:08:05
    were transforming a customer service uh
  • 00:08:07
    Center this is pre gen even right uh and
  • 00:08:10
    we had a a customer service center that
  • 00:08:13
    supported calls coming in and we were
  • 00:08:16
    always struggling with that the number
  • 00:08:17
    of calls were far too many we couldn't
  • 00:08:19
    service all of them we were being
  • 00:08:20
    selective what only being able to
  • 00:08:22
    support our biggest
  • 00:08:25
    customers uh and even there we were not
  • 00:08:27
    really serving them in a way that was
  • 00:08:29
    sufficient so we put in an AI capability
  • 00:08:32
    to improve to first of all to help
  • 00:08:36
    answer the simpler questions
  • 00:08:37
    automatically and then for the more
  • 00:08:39
    complex one gives a lot of support to
  • 00:08:41
    the agents so they can answer those
  • 00:08:42
    questions more easily and in situation
  • 00:08:45
    like that it was interesting there were
  • 00:08:46
    there were a group of people who leaned
  • 00:08:48
    into this and said hey I know how to
  • 00:08:50
    train this thing I can help you pick the
  • 00:08:54
    relevant historical data that will help
  • 00:08:56
    train this model to be more effective uh
  • 00:08:59
    and those are the folks who therefore
  • 00:09:01
    got into the center of this they got
  • 00:09:03
    they learned a lot from it and I mean
  • 00:09:05
    there were a few of our customer support
  • 00:09:07
    agents who became chatbot trainers and
  • 00:09:11
    this is not just a onetime thing they
  • 00:09:12
    they were involved on a weekly basis
  • 00:09:14
    analyzing what kind of questions are
  • 00:09:16
    coming in which ones are being answered
  • 00:09:18
    uh early and and in a high quality way
  • 00:09:21
    which ones need more training what data
  • 00:09:22
    can we pull out and that became their
  • 00:09:24
    new career so it was just a great
  • 00:09:26
    example and as Leaders therefore it was
  • 00:09:29
    important for us to even though they
  • 00:09:32
    knew nothing about the AI to say you
  • 00:09:35
    know about the business process we want
  • 00:09:37
    you to learn how AI can make this
  • 00:09:40
    process better rather than bringing
  • 00:09:41
    someone from outside to do this to you
  • 00:09:44
    so we created the those opportunities
  • 00:09:46
    not everyone lean into it but the ones
  • 00:09:48
    that did got a lot of support and
  • 00:09:50
    training and then very importantly and
  • 00:09:52
    my message to leaders is when you do see
  • 00:09:55
    situations like that amplify it make an
  • 00:09:58
    example out of that so we actually went
  • 00:10:00
    we we did newspaper articles about this
  • 00:10:03
    individual who had progressed from being
  • 00:10:04
    a customer support agent to being a
  • 00:10:07
    chatbot trainer and that created this
  • 00:10:09
    energy around this this whole Trend and
  • 00:10:11
    many more people came forward and said
  • 00:10:13
    hey for your next project I want to be
  • 00:10:15
    the one who who plays a role like this
  • 00:10:17
    so I think that's really important for
  • 00:10:18
    leaders to take that position I that was
  • 00:10:20
    a very fascinating story especially
  • 00:10:22
    about the fact where we got to know
  • 00:10:24
    where somebody who has been working for
  • 00:10:25
    so many years then ended up training
  • 00:10:27
    chat Bots themselves great great story I
  • 00:10:30
    want to ask you because there's a
  • 00:10:31
    question that has come to me as well we
  • 00:10:33
    have talked about this question multiple
  • 00:10:34
    times as an industry leader yourself
  • 00:10:36
    have you ever heard this from Junior
  • 00:10:38
    employees colleagues that sir is our job
  • 00:10:40
    in danger because of
  • 00:10:43
    AI it is happening a lot of people are
  • 00:10:46
    having this in their mind even if
  • 00:10:47
    they're not vocalizing it but I want to
  • 00:10:49
    tell you a very personal story on this
  • 00:10:50
    one so my older son is in second year of
  • 00:10:54
    college learning computer science and
  • 00:10:56
    he's a geek he's he's he's trying to get
  • 00:10:58
    into programming
  • 00:11:00
    and we've had more than one occasions
  • 00:11:02
    where he said did I make a mistake going
  • 00:11:04
    into this field because I don't see a
  • 00:11:05
    future for myself ouch right so so here
  • 00:11:09
    is how I think about it okay and there's
  • 00:11:12
    a longer term view to this and there's a
  • 00:11:13
    shorter term view to it the longer the
  • 00:11:15
    the slightly longer term view to this I
  • 00:11:17
    see it is if you think about it software
  • 00:11:21
    development is
  • 00:11:22
    still like an artisanal craft okay every
  • 00:11:27
    piece of software that is created has
  • 00:11:29
    someone's personal energy and thinking
  • 00:11:33
    in it it is like we still making shoes
  • 00:11:36
    by hand is how the software industry
  • 00:11:39
    works the thing which I'm excited about
  • 00:11:42
    is that with generative Ai and coding
  • 00:11:46
    assistants
  • 00:11:47
    Etc this is the opportunity for software
  • 00:11:50
    development to go from an artisanal
  • 00:11:52
    craft to an
  • 00:11:54
    industrialized production model right
  • 00:11:58
    what that will do is that in the short
  • 00:12:00
    in in the immediate term you may think
  • 00:12:01
    oh my God that means massive job losses
  • 00:12:04
    but in there's another dimension this
  • 00:12:06
    which is the unit cost of software will
  • 00:12:08
    go down dramatically and if you look at
  • 00:12:11
    Industries and I've been a CIO for so
  • 00:12:12
    many years the reality is although we
  • 00:12:14
    have a lot of technology in
  • 00:12:16
    organizations I would say 50% of the
  • 00:12:18
    work still happens in Excel and emails
  • 00:12:20
    and slack messages right why is that the
  • 00:12:23
    case is because there is not sufficient
  • 00:12:27
    software actually out there that touches
  • 00:12:29
    all of the business scenarios and cases
  • 00:12:32
    and humans are still intervening and and
  • 00:12:34
    connecting the dots so many places what
  • 00:12:38
    reduced unit cost of software will do is
  • 00:12:41
    lead to an explosion in the demand for
  • 00:12:44
    software because it will be much easier
  • 00:12:45
    to create and even more importantly to
  • 00:12:47
    maintain and modify software to match
  • 00:12:50
    the evolving needs of businesses we are
  • 00:12:53
    not even scratching that surface of that
  • 00:12:55
    of that opportunity so yes the unit cost
  • 00:12:58
    will go down the number of software
  • 00:13:00
    developers needed per million lines of
  • 00:13:02
    code will go down but the number of
  • 00:13:04
    millions and billions of line of code
  • 00:13:06
    that will actually start existing will
  • 00:13:09
    explode and that will create a large
  • 00:13:12
    amount of employment as well I think so
  • 00:13:16
    I think you know it all depends on how
  • 00:13:18
    soon we get to that but there are these
  • 00:13:20
    two opposing forces one that's reducing
  • 00:13:22
    the number of people needed per unit per
  • 00:13:24
    per line of code but the number of lines
  • 00:13:27
    of code are exploding so in the long
  • 00:13:29
    long term I think there is a going to be
  • 00:13:30
    a huge demand for people to run these
  • 00:13:33
    industrialized production shops for
  • 00:13:35
    software and much more software to
  • 00:13:38
    happen so that's what makes me uh very
  • 00:13:41
    very optimistic about this this is about
  • 00:13:43
    software but in other fields as well
  • 00:13:45
    there's a similar CH challenge with
  • 00:13:47
    creative fields and people creating copy
  • 00:13:49
    for
  • 00:13:50
    advertising I think there'll be a lot
  • 00:13:52
    more personalized advertising out there
  • 00:13:55
    so we it remains to be seen how the
  • 00:13:57
    balance of this will be but I feel this
  • 00:13:59
    is just like many other innovations that
  • 00:14:01
    have happened in the past I think for
  • 00:14:03
    Humanity this is a positive force I hope
  • 00:14:07
    that the Elder son was also satisfied
  • 00:14:09
    with the answer no he was not he wasn't
  • 00:14:12
    not at all that's too far out I care
  • 00:14:15
    about the job scene two years from now
  • 00:14:17
    so I nobody can predict that we'll see
  • 00:14:19
    we'll see uh you mentioned the job scene
  • 00:14:21
    uh I want to ask you a question which
  • 00:14:23
    pertains to something that is talked
  • 00:14:24
    about so often and that is performance
  • 00:14:27
    review start starting from your first
  • 00:14:29
    interview to your 10th interview
  • 00:14:31
    performance review is something that is
  • 00:14:33
    talked about I want to ask you as an
  • 00:14:34
    industry leader yourself when industry
  • 00:14:36
    leaders are doing as simp in simple
  • 00:14:39
    words they're doing a performance review
  • 00:14:40
    of their employees in which Best Way can
  • 00:14:43
    this differentiation be done as to how
  • 00:14:46
    much of the performance results were AI
  • 00:14:48
    driven and how much of it was the
  • 00:14:50
    employee herself or himself okay my my
  • 00:14:54
    my thinking on this
  • 00:14:56
    is do you today in your performance
  • 00:14:59
    evaluation say so much of my performance
  • 00:15:01
    was because I had access
  • 00:15:04
    to Office Solutions true or laptops or a
  • 00:15:09
    phone or I had a mobile phone we don't
  • 00:15:13
    right to me AI is just another one of
  • 00:15:15
    those
  • 00:15:16
    tools so I don't think we need to
  • 00:15:19
    separate that out I don't think so what
  • 00:15:22
    matters end of the days at an individual
  • 00:15:25
    level what impact am I making on
  • 00:15:27
    business and the
  • 00:15:31
    smarter folks and the ones who will be
  • 00:15:34
    rewarded more are the ones who figur out
  • 00:15:36
    all of the tools that are available to
  • 00:15:37
    me and make good use of that at the
  • 00:15:39
    right time for the right
  • 00:15:41
    problems so end of the day it's still
  • 00:15:43
    what it boils down to is what impact am
  • 00:15:45
    I making not how much of that was
  • 00:15:48
    because of this tool or that tool this
  • 00:15:50
    just a tool you know what you said
  • 00:15:52
    instantly reminds me of this one quote
  • 00:15:54
    that I read on the first day that I
  • 00:15:56
    joined aim media house in one of the
  • 00:15:59
    desks there's this quote written where
  • 00:16:01
    your job will not be replaced by AI but
  • 00:16:05
    maybe somebody who uses AI better
  • 00:16:07
    absolutely I completely agree with that
  • 00:16:10
    absolutely I want to ask you what are
  • 00:16:12
    some of the common AI misconceptions
  • 00:16:15
    that industry leaders should be aware of
  • 00:16:18
    so and and maybe we'll focus on gen
  • 00:16:21
    because there's a lot of that going
  • 00:16:22
    around right now uh but for example one
  • 00:16:25
    of the biggest ones is that uh that
  • 00:16:30
    somehow this is going to steal your data
  • 00:16:32
    I think that is something which which
  • 00:16:34
    which is a which is a going around a lot
  • 00:16:37
    and and the reality is what I would go
  • 00:16:40
    back to is the conversation we had
  • 00:16:41
    earlier that end of the day this is an
  • 00:16:43
    engine and there's a whole set of other
  • 00:16:46
    supporting environment around that that
  • 00:16:48
    turns it into a usable vehicle right now
  • 00:16:53
    some of those Vehicles would be designed
  • 00:16:54
    in a way that that use that data but
  • 00:16:57
    that's what the chat gpds or the World
  • 00:16:59
    potentially could be right but the
  • 00:17:01
    reality is that that's not an inherent
  • 00:17:03
    feature of all AI it is how you design
  • 00:17:06
    your AI do you design your AI in a way
  • 00:17:09
    that every piece of data it observes and
  • 00:17:12
    is also uses to train itself that's not
  • 00:17:15
    the norm that's the exception right so
  • 00:17:18
    so that's that's one misconception which
  • 00:17:20
    is really holding people back right uh
  • 00:17:23
    if you
  • 00:17:24
    create if you deploy llms in your own
  • 00:17:28
    envir environment and you design that
  • 00:17:30
    environment in the way that it only uses
  • 00:17:32
    that to provide recommendation but does
  • 00:17:34
    not use that to somehow leak information
  • 00:17:37
    then it won't do that right uh and
  • 00:17:39
    therefore you need those uh those kind
  • 00:17:42
    of setups the Bedrock to create your
  • 00:17:46
    solution in a way that doesn't doesn't
  • 00:17:49
    make that possible so that's one which
  • 00:17:50
    is very commonly heard another very
  • 00:17:52
    common Mis uh sort of concern I hear is
  • 00:17:55
    that it's too expensive right and the
  • 00:17:58
    reason why that that happens is because
  • 00:18:00
    again we end up using the wrong engine
  • 00:18:03
    in the wrong car it is like putting the
  • 00:18:04
    engine for a truck in a small sports car
  • 00:18:08
    it doesn't fit of course it'll consume
  • 00:18:10
    too much fuel because you're just using
  • 00:18:13
    something uh that's not fit for purpose
  • 00:18:17
    so with llms there are so many llms now
  • 00:18:20
    with very different price performance
  • 00:18:22
    ratio equations so if you're trying to
  • 00:18:24
    do document summarization don't use the
  • 00:18:27
    most complex and EXP expensive uh llm
  • 00:18:30
    use one that is very lightweight and all
  • 00:18:34
    you need is the linguistic capabilities
  • 00:18:37
    not the inherent knowledge of the llm
  • 00:18:39
    right so those are some examples couple
  • 00:18:41
    of of of them which these are the two
  • 00:18:43
    biggest ones I come across all the time
  • 00:18:46
    you know uh in the last few days and
  • 00:18:49
    weeks I've got the chance to speak to a
  • 00:18:51
    lot of Industry leaders so many of them
  • 00:18:54
    working at AWS itself uh a couple of
  • 00:18:57
    days back itself in Bangalore I was
  • 00:18:58
    talking to your colleague at AWS ishit
  • 00:19:00
    yeah and we got to talking about a lot
  • 00:19:03
    of the very very fascinating stuff
  • 00:19:05
    fascinating work fascinating projects
  • 00:19:07
    that is going on hand inhand with AWS
  • 00:19:10
    and gen and other companies one of them
  • 00:19:12
    being F1 the latest car so you have had
  • 00:19:15
    about uh 9 months at AWS up until now so
  • 00:19:19
    if you can talk about some of those
  • 00:19:20
    really exciting projects that you've got
  • 00:19:21
    to know and some that are lined up as
  • 00:19:23
    well so uh I I I talked about the fact
  • 00:19:27
    that I spent some time in in Insurance
  • 00:19:29
    in the
  • 00:19:30
    past uh so I'll talk about that story we
  • 00:19:32
    were trying to to transform the whole
  • 00:19:34
    claims process and historically the way
  • 00:19:36
    it was done was you get these hospital
  • 00:19:38
    bills some people type it in then some
  • 00:19:42
    then a claims assessor figures out
  • 00:19:43
    whether this is payable or not six seven
  • 00:19:45
    years ago I I did this project to
  • 00:19:48
    transform that whole process and use
  • 00:19:51
    what you would call traditional AI to do
  • 00:19:53
    that okay OCR and uh models uh which
  • 00:19:57
    were doing mathematical statistical
  • 00:20:00
    models
  • 00:20:01
    mostly it was so difficult to do that at
  • 00:20:04
    that point in time because OCR needed to
  • 00:20:07
    be exactly designed for different
  • 00:20:10
    hospital builds and even a small place
  • 00:20:12
    like Singapore with relatively limited
  • 00:20:14
    number of hospitals we ran into hundreds
  • 00:20:16
    and probably thousands of different
  • 00:20:18
    templates that had to be maintained all
  • 00:20:19
    the time now with Gen we now working
  • 00:20:22
    with some insurance companies where
  • 00:20:24
    there's no need for a templatized
  • 00:20:25
    approach you take any hospital bill and
  • 00:20:28
    it can extract the relevant information
  • 00:20:31
    from them so this just a very simple
  • 00:20:33
    example of how things have improved
  • 00:20:35
    dramatically now with gen versus the
  • 00:20:38
    earlier ways we used to do these things
  • 00:20:40
    I mean I can just think of the top of my
  • 00:20:41
    head all the information that is that
  • 00:20:44
    gen is taking from that one hospital
  • 00:20:46
    bill and then converting it into Data
  • 00:20:48
    makes the followup meetings with the
  • 00:20:51
    doctors and the follow-up appointments
  • 00:20:52
    at the hospital is just so much easier
  • 00:20:53
    for the people dramatically easier
  • 00:20:55
    absolutely and the decision- making
  • 00:20:57
    process you can figure out which
  • 00:20:58
    hospital was this from what was the
  • 00:21:00
    diagnosis what treatment was done and
  • 00:21:02
    then you can provide that that
  • 00:21:04
    structured data to other models that
  • 00:21:07
    figure out whether this is a payable uh
  • 00:21:10
    uh sickness is a treatment covered is
  • 00:21:14
    there any exception Etc so it makes
  • 00:21:16
    Downstream processing so much easier now
  • 00:21:19
    with uh with this Arin before we almost
  • 00:21:22
    towards the end of the podcast before we
  • 00:21:23
    end I want to ask you one question that
  • 00:21:25
    is pretty similar to the first question
  • 00:21:28
    that I ask you that was about what has
  • 00:21:30
    happened in the last 2 three years how
  • 00:21:32
    did this happen but now I want to ask
  • 00:21:34
    you say if you want to look at the
  • 00:21:36
    future not very distant future slightly
  • 00:21:38
    near future the next half a decade the
  • 00:21:40
    next decade what roles what do you think
  • 00:21:43
    is possible how much can happen when it
  • 00:21:45
    comes to generative AI in the workplace
  • 00:21:48
    you you really called the next decade
  • 00:21:49
    short-term future half a decade half a
  • 00:21:53
    deade okay good see we we talked about
  • 00:21:56
    that this is the the iPhone moment for
  • 00:21:59
    AI right now think about that over the
  • 00:22:02
    years over the decade since iPhone came
  • 00:22:04
    out how has this changed our lives right
  • 00:22:07
    today you you connect with your friends
  • 00:22:10
    and family on the smartphone you order
  • 00:22:13
    your food you learn you um you know plan
  • 00:22:18
    your holidays everything it's changed
  • 00:22:21
    our lives we could not even imagine all
  • 00:22:23
    of that stuff when the phone was
  • 00:22:26
    launched right I think that's kind of
  • 00:22:28
    what will happen over the next half a
  • 00:22:29
    decade or decade like you ask me we
  • 00:22:32
    cannot even imagine right now but here
  • 00:22:33
    are a few things that that excite me
  • 00:22:35
    I'll take two which absolutely blow my
  • 00:22:38
    mind the first one is in the yeara of
  • 00:22:40
    Education okay I'm sure you will have
  • 00:22:43
    you will remember that teacher from your
  • 00:22:45
    fifth grade or seventh grade who was
  • 00:22:47
    really special who took an interest in
  • 00:22:49
    you understood you gave you exactly the
  • 00:22:52
    the stuff that that made a difference to
  • 00:22:54
    your learning and interest in that
  • 00:22:55
    subject and that's why you love whatever
  • 00:22:57
    that subject was math or
  • 00:22:59
    right but it's very rare why is it so
  • 00:23:02
    rare because education is a scale
  • 00:23:06
    problem right there are great teachers
  • 00:23:08
    but none of them have the capacity to
  • 00:23:10
    understand every student at that level
  • 00:23:12
    of granularity it just sometimes the
  • 00:23:14
    chemistry happens and that happens it's
  • 00:23:16
    special with Gen and the wayi is
  • 00:23:20
    developing I think every teacher can be
  • 00:23:23
    like that special teacher for you why
  • 00:23:27
    because through data and through
  • 00:23:29
    extraction of this understanding they'll
  • 00:23:32
    every teacher will be able to understand
  • 00:23:34
    what your special interests are what
  • 00:23:37
    your talents are what little nudges and
  • 00:23:40
    training input you need to develop in
  • 00:23:44
    that particular area imagine how that
  • 00:23:47
    will change the lives of students who
  • 00:23:49
    entering the education system now this
  • 00:23:51
    is a very special point I predict that
  • 00:23:55
    students kids who are entering the
  • 00:23:57
    system now will have a completely
  • 00:23:59
    different experience every teacher every
  • 00:24:02
    subject every grade they will have that
  • 00:24:04
    special experience and I believe that
  • 00:24:08
    human potential will completely
  • 00:24:09
    transform because people will be able to
  • 00:24:12
    go in the direction that they're really
  • 00:24:14
    good at where their natural abilities
  • 00:24:17
    are and they will have all of the
  • 00:24:19
    support and nourish nourishment that
  • 00:24:22
    that's needed to develop into the areas
  • 00:24:24
    that they're they're specially inclined
  • 00:24:25
    towards I think that's that's one area
  • 00:24:28
    which completely blows my mind right
  • 00:24:30
    another one is Healthcare yeah right
  • 00:24:32
    which we kind of touched on earlier as
  • 00:24:34
    well imagine today same challenge right
  • 00:24:36
    doctors just don't have the capacity to
  • 00:24:39
    understand all of the Dynamics and all
  • 00:24:40
    the history of every patient and they're
  • 00:24:42
    trying to see patient every 2 minutes so
  • 00:24:45
    if the doctor can very quickly
  • 00:24:48
    understand all of the history and be in
  • 00:24:50
    touch with every piece of research that
  • 00:24:52
    is coming out and be able to recommend
  • 00:24:55
    the right thing based on the latest
  • 00:24:57
    research for every
  • 00:25:00
    individual patient imagine the impact
  • 00:25:03
    this will have on Healthcare right so
  • 00:25:05
    two examples I think in the next 5 years
  • 00:25:07
    for sure 10 years
  • 00:25:10
    guaranteed Healthcare and education will
  • 00:25:12
    be a totally different space I mean it's
  • 00:25:13
    great that this is the second time that
  • 00:25:14
    you mentioned Healthcare I just want to
  • 00:25:16
    mention this one thing I was talking to
  • 00:25:18
    doctors very senior doctors from as at
  • 00:25:20
    the global partnership on artificial
  • 00:25:21
    intelligence in Delhi earlier this year
  • 00:25:24
    and they were showing us models and they
  • 00:25:26
    were showing us uh products that they
  • 00:25:27
    have created
  • 00:25:28
    where with the help of AI they are being
  • 00:25:31
    able to detect breast cancer at the
  • 00:25:33
    earliest of stages and they said the
  • 00:25:35
    earlier the detection the better the
  • 00:25:37
    chances of cure and they're doing that
  • 00:25:39
    in other forms of cancer as well right
  • 00:25:40
    now so yes healthare and and we can go
  • 00:25:43
    down this in a lot of uh drug research
  • 00:25:46
    is being improved uh patient histories
  • 00:25:49
    are being improved preventive healthare
  • 00:25:52
    is being improved so across the entire
  • 00:25:54
    spectrum of healthare I think this just
  • 00:25:56
    incredible opportunity incredible
  • 00:25:58
    opportunity indeed AWS themselves doing
  • 00:26:01
    such pathbreaking work in so many of
  • 00:26:02
    these industries Arin it was an absolute
  • 00:26:05
    absolute pleasure to talk to you today
  • 00:26:07
    to get your time thank you so much how
  • 00:26:09
    was the podcast for you wonderful
  • 00:26:10
    chatting with you thank you thank you so
  • 00:26:12
    much Arin guys do leave in the comments
  • 00:26:15
    how did you like this podcast episode of
  • 00:26:17
    simulated Reality by a media house this
  • 00:26:19
    is me your friend korak see you in the
  • 00:26:21
    next one
  • 00:26:25
    [Music]
Tags
  • Generative AI
  • AI Implementation
  • Job Displacement
  • Misconceptions
  • Corporate Leaders
  • Education
  • Healthcare
  • Technology
  • Employee Empowerment
  • AI Benefits