Major breakthroughs in artificial intelligence research often reshape the design and utility of AI in both business and society. In this episode of Smart Talks with IBM, Malcolm Gladwell and Jacob Goldstein explore the conceptual underpinnings of modern AI with Dr. David Cox, VP of AI Models at IBM Research. They talk foundation models, self-supervised machine learning, and the practical applications of AI and data platforms like watsonx in business and technology.
Visit us at: https://www.ibm.com/thought-leadership/smart/talks/
Learn more about watsonx: https://www.ibm.com/watsonx
This is a paid advertisement from IBM.
See omnystudio.com/listener for privacy information.
This is an unofficial transcript meant for reference. Accuracy is not guaranteed.
The welcome to textile production from I heart. Radio
The today we are witness to one of those rare moments in history: the rise of an innovative technology with the potential to radically transform business and society forever. That technology, of course, is artificial intelligence and it's the central focus for this new season of smart talks with idea join hosts from your favorite pushkin podcasts, as they talk with industry experts and leaders to explore how businesses can integrate ai into their workflows and help drive real change in this new era of ai and, of course, host malcolm gladwell will be there to guide you through the season and throw in his two cents as well. Look out for new episodes of smart talks with I bm every other week on the iheart radio, app apple podcast.
Wherever you get your podcasts and learn more at I b m dot com, slash smart talks, hello, hello, welcome to smart talks with I b m a podcast from pushed
industries, Iheart, radio and ibm Malcolm harbour,
This isn't worth continuing our conversation with new creators vision.
It's? U, r creatively applying technology in business to drive change, but with a focus on the transformative power
rather artificial intelligence and what it means:
still leverage ay I
changing multiplier for your business. A guest today is doctor David Cox, vp of ay models at ibm, research and ibm d.
Two of the mit ibm watson s island, a first of its kind industry, academic collaboration between ibm mit focused on the fundamental research of artificial intelligence over the course of decades.
David Cox watched as the aid. I revolution steadily grew.
from the simmering ideas of a few academics and technologists into the industrial boom we are experiencing today.
Having dedicated his life to pushing the field of aid towards new horizons. David is both contributed to and presided over many of the major breakthroughs in artificial intelligence. Today's episode, you'll, hear David, explain some of the conception
underpinnings of the current, a landscape things like foundation models. In surprisingly, comprehensible terms, I might add,
will also get into some of the amazing practical applications for a business as well as what implicate
I will have for the future of work and design. David spoke with Jacob goldstein host of the Pushkin podcast. What's your problem?
a veteran business journalist, Jacob S, reported for the wall street journal
the miami Harold and was a
tat most of the npr programme. Planet, money. Ok, let's get to the interview,
tell me about your job idea, so I were too
I said I b m. Ah so one I'm the I b m director of the mit I b m watson, ai lab. So that's a joint lab between
ibm at mit, where we re trying to invent. What's next and I've been running for about five years,
and then more recently, I started as the vice president for I models and I'm in town
the building ibm foundation models of you, no building these fifty miles down models that allow us to have all kinds of new, exciting capabilities and air. So
When I talk to you a lot about foundation models about gender today I but before we get to that lets, you spend a minute on the
on the I m mit collaboration, where, where did that partnership start? How did it originate?
So actually, it turns out that mit and I began have been collaborating for a very long
the area they ai. In fact, the term artificial intelligence was coined. Nineteen. Fifty six workshop.
EL the dartmouth, who has actually organised by an idea that annual rochester, who led the dinner
one of the ibm son one, so we ve really
together in the eyes since the beginning, and as they are
kept accelerating more and more and more
I think it was really in interesting decisions. So let's make this a formal partnership. So I vehemently seventeen and also to be committing close to a quarter billion dollars over ten years to have this joint bob moved with mit and
The way we look at it sells run on the campus, and we ve been developing very, very deep relationships where we can really get to know each other work shoulder to shoulder conceiving what we should work on maximum executing the problem.
And is really very
Few entities like this exist between academia and industry. It's been really fun of the last five years to be a part of it, and what do you think are some of the most important outcomes of this collaboration between I b m and mit yeah. So we're really
the tipp of his spear fur, for I b m's strategy
so we're were really looking way. You know, what's coming ahead,
Anyone areas I professionals as the few changes mit before.
I have been working on your faculty. Students and staff are interested in working on. What's the latest thing? What's the next thing we at ib research are very much interested in the same, so we can kind of put out feelers. You know
doing things that were seeing in our research. Interesting things were hearing the feel we can go and chase those opportunities. So when something big comes like a big change that
it happening lately with foundation. Models were ready to jump on it. That's what the purpose of that is the lab functioning the way it should we're, also really interested in. How do we advance? You know ai that can help with climate change
sure you know build better materials and all these kinds of things that are you know a broader aperture, sometimes than than what we might consider just looking at the product portfolio of I b m and that that gives us again a breadth where we can see connections that we might not have seen. Otherwise, we can, you know, think things that help out society and also help out our customers. So the last, whatever six months say there has been this while
rise in the public public's interest in I write clearly coming out of these generative ay. I models that are really accessible. You know certainly chat jpg language models like that, as well as models that generate images like mid journey, I mean: can you just sort of briefly talk about the the break
who's in a high that have made this moment feel so exciting, so revolutionary for artificial intelligence yeah. You know
then studying I? Basically, my entire life
breaking by the ama as a professor at harvard. I have been doing this a long time and I've gotten used to being surprised. It sounds like a joke, but it's a serious like a. I use, getting used to being surprised at the acceleration of the pace,
again, it tracks are actually a long way back. There's lots of
where there is an idea that just simmered for a really long time, some of the key math behind this,
the stuff that we have today, which is amazing,
as an algorithm called back propagation, which is key to training
networks have been around since the eighties and wide use and really what happened. Was it simmered for a long time and then enough data than enough compute came serbia
enough data, because we all started carrying model
cameras around with us. Are mobile phones, above all, do not always cameras in this way
everything on the internet and this all this out there we call the lucky break that there was something called graphics processing unit which
turns out to be really useful for doing these kinds of algorithms, maybe even more useful than it is fitting graphics. They re graphics too, and things just kept kind of adding to the snowball. So we had deep learning
which is sort of a rebrand of neural networks that I mentioned
from the eighties and that was enabled again by data cause. We did the world and bucharest cause we kept building faster and faster and more powerful,
and then that allowed us to make this this big breakthrough and then
you no more recently using
the same building blocks that he knows
well rise of more and more and more data map, the technology called self supervised learning, where the key difference there in traditional, deep learning for classifying images, you know like is as a cat or is this a dog and a picture. Those technologies requires revision, so you have to take what you have and then you have to label it. So you have to take a picture of cat and then you label it as a cap and it it turns
you know that's very powerful, that it takes a lot of time to label cats and to label dogs, and- and he was always so many labels that he was small. So what really changed more recently is that we have self supervised learning where you don't have to have the labels. We can just take an annotated data and what that does is allow you to use even more data and that's really
drove this. This latest ass with rage and then all of a sudden we start getting these these really powerful models, and then really this has been simmering technologies ride. This has been happening
for a while and end progressively, getting more and more powerful. I'm one of the things that really happens with catchy bt
in technologies like stable
The future and admit journey was that they made it visible to the public utility. Put it out there, the public and touching field
I like wow. Not only is there palpable change and while the stood, I could talk to this thing. While this thing can generate an image, not only that, but everyone can touch and feel and try my kid
can use some of these ai, our generation technologies, and that's really just launched. He notes like propel
when you shouted us into a different regime in terms the public awareness of these technologies
mentioned earlier in the conversation foundation models, and I want to talk a little bit about that. I mean. Can you just tell me you know what our foundation models for I am and why are they a big deal?
yeah, so this term foundation law was coined by a group at stanford, and I think it's actually a really apt term because remember, I said you know one of the big things that unlocked this latest excitement was the fact that we could use large amounts
of an animated data. We can be a train, a model we don't have to go through. The painful effort of labels
each and every example you saw me to have her mother. Do something you want to do is tell me to tell him what do you want to do it? You can't have em all that doesn't have any
purpose, but what he found his models. It provides. A foundation like a literal foundation
We can throw stand on the shoulders of giants. Haven't is massively trained models and then doing
little bit on top you know you could use just a few examples of what you're looking for and you can get what you want from the model. So just a little bit on top now gets you the results that a huge amount of effort used to have to put in you know to get from the ground. Up to that.
novel. I was trying to think of of an analogy for sort of foundation models verses. What
before, and I don't know that I came up with a good one, but the best I could do was this. I want you to tell me if it's plausible,
like, before foundation models, it was
like you had these
If single use, kitchen appliances, you could make a waffle iron. If you wanted waffles or you could make a toaster, if you
to make toast, but a foundation model is like, like an oven with a range on top. So it's like this machine and you can just cook anything with this machine death. That's it that's a great analogy: it they're they're, very versatile
the other piece of the tools that they dramatically lower the effort that it takes time to do something that you wanted
and tat is to say about the old
However, I would say you have. The problem of automation. Is that its to labour intensive has itself
I am making a joke. Indiana famously of automation. Does one thing at substitutes machines
computing power for labour rights. Oh, so what does that mean to say? I is our obligation is to labour intensive it. It sounds like an angel.
Actually says, and what I mean is that the effort it took the older,
to automate something was very, very high. Aha, so if I
it go in here and all this data
all this data and carefully. We have all these examples.
that labeling itself might be incredibly expensive entire to me.
We estimate anywhere between eighty to ninety percent of the effort it takes to feel the name
solution actually is just spent on data, so that that has some consequences, which is the threshold for bothering to you know if you're going to only get a little bit of value back from something or
The is huge effort, turn curate all this data and then comes
the train, the model you did highly skilled people and serve hot.
Find in the labour market
What are you really going to do something? That's just a title and criminal thing now: you're you're going to do that only the highest value, things that weren't right,
I level then, because you have to essentially build the whole machine from scratch, and there aren't many things where it's worth that much work to build the machine. That's only going to do one narrow thing: that's right!
and then they d and tackle the next problem, many best where you have to start over at dinner there.
For some nuances here, like for images, you can pretend I'm
I want another task and change it around too. There are some examples of this leopard non recurring cost there we have,
the old girl too, but by and large, it's just a lot of effort, it's hard. It takes. You know a large level of skill to to
One analogy that I like is you don't think about it? As you know, you have a river of data are running through your company area,
Annotation traditional ai solutions are kind of like building a dam on that river. You know dams are very expensive things to build. They require highly specialized skills-
lots of planning- and you know- you're gonna put it down on a river- that's big enough, though young enough energy,
meant that it was worth the trouble
there are a lot of value that down. If you have a river light back you in a river of data, but it's the actual. The vast majority of the water
You know in your kingdom, actually isn't in that river, it's in puddles and creeks and babylon sim, and you know, there's a lot of value left on the table because it's like like I can there's nothing. You can do about it. It's just that that's too low value, so it takes too much effort. So I'm just not going to do it. The return on investment just isn't there. So you just end up not all,
seeing things says it's too much of a pain now what foundation models do as they say? Well, actually, no, we can train up a base model of foundation, they can work on and we don't we don't care. We have specified what the task is ahead of time. We just need to like learn about the domain of data, so if we want
build something they can understand english language? There's a ton of english language text available around the world. We can now train models on shoes
quantities of it. And then it learned the structure learned how language a good part of howling, which works out
I don't label data and then, when you roll up with your task, you know how to solve this particular problem. You don't have to start from scratch, you're starting from a very, very, very high place, so that just gives you the ability to, as you know, vowels and everything is accessible. All the puddles and creeks and babbling brooks and cal ponzi, although those are all accessible now- and that's that's very exciting,
just changes the equation. What kinds of problems you could use producer to solve an end, so foundation models basically mean that all
waiting. Some new task is much less labour intensive. The sort of marginal effort to do some new automation thing is much lower cause your building on top of the foundation model, rather than starting from scratch, absolutely
that is that is like the exciting, a good news. I do feel like there's a little bit of a countervailing idea- that's worth talking about here, and that is the idea
even though there are these foundation models that are really powerful that are relatively easy to build. On top of it still the case right that there is not some one size fits all foundation model, so you know what does that mean, and why is that
important to think about in this context, yeah. So we believe very strongly that there is just one now to rule them.
There's a number of reasons why that could be true one which, having it,
important and very relevant today, is how much energy these models can consume. So these models, you know, can get very very large. So one thing that that we're starting to see or seem to believe is that you probably shouldn't use one giant sledgehammer model to solve
every single problem you know like. We should pick the right size model to solve the problem and believe me, we shouldn't necessarily seem that we need the biggest based model fur for every little use case, and we are also seeing the small models that are trained
to specialize on particular domains can actually outperform.
Much bigger models so bigger, isn't always even better so they're more efficient and they do the thing you want them to do better as well. That's right! So sir
For instance, you gotta, Sanford trained a model, was giza
two point: seven billion parameter model which isn't terribly big by today's standards. They trained it just on the biomedical literature. This is the kind of thing that universities do and what they showed was that this model was better at answering questions about the biomedical literature, then some models
Two hundred billion prouder soon many times larger. So little
about like asking and expert for help on something versus asking the smartest person you know
gets monstrous. You know maybe very smart, but they're not gonna, be expertise and then
as an added bonus. You know this is now a much smaller model. It's much more efficient to run. We are you know it's cheaper, so there's lots of different advantages there. So I think we're going to see attention in the industry between vendors that say hey. This is the one big model and then others say well, actually you know, there's there's you know lots of different tools. We can use that all have this nice quality that we all find at the beginning, and then we should really pick the one that makes the most sense for the
the task at hand, so there's sustainability, basically efficiency, another kind of set of issues that come up a lot with I r r bias hallucination. You talk a little bit about bias, an hallucination what they are and how you're working too to mitigate those problems.
So there are a lot of issues stars amazing, as these technologies are in and there they are amazing. Let us be very clear: lots of bread,
we're gonna unable with these kinds of knowledge, is biased, isn't a new problem, so you know basically everything
just since the beginning of ai, if you train a model on data that has a bias in it, the model is going to recapitulate that bias when it provides it's answers. So every time you know, if all the texts you have,
says you know, is more likely to refer to female nurses and male scientists. Then you're going to get models that you know. For instance,
There was an example where
She learning based translation system translated from hungarian to english hungarian,
I have gender pronouns english girls and when you asked to translate to translate, they are nurse too. She is a nurse translate. They are a scientist too. He is a scientist and that's not because the the people who wrote the algorithm were building and bias and coding in like. Oh, it's going to be this way. It's because the data was like that you know we have biases in our society and their reflect
and our data are taxed in, are images everywhere and then the models mapping from what they ve, but they ve seen their jane debt or two to the result in your track at them to do it and to give, and then these by
He's come out, so there's a very active programme of research. In a week.
A point about it. I've been research, my tea, but also all over your community and and strain academia trying to figure out. How do we explicitly remove these biases? How do we identify them out of here? How we build
Was that allow people to audit their systems to make sure they are biased? So this is a really him.
thing, and you know it again. This was here since the beginning
you know, of of machine learning any eye, but
today some models and large I'm with models and charity. I d spring it into sharper even sharper focus because there is so much data in its bid to building in baking in all these different policies. We have to that. That's that's
The key problem that these matters have another one, that emissions hallucinations
so even the most impressive. Our models swore often just make stuff up
the technical term that the field was chosen as a solicitation to give you
ample. I asked chauncey bt to create a biography of david talks, idea and he'll get started off.
Well, you know you identified that I was the director of the mit ib I'm watching one hundred and seventy words about that, and then it proceeded to create an authoritative book, complete, be fake, biography of me, whereas Britain,
I was born in the uk k. I went to british unit,
I see it in our universities in the uk, was founded, ready authority right. It's the certainty that it that is, is weird about right. It's it's dead, certain that you're from the uk etc.
Absolutely yeah. That's all kinds of flourishes like I won awards in the uk, so yeah it gets and it's problematic because it kind of pokes a lot of weak spots in our human psychology or if something sounds coherent, we're likely to assume it's true we're not used to interacting with people who eloquently and authoritatively
here met complete, complete nonsense. Idea we forget about that, but we could debate have had, but yes it that the sort of blind confidence
rose you off when you realize it's completely raw grant, that's right in and we do
a little bit of a great and powerful? I was sort of live goings, sometimes very like well. You know,
the eyes ongoing and am therefore whatever its as it must be true, but about these things will make up stuff you're very aggressively.
And you know, if you ever could try asking it for their their bio. You, you you'll, you'll, get something that you always get something that's of the right form that has the right tone but know the facts. Just aren't necessarily there so that that that's obviously a problem. We need to figure out how to close those gaps fix those problems. There's lots of ways we could use them much more easily. I just like to say faced with the awesome potential of what these technologies might do, it's a bit encouraging to hear that even chat, gp tea?
the weakness for inventing flamboyant. If fictional versions of people's lives and while entertaining ourselves with chat, djibouti and mid journey is important, the way lay people.
Use consumer facing chat and generative a is just fundamentally different
from the way an enterprise business uses a I. How can we harness the ability
of artificial intelligence to help us solve the problems we face in,
this in technology. Let's listen on as David Jacob continue their conversation.
we ve, been talking a somewhat abstract way about a high in the ways that can be used. Listen
a little bit more of a specific way. Can you just talk about some examples of business challenges that can be solved with automation, with this kind of automation were taught him.
Yeah. So the really is caused to limit that there's a whole set of different applications to he's mouser a really good at it.
Basically it's a superset of everything we used to use ai for in business. So you know the simple kinds of things are like hey. If I have text- and I know I have- I have like product reviews- and I want to be able to tell these- are positive or
It is, you know like let's look at all the negative reviews, so we can have a human look through them and and see what was up very common business use case. You do with traditional, deep learning a stay. I suggest things like that that are very prosaic sort of we were already doing. Have you been doing it for a long time, then you get situations that are that were harder for the old ai like if, if I'm I want to compress something like I want to, I have like say a half a chat
inspector your customer called in and they had complaint? The call back. Ok now knew. How do you know
Samantha line needs to go, read the old transfer to catch up when it'd be better. If we could just summarize that does condense it all down quick little paragraph, you know customer call they're upset about this rather than having to read the blow by blow
this has lots us settings like ours. Summers ation is really helpful. Hey you have a meeting and I'd like to just automatically
I haven't meeting or that email or whatever I'd like to have a condensed as I can really quickly get to the heart of the matter. Now these
cause. There are really good at doing that. There are also really good question answering so if I want to find out what's how many vacation days do I have, I can now interact in natural language with
system that can go when that has access to our age, our policies and I can actually have a multi.
or conversation where I can you like? I would have you know somebody you know actual
fashionable or customer service representative. So a big part. You know this of what this is doing is it's it's.
putting an interface see where we think of computer interfaces, where you are thinking about you, I
a user interface elements, right, click on menus and those buttons and all the stuff. Increasingly now we can just talk,
you just in words, you could describe what you want. You want to answer, ask a question. You, and I recommend,
system to do something, rather than having to learn how to do that, fucking buttons, it might be inefficient. Now we can just sort of spell it out
interesting, right, the graphical user interface that we all sort of default to that's, not like the state of nature right, that's the thing that was invented and just cause.
to be the standard way that we interact with computers, and so you could imagine, as you're saying, like chat essentially
chatting with the machine could could become a sort of standard user interface
it's like the graphical user interface. Did you know over the past several decades? Absolutely and- and I think, those
The conversational interfaces are gonna, be hugely important for increasing our approach.
Every just a lot easier. If I have had to learn how to use at all- or after I have offered you know, interactions,
I can just tell her what I wanted. I could understand it. Could you know potentially even ask questions back to clarify and have those kinds of conversations that can be extremely powerful. In fact, one area where that's going to be absolutely game changing is in code. When we write code, you know, programming languages are a way for us to sort of match between our very sloppy way of talking,
and if the very exact way the you need commander can get her to do what you wanted to do their cumbersome to learn. They can be a great, very complex systems that are very hard to reason about, and I were already starting to see the latest right now and you want and bay. I will generate the code for you.
I think we're gonna see a huge revolution up like we just converse.
We can have a conversation, that is to say what we want them to computer factually, not only do fixed actions and do things for us, but it can actually even write code to do new things in urban and direction.
for itself, given how much software we have, how much craving we have for software legal will
never have enough software in our world
really to have it. Yeah systems is as helper in that come. I think we're gonna see a lot of of either. So, if you, if you think about the different ways,
I might be applied to business and you ve talked about a number of the sort of classic use cases, waters
of the more out their views cases. What are some you know unique?
ways. You could imagine I being applied to business in others,
really discuss the limit. I mean we have one product that I'm kind of a fan of where we actually were working with me
people engineering, professor at mit, working out a classic problem. How do you build linkage systems
I can imagine bars and joint sin burgers.
The other things that are rebuilding the thing: building a physical machine of some of my real, like metal.
century just all industrial revolution
by the other. The little arm that is holding up my microphone in front of me cranes, uphold your buildings in a part of your engines. Does a classical stuff. It turns out that your humans, if you want to build an advanced system, the eu, decide what by her. If you want to create
then up a human together with computer program, can build a five or six your bar linkage and then thus
are you top out? His excellent gets too complicated to ward off more than that,
he built a jar? Our system, the can twenty bar linkages like arbitrarily complex business,
This means that are beyond the capability of the human to design her himself
For example, we have a system that can generate electronic circuit sooner. We had a product.
very working where we are building better power converters which allow are your computer.
it is on our devices to be more efficient, save energy, less carbon up,
in the world around us is always been shaped by technology. If we look around here, just think about how many steps at home
people and how many designs went into the table
the chair and the way it is really just astonishing and with that's already. You are the fruit of automation.
And peers and those kinds of tools, but where, as he that increasingly be okt, also have a I it's just gone everywhere around us. Everything we touch is going to happen. You know helped in some way to get get to you by
You know that is a pretty profound transformation that you're talking about in business. How do you think about the implications of that both for the sort of you know, business itself and also for for employees yeah? So I think for businesses this is going to
cut costs make new opportunities, delight customers, you know like that, there's just it is sort of all upside right like for the for the workers. I think the story is mostly good too.
Look at how many things do you do in your day that you know
and rather not trade deal? I'm ever use doubting things you don't like automated away. You know we waited maiden
He didn't like walking many miles to work. Then you know like in a car and control.
Where are you? Are you still have a few traction over ninety percent of our europe
I should engage in agriculture and then meet mechanized. Edna are very few people work in agriculture. A small number of people can do the work of a large number of people
and then you know things like email and yet they ve latitude productivity enhancements because I
Don't need to be writing letters and sending alone the mail. I can just instantly communicate with people. You just become more effective, like our jobs have transformed our.
the physical job like agriculture, with whether it's a knowledge were her job, where you're sending ye males in and unity with people and remaining teams we'd just gotten better than any other technologies has made us more productive, and this is just another example now we know there are people who worry that you know will be so good at that that maybe jobs will be displaced and that's that's a legitimate concern, but just like
How would I refer in atomic some may be had ninety percent of the population unemployed, young people, transition to two other jobs, and the other thing that we found to is that our appetite fur for doing more things?
is ass. Humans is a sort of insatiable so you'd, even if we can dramatically increase how much you're one human can do.
That has surrounded me. You're gonna do a fixed amount of stuff. There's an appetite have even
our somewhere get out. You can continue to grow rather pie, so I had to get these. Certainly then,
the term. You see a lot of drudgery go away from work, really see people be able to be more effective at their jobs. Here we will see some transformation in jobs and looked like a be seen that before mountain term, the technology is has the potential to make our
lot easier, so ibm recently launched watson x, which includes what the dot day. I tell me about that:
Tell me about you know what it is and the new possibilities than it opens up. Yes, so so lots in
says obviously a bit of a
new branding the watson brand here today about in those areas
replied the young,
and already I technologies of the watson brand lots of acts
is a recognition that that there's something new there's something that
she has changed the game we ve gone from this old world of
is to labour intensive to this new or other possibilities where its
it's easier to use a lie and what watson ex does that brings together tools to
our businesses to harness that power. So watson study, I said foundation as that our customers can use. It includes
tools and make it easier to run. It applies easy to experiment. Does r watson extort data,
internet which allows you to to organize and access your data, so
we're. Trying to do is give our customers a cohesive set of tools to harness the value
these technologies and, at the same time, be able to manage the risks and other things the the you have to keep an eye on in an enterprise context. So we talk about the
the guests on the show, as as new creators by which remain
people, who are creatively applying technology in business, to drive change
and I'm curious how creativity
plays a role in the research that you do. I honestly, I think, the creative aspects of this job here this is by pigs. This works exciting I've only. I should say the other, though the folks who work on my own
Jason R, r R, during the creating and I like you're doing the managing it so that the architect, the creator and I'm helping them be their best and do I still get to get involved and in the weeds of the research as much as I can, but you know, there's something really exciting about inventing you know,
see one of the nice things about doing, invention and doing research on industry is it's usually
It is a real problem that somebody's haven't. You know the customer
to solve this problem nets, losing money empty others to be a new opportunity. You identify that problem.
And then you you, you build something. That's never been done before to do that and I think that's honestly the adrenalin rush that keeps all of us in this field. How do you do something that nobody else on earth says as done before? I tried before so that that kind of creativity- and
there's also creativity as well- and I doubt if I got those problems- are being able to understand the places where the technologies close enough to solving a problem. The do that matchmaking
between problems that are now solvable.
nay, I wear the is moving so fast. This is constantly growing horizon of things that we might be able to solve said that match making. I think there's also a really interesting creative problem, so I think I that's
that's why it is so much fun and it is a fun environment. We have here too, as it's people drawing on white boards in writing. On pages of now,
and like in a movie like in a movie yeah stray from several casting
drawing on the drawing of the window riding on the window and sharply absolutely so solid
close with a really long view.
How do you imagine a sigh and people working together? Twenty years from now yeah, it's really hard to make predictions of the vision, the ay
I like actually on this came from mit economist named David otter, which was
didn t I almost as a natural,
resource, we have made our natural resources were crowded with vision, or we can dig up out of the earth.
it comes from a cost springs from the earth. We usually think of that in terms of physical stuff, with a sigh
think of as like there's a new kind of abundance, potentially twenty years, not only for me
we can build or either user burn or whatever. Now we have. You know this ability to do things and understand things and do intellectual work, and I think we we can get to a world where automating things
This is seamless spoiled were surrounded by capability to
I meant ourselves to to get things done and you could think of that in terms of like, oh, that is going to displace our jobs, because eventually the ai system is going to do what it can do, but you could also think of it in terms of like wow. That's just so much abundance, then you now have and really how he used that abundance is.
It is up to us in our work and even writing. Software is super easy and fast in anybody can do it just think about all the things you can do now like it.
Think about all the new activities and go to all the ways we could use that to enrich our lives. That's where I'd like to see us at twenty years. You know we can you can do just so much more than we were able to do before abundance. Great. Thank you. So much for your time. Yeah. It's been a pleasure thanks for inviting me.
what a far ranging deep conversation I mesmerized by the vision David just described a world where now
sure conversation between mankind and machine can generate creative solutions to our most complex problems,
a world where we view I not as our replacements principle for resource we tap into an exponentially boost our innovation and productivity
thanks so much to David Cox for joining us on smart talks. We deeply appreciate
him sharing his huge breath. They I knowledge with us and forks,
ending the transformative potential, a foundation models in a way that even I can understand. We eagerly await his next great breakthrough. Smart talks about the aim is produced by Matt Romano gave a jaw me should then cat and raised him preserve with Jacob
we're edited by lydia jean caught. Our engineers are chasing embryo cerebral care and bent holiday theme song, but grammar scope. Especial thanks to kali make liore Andy Kelly cat
Callaghan and the eight bar and ibm teams, as well as the Pushkin marketing team. Smart talks with ibm is a production of pushkin industries and I heard media.
to find more pushkin podcast this money, I heart radio, up apple pie, guess where ever you listen to park ass? I now combat. Well, this is a paid advertisements from idea.
Transcript generated on 2023-12-14.