AI Wins Nobel Prizes in 2027
Key Points
- The hosts open the episode with a tongue‑in‑cheek “2027” scenario where an AI‑generated work wins the Nobel Prize for literature and AI also sweeps major entertainment awards, setting up a debate on AI’s cultural impact.
- Recent real‑world Nobel wins are highlighted: the 2024 Chemistry prize went to David Baker, Demis Hassabis and John Jumper for AlphaFold‑related work, and the Physics prize honored Geoffrey Hinton and John Hopfield for advances in neural networks.
- Chris, the CTO of Customer Transformation, argues that the Nobel recognitions signal a shift from theoretical awards to honoring AI’s tangible contributions across science, suggesting AI‑human collaboration will dominate future breakthroughs.
- Edward, VP of Product Management for Watson X, offers a more skeptical counterpoint, questioning whether the Nobel Committee is simply riding the AI hype wave rather than acknowledging lasting scientific merit.
- The show also teases other AI news—OpenAI’s new DGX B200 hardware and fresh funding for the startup Unstructured—indicating broader industry momentum beyond the Nobel discussion.
Full Transcript
# AI Wins Nobel Prizes in 2027 **Source:** [https://www.youtube.com/watch?v=v9go9rttLO8](https://www.youtube.com/watch?v=v9go9rttLO8) **Duration:** 00:37:23 ## Summary - The hosts open the episode with a tongue‑in‑cheek “2027” scenario where an AI‑generated work wins the Nobel Prize for literature and AI also sweeps major entertainment awards, setting up a debate on AI’s cultural impact. - Recent real‑world Nobel wins are highlighted: the 2024 Chemistry prize went to David Baker, Demis Hassabis and John Jumper for AlphaFold‑related work, and the Physics prize honored Geoffrey Hinton and John Hopfield for advances in neural networks. - Chris, the CTO of Customer Transformation, argues that the Nobel recognitions signal a shift from theoretical awards to honoring AI’s tangible contributions across science, suggesting AI‑human collaboration will dominate future breakthroughs. - Edward, VP of Product Management for Watson X, offers a more skeptical counterpoint, questioning whether the Nobel Committee is simply riding the AI hype wave rather than acknowledging lasting scientific merit. - The show also teases other AI news—OpenAI’s new DGX B200 hardware and fresh funding for the startup Unstructured—indicating broader industry momentum beyond the Nobel discussion. ## Sections - [00:00:00](https://www.youtube.com/watch?v=v9go9rttLO8&t=0s) **When AI Claims the Nobel** - A satirical podcast segment imagines an AI‑generated work winning the 2027 Nobel Prize for literature, prompting expert guests to debate the milestone alongside other AI triumphs in science and entertainment. ## Full Transcript
it's 2027 has an AI generated work won
the Nobel Prize for literature chrisy is
a
distinguished let me start that again
give me that pause man just starts
yelling even before I ask the question
and do the in it's 2027 has an AI
generated work won the Nobel Prize for
literature Chris Haye is a distinguished
engineer and the CTO for customer
transformation Chris welcome to the show
what do you think absolutely and while
we're at it AI is going to win a few
Oscars and an em as well okay all right
and uh next up Edward calbar is a vice
president product management for the
Watson X platform Edward welcome to the
show uh what do you think no way uh I
don't think the noble institution will
outow work all right cool well with a
difference opinion like that you know
it's going to be a good show all that
and more on today's mixture of
[Music]
experts I'm Tim hang and it's Friday
which means that it's time again to take
a deep dive with our experts into the
week's news in AI we're going to talk
about open ai's new dgx b200 the new
round of funding for a company called
unstructured but we're going to start
today with the big news story of the
week which is basically that AI has been
taking the Nobel prizes by storm this
year in the prize for chemistry David
Baker Demis hbus and John jumper took
the prize with hbus and jumper winning
in part for deep minds work on Alpha
fold and then in the Nobel Prize in
physics one of the founding fathers of
modern AI Jeff Hinton and John hopfield
another uh Leading Light of the field W
for their work in neural networks so I
think Chris I want to start with you
first is what do we make of this are the
Nobel prizes basically scumming to like
AI hype or is this the start of
something way bigger I love it actually
um I think well I think the Nobel prizes
is if I'm being completely honest has
been a little theoretical and not
hitting with the real world for a while
so actually you know recognizing Ai and
the impact that it's going to have in
multiple Fields such as physics and
chemistry and if we think about the big
innovations that are going forward in
the next few years it is going to be
more AI Le right so it's going to be Ai
and humans in collaboration and how do
you distinguish that well is it fair to
say that the the people who founded AI
in the first place then don't get
rewarded for that work of course not so
I actually I think it's a good thing
because this is the Cornerstone over you
know the next few years where AI is
going to massively help in these areas
so I'm I'm all for it go for it and
while we're at it uh I think my next AI
means I'm going to become the MVP of the
NFL soon as well you know so Aaron
Rogers watch out here I come I guess
Edward to bring you into the
conversation I think Chris has already
taken a very strong stance that
basically you know it's G to be a few
years from now and you'll just be AI
winning every single Nobel Prize award
um I guess Edward there's two questions
I think what you said in the opening
question you were like well I don't know
if the dobell institution will really
allow it does that mean that like you
don't think it's deserved or basically
that you think like the institution
won't really you know be into being into
awarding and giving AI it's do yeah I
mean I think I think um mainly the the
former uh I mean I do think that that AI
is going to be an incredible Tool uh in
in almost every aspect of our lives and
it's going to do amazing good for
society and and uh well-being and
quality of life and and and basically
all the things that they institution
stands for uh but but I I do think uh
you know the human contribution to uh to
to to those outputs that deliverable is
really essential right so whether it's
10% AI or 90% AI I think if it's 100% AI
maybe it goes a little bit too far uhuh
right um I think Chris one thing I
wanted to do is you know I think I I
take a lot of pride in the fact that
mixture of experts is like a good way
for people who might not be like reading
every single archive paper working their
way through every single machine
learning textbook to kind of learn about
sort of like what's happening more
deeply in the AI space and you know I
think what's kind of interesting is you
can often get lost with all the stuff
that's happening like at the Enterprise
layer around AI like I think there's a
lot of people listening to this who may
actually not even know Jeff Hinton right
um and um I guess I'm curious if you
feel comfortable you know if you want to
give our listeners just kind of a quick
explanation for why someone like Hinton
is so kind of important to the field and
and what exactly he sort of like
contributed here no absolutely so I
think the first thing I would say is
Hinton really is kind of the he's
considered as the the OG goat of uh
Godfather of AI which is which is I love
that term in that sense and and he's
been doing AI for a very long time U
machine learning specifically he has uh
been there before it was cool so he was
doing that right back in the sort of
1980s right so if we think really uncool
basically right yeah very uncool exactly
so so but if we think of the modern
foundations of what we've got today so
we think of things like deep learning um
that all comes down into these really
deep massive neural networks with with
billions and even trillions of
parameters these days right and I'm not
going to go into the the massive details
of that but if we look at the work that
Hinton has done there even as far back
as sort of 2011 right when uh Alex snack
came out um you know and IL were were
was part of that as well then that was a
time where really the Deep learning
Revolution kicked off which was the sort
of first kind of CNN on a on a GPU for
training uh against images and if we go
further back uh in time there as well so
if we look at the the work that he did
things like back back propagation which
is a key Cornerstone of what we even do
today with deep learning so all of this
goes back to the 80s and Jeff Hinton
Wasing doing this when it was uncool
right so if he hadn't done that work we
wouldn't be where we are today so I I
think you have to sort of recognize that
fact and as I said earlier the impact
that AI is having and going to have in
the future is going to be incredible so
I think actually the impact that it's
having in these different fields he
should be recognized for the work that
he's done right even even in physics
right because I know there's some
physicists I saw on my Twitter feed
grumpy about like well what's this
computer science person doing in here I
guess kind of ultimately you're like
actually this is significant enough that
like it it actually should be recognized
in this context absolutely and and I
think it does open this up right so and
I think it's a good thing for physics as
well right you don't want to be seen as
this kind of boring thing here's a bunch
of formulas you know oh look here's
another telescope in the sky do you know
what I mean it's like this is stuff
exactly AI is impacting every field and
and and therefore um I think I think
it's a really good move by by the Nobel
uh instit Edward I'd love to bring you
into this because I think one of the
things I I love about Jeff Hinton in
particular is just kind of how sort of
down to earth and like kind of open he
is about his sort of views um there's a
great quote that the Nobel committee had
posted on Twitter about how he was like
oh yeah like I'm I was just in this like
low rent hotel room when I got the news
and like I I had to like reschedule my
medical appointments to go deal with
this Nobel Prize win um I think one of
the things that Hinton has been sort of
very kind of um sort of strong on I
would say in the last few years is kind
of warning about the sort of risks of AI
and I think people have taken him very
seriously because he's been at this of
course for a very long time as Chris
explained he's sort of like the goat
hipster of neural Nets um and I guess
I'm kind of curious about how you sort
of think about those as kind of like a
Leading Light in the fields you know do
you take do you take his kind of sort of
dark warnings about where AI is going
seriously do you think he's on the right
track you know I'm curious about how you
kind of think about those those kind of
risks and he I think made it actually a
center of like some of his comments
during his some of his interviews around
this prize and so I did want to make
sure that we talk about it before we
move on to our next topic yeah I mean I
think he's he's raising the warning uh
just to make sure that that that that
that voice is you know always considered
right that that that that risk is always
kind of part of the part of the math
that uh that we're you know that
enterprises individuals
institutions uh you know do when when
they're applying AI to to the particular
problem or use case that that they're
applying it to and I really think that's
what it comes down to right so when we
think about uh risk assessments or AI
governance it's really in the
intersection of the technology and the
use case right it's not the same thing
to apply AI to do creative writing right
as we mentioned this morning uh than it
is to do credit underwriting uh for for
a bank right or to do a hiring decision
for for an organization uh so these are
very different use cases very different
impact on on individuals right uh on on
Society uh and and they they pose
totally different uh different risks so
it's really not just about that
technology it's really what the
technolog is being applied to uh that I
think uh that I think is is needs to be
assessed you know the more this
technology makes its way into into
National Security into defense right
obviously it's a much different
consideration uh than uh than than
poetry yeah for sure and is this are you
hearing this like um you know because
you work very close to the metal in some
ways right like Watson X platform is
something that like customers are using
and relying on I mean sometimes I think
I feel like a lot of the discussion
about like oh AI is really dangerous
kind of takes place in like a totally
different domain but I I guess I'm kind
of curious I mean it sounds like you're
sort of suggesting that like even
day-to-day you're sort of hearing from
you know customers and and the market
that like these kinds of risks and these
kinds of concerns are are things that
people are thinking about yeah I mean
they're not existential risks right but
but there definitely risks uh to uh to
Brand to Brands right uh their their
their business risks uh their Regulatory
Compliance risks uh and and and managing
these risks uh is definitely one of the
the top considerations that enterprises
are uh uh that that that's that's really
acting as an inhibitor right to to more
WID scaled adoption of the technology uh
and and it's something you can't really
do after the fact right because so much
of of uh of managing this risk is is the
endtoend life cycle right it starts with
the data that goes into the model and
the
and you know what model you selected and
how you customized and tuned it all the
way to monitoring and guard rails um
separation of Duties right between
development deployment so so all these
things that uh that you kind of have to
start thinking about from the beginning
because if you don't then at the end
they become a wall or real obstacle to
try to reconstitute Posta uh so so
that's we've been working with clients
uh uh you know in that perspective uh
with that approach uh and that's what's
leading to you know some of these use
cases making their way into production
and I wasn't being hypothetical when I
was talking about credit risk
underwriting and and and hiring
decisions right these These are these
are real real world use cases where
where the risk is being assessed
mitigated uh in order to implement uh
these uh these work fors yeah for sure
um Chris do you want to get a final shot
here curious about what you think about
sort of you know I guess hinton's kind
of late career turn as being sort of
like a voice of warning around some of
these Technologies I I like to think of
this as like the difference between
waterfall and agile you know which is
probably a weird way of putting it which
is go into that more if if we think of
waterfall projects nobody does waterfall
projects anymore because we realize that
we are too dumb and I mean this in the
nicest possible way to figure out every
requirement in advance and be able to
plan everything because the world is too
complicated I sort of feel that way
about AI risks I think we are too dumb
to figure out every single risk and
every exploitation and be able to get
ahead of everything in advance and
pre-planned so therefore I kind of think
like a software project I think we need
to be agile which is you need to
experiment and then you need to discover
in a safe and controlled fashion what
those risks are and let them evolve and
that means we're going to do dumb things
we really are right but then in the
process of doing dumb things like you
know sticking your fingers and you know
a wall socket or whatever you realize oh
I better not do that right and then you
put safety things in there now I'm not
saying that we should go that far with
AI but I I hope that history tells us
that in human existence um we've done
enough dumb things that we shall do a
enough dumb things before they become
catastrophic dumb things so I think a
little bit of agility will help us
discover that stuff we need to have
control but I don't think we are going
to all blow ourselves up because I think
we're going to do much dummer things
much earlier that is my
[Music]
opinion I'm going to move us on to our
next topic um there was an incredible
photo that uh open AI uh put out onto
social media it's of the team
celebrating their receipt of the new
Nvidia dgx b200 um and it's a great
photo because it's like you can see
clearly that everybody is so jazzed to
be standing next to this fresh new piece
of compute that it's like Christmas
morning you know it's like people are so
thrilled to get this computer in their
hands um and I think it's a nice cook to
talk a little bit about this kind of
next generation of uh platform that
Nvidia is rolling out um and is actually
having a really material effect on the
market for compute right so um there's a
great chart I saw earlier in the week
about how sort of the prices for NVIDIA
h100s right which were sort of like last
season's got to have it Hardware those
compute costs are just dropping all of a
sudden right as these new boards are
kind of coming available and online and
so I think it's a nice hook to talk a
little bit about what's happening in the
hardware markets and I think you know
maybe Edward I'll turn to you first you
know is what we're seeing here just more
speed right like I guess there's one
kind of point of view which is it's
Christmas morning because it's really
cool to be standing next to what's
basically like an F1 racing car for
compute but like is what we're getting
here largely just faster and if not you
know what's different about it yeah I
mean I think I think it connects back to
the first topic we talked about right
the the the the the evolution of this
technology and really trying to to build
it in a way that uh somewhat models the
way our brain works right and and and
this kind of almost Infinity uh I know
that's a big word but uh uh of of of
nodes and connections and and relative
strengths between them right so so I
it's not just speed right it's it's it's
it's scale right and the capacity to to
consume more data and to have more
nuanced uh relationships between between
that data um so I'm I'm not a hardware
expert uh but uh but I definitely uh I
think it's I think it's a great time in
technology when Hardware matters again
right I think we go through Cycles where
like Hardware becomes totally
commoditized and then and then it
matters again uh and then eventually
becomes commoditized again right uh so
we're definitely in a in a stage where
it matters um I think that I think
that's a signal right that that the
Innovation right the The Innovation
Frontier is is active and and moving
rapidly and I think that's all very
positive yeah I mean I think Chris it's
it's stunning I was talking to a friend
of mine who is working on some of these
clusters and he's basically like the
hardware is literally moving so quickly
here that they can only really afford to
do like one big training run on a
cluster they've built and then almost
immediately they start moving to
building the next cluster that they're
going to do training right on um I guess
I'm kind of curious here as someone who
kind of like thinks about this and
researches in this space you know
where's this all going right like are we
just you know is is the cluster just
going to get bigger and bigger and
faster and faster you know like is there
a top this this top out at some point or
you know what what is the trend here in
the next like 12 to 24 months I I think
there's a couple of Trends going on and
and I I and I think I might have said
this on another episode but I'm going to
say it again it's like it's almost like
following the Bitcoin Trend right which
is if you follow a Bitcoin Trend
everything started on CPU then it moved
from gpus and then it moved to fpgas and
it moved to as6 right so so basically
you went from kind of CP you went from
you know compute being CPU to being GPU
bound and then you were going to custom
hardware and we we're kind of seeing the
same thing again because people you need
to bring the cost of compute down you
need to bring the cost of training down
I.E you've got bigger and better models
to train but actually I think the bigger
thing is on inference right so you got
to run these models at a low cost and
speed now if if there is one criticism I
would say of Nvidia uh over anything is
that the speed of tokens per second and
the cost on these gpus are quite
expensive and you've seen this in the
marketplace already this is where folks
like grock have been coming in right and
they've been sort of releasing these
chips that go really really fast and
then IBM's got their North Pole chip as
well right um and then Google's got
their TPU chip so everybody's trying to
bring down the cost of inference because
if you're running these massive models
on the cloud everybody's consuming
compute you that to be as cheap and as
fast as possible the big thing if you
look at these new Nvidia boxes right is
yes the the the training speed was much
faster but actually if you look at that
chart the cost of inference the speed of
inference came down massively right so
they've obviously put a focus on that as
well because they know that if they
don't improve the inference speed if you
don't improve that influence cost then
all of these other providers are going
to start eating their lunch as well
right because everybody's is going to go
cheaper and I and I and I but I think
this push and pull between kind of
general purpose GPU and sort of custom
chips is really important but again in
the training point of view different
from inference everybody's just focused
on I need to get the biggest and fastest
mod I need to get my model out really
quickly and therefore you know throw
away your last card put in the the
latest card because I just need to get
my model out all the time so there's a
different Dynamic that's going on
over time you know you're going to get
faster architectures you're going to get
different architect it's going to get
cheaper and and these cost uh speed
performance ratios are going to change
over time yeah the architecture I think
bit of this component I think is a
really interesting part of the market
right I think like one one theme that
we've had pop up on a lot of mixture of
experts episodes is customers want
smaller models they want faster models
uh they don't want the gigantic model
that's really expensive right um and so
there is that pressure there but it
feels like there's kind of two ways of
getting there right one of them is well
we start marketing just smaller models
right where we like lower our demand the
other one which you're arguing is well
the chips get good enough that the cost
of inference finally Falls for running
larger and larger models and the two are
kind of like in a little bit of a race
it sort of seems like um and I don't
know predictions on kind of who wins
that race in the end because you can
imagine like the the market might
eventually settle and say hey look these
models do 99% of what we need them to do
we don't need cre near AGI models to do
this work so at a certain point you just
don't need the chunkier model right I
guess there's another point of view
which as well but if the cost was cheap
enough you would still go bigger um and
I guess I'm kind of curious about like
how you think about that relationship
it's a little bit complex and it's
unclear where it lands in the market
today I think it's just going to keep
pushing and pulling right because we are
going to want to run our models on
device right if you think of things like
apple intelligence Etc so I think
smaller models and faster compute are
just they're you're going to need both
for a while will one win it yeah and
will one it win out I don't think so
because the smaller that you can make
the models and the faster you can make
and smaller you can make the chips then
the more you can put them on embedded
devices which open up a whole set of
other scenarios which are kind of low
latency and and again you even see that
like this week so what uh llama 32 was
out last week and they released their
their 1 billion model and their three
billion model I think it was right and
and again just smaller models and and I
think the big thing there is folks are
getting really good at taking these
larger models and Distilling them down
into into much smaller models and that's
going to continue and I and I think
we're looking at 1 billion parameter
models but let's project forward maybe
uh six months a year you're going to
then start to be back into the million
parameter models and then the chips are
going to get faster and we're just going
to go back and forward back and forward
and it's forever yeah um yeah Edward are
you seeing that in the market I mean it
feels like one kind of interesting
outcome of what Chris is talking about
is that there's a lot of Market pressure
to like have a lot of the models just
more on edge devices everywhere um and
it strikes me that like you know part of
the idea of a platform is you're running
it in the cloud and all the advantages
of cloud but it does seem like there's
actually really powerful kind of
economic incentives eventually kind of
pushing us to sort of all on device here
you know not really like in the model
that we're familiar with do you think
that's like a real possibility going
forward I mean I think it's going to be
all the above uh and you know and we're
we're the hybrid Cloud company right so
so Edge to us you know is definitely a
Continuum right uh uh the data center
right compared to the hypers skater
cloud is is effectively a type of edge
um and then you go down to facilities
and and eventually you know devices um
so so yes it's going to be it's going to
be all of the above and and and finding
the right balance is always uh is always
very specific to the requirements of the
of the use case I mean I think what what
what we see a lot right is
clients to get started use a big model
because that's a that's a way of you
know accommodating a very broad range of
of requirements use cases
languages uh all sorts of things right
so so so so you kind of prove out the
business case with a with a big big
model uh that's that's going to help you
accelerate right but then when you're
there you're like okay how can I do this
as cheap cheaply and with the least
latency as possible right uh and and now
you start to really kind of optimize um
and customize right uh once you once
you've validated that uh that that
business case and really want to want to
scale it uh with uh with with real
economics behind it so you know it's
it's it's like you use the the Swiss
army knife right uh it's going to give
you a lot a lot of flexibility but
eventually you know you're going to want
to use that fit forp purpose tool to get
the job done yeah that's super
interesting I never really thought about
it as kind of like this life cycle but
it's sort of very interesting that like
well just for the pilot we use the
biggest baddest model because it gives
us the most optionality and then as an
organization kind of Tunes in the use
case it gets kind of like much more
discret and smaller and you're
optimizing for cost and all these other
sorts of things um it's very interesting
is this the time to mention agents I
realize we haven't mentioned the word
agents in this episode yet so I mean
we're not contractually obligated to
talk about agents but if you want to
mention agents Chris go for it you can
do the final uh uh hot take before we
move on to last topic this is needed for
agents because you need your agents are
going to be highly specialized they're
going to work together and they need to
have low latency Etc so actually the
smaller model and being being able to
run on device and being able to run in
you know whether it's on data center on
device and running different locations
that is 100% necessary for this agentic
world that we're in so uh you know so
it's a good thing agents for sure
agents uh a lot more to get into there
for sure
[Music]
sure so for our final story of today I
really wanted to make sure that we had a
chance to talk about a company called
unstructured which recently closed a $40
million round um and uh this round was
led by IBM and Nvidia and a long list of
kind of prominent companies and
investors in the space what's most
interesting about inst structured is
it's a company that focuses purely on
transforming unstructured data into
structured data which is not something
that you normally think of as being
something that you'd invest $40 million
in so I want to make sure that we talked
about it first Edward if I want to bring
you in just like if you want to resolve
that mystery for some of our listeners
like why is unstructured data important
and why is structuring it incredibly
incredibly important for AI yeah well
unstructured data is most data uh
nowadays right and and I think the most
relatable type of unstructured data the
most usable type of unstructured data
today for for llms is uh is document
data right so so
um could be could be the the content on
the Internet or or or word docs or
PowerPoint presentations right but
effectively document data and that is
that is Enterprise knowledge right that
is that is what runs the world right
it's it's these documents uh in this
language information um and and and
that's what large language models are
built on right that's what they're
trained on and that's what they're
excellent at processing
summarizing uh and and and making making
usable right um so so bringing bringing
that data bringing that Enterprise and
institutional
knowledge to the models is really the
way in which uh in which an organization
can make it their own right customize it
to the knowledge of their business the
language of their business the tonee the
entities the relationships right the the
values all all all the things that you
need to to put a a model uh in service
of of a business right or a goal you
need to do that by by effectively
teaching it right uh with with uh with
your data and and that's what this
company uh focuses on i' I've met them
uh they're very focused I think that's
really been part of their strength and
success they're very focused on taking
that unstructured data that relies you
know in different locations and and and
and different formats and then make it
available for for the models
particularly uh in Vector stores right
for uh for retrieval augmented
generation as an initial uh use case
which is effectively Universal at this
point but then beyond that you know
identifying relationships in the data
for graph rag taking the data and
putting into structured format to really
increase the Precision uh and accuracy
of some of those queries so I think I
think rag is uh you know very popular
really valuable uh but already kind of
running out of uh out of gas a little
bit for the next evolution of uh of use
cases and and that's really all about
kind of continuing to unlock the value
of the data in those documents huh yeah
that's really interesting can you go
into that a little bit more is like why
why is rag running out of steam it's
kind of like again it feels like 12
months ago is like the new hotness right
or like people were still definitely
like leaning into it as the kind of key
strategy for doing retrieval what's
what's missing I guess what's what were
the cracks appearing yeah I mean I think
I think it's a great starting point and
I and I think think it's I think it's
essential in in most cases right but but
for example graph rag um is going to
give you the ability to have richer
contextualization right by identifying
non- obvious relationships if I prompt
the model with you know a certain set of
words uh it's it's really only going to
limit uh its ability to to to reason
right including uh retrieving the the
the uh the knowledge base to to that
domain right and there may be hidden
relationships right there may be for
example if going to search something
about Facebook but I don't get a
response about Instagram then I'm not
really getting the whole picture right
uh but the model is not necessarily
going to know that Facebook and
Instagram are are related right because
those relationships could potentially be
non-obvious right uh so so the graph rag
um pattern right it's going to give you
uh strength in relationships uh that are
nonobvious and in doing so provides you
richer contextualization that will be
more relevant right to the question
being asked even if it's not asked with
those specific words right so it's it's
again MIM mimicking a little bit of how
how how our brain works uh in in in
identifying those relationships that's
that's one example and but even that is
not necessarily going to be perfectly
accurate right because there's data
about transactions that may have like a
skew a skew number or a particular you
know ID has no semantic value it's just
a bunch of characters it's like it's
like your license place it doesn't
really doesn't really mean anything
right so so so you need to have that
type of data in structured formats and
really combine rag or semantic search
with with SQL right was with structured
query queries and that's going to give
you more accurate responses to questions
that have you know to do with uh
transactions or or other types of data
that are that are very you know very
important to a particular business very
important to particular domain but don't
have semantic value in a conversational
or language sense right so now you have
to complement rag with a different
dimension
uh which is structured data so those are
just two examples right of of how how
you really need to complement kind of
classic rag uh to make it more more more
accurate that's really helpful Chris you
know again I think there's another I
think this story made me think a little
bit about like the market for data
structuring which I think is really
interesting which like we normally think
about like okay the people who generate
data the people who do the training the
people who offer the models to the
consumer as kind of the supply chain and
one Link in that chain I have really
thought about is just like this layer
that exists between like the data that's
out there and like the data that's
usable um and I guess one question I
want to ask of view is that it feels
like there's lots of different potential
ways you could go about doing that right
there's companies like unstructured
where like we have a Specialized Service
that does the structuring of data for
you you might imagine that um you know
that the the foundation models
themselves become good enough that they
can do the structuring kind of out of
the box you don't actually have to do
much additional post-processing to make
it happen you could imagine that uh
synthetic data gets good enough we don't
even need the UN structure data because
we can just generate a purely you know
out of nowhere um and it feels like
there's a lot of contenders to the
throne of getting data that's usable um
I guess how do you size that up like do
you think that at some point say like
synthetic data just gets good enough
that you know you don't need to do this
data structuring anymore or is there
like always going to be a niche for this
kind of structuring kind of business
just kind of curious about how where you
think this Market is going oh goody I
get to say the word agents again my fa
yes please do yeah we got to get a few
more in before the episode's up so so
actually I think everything is GNA move
into a Marketplace in the future so I do
think we're going to have a Marketplace
of data we're going to have marketplaces
of agents and we're going to have
marketplaces and models and I think we
are going to get more outcome focused so
specifically on the data I think we're
doing a lot of human work at the moment
to curate that data and even if you look
at things like stretcher Etc they do
great work because they're actually
taking away a lot of the complexity to
get your data into your kind of vector
databases to follow rack right because
it's it is really hard you have to do
things like chunking you're constrained
by the uh the context of the model I.E
the short-term memory that it can work
within you have to work out which data
is going to be associated with wall as
Edward's saying you need to start
building up things like relations and
then you've got to understand okay I've
got to get this data from this form I'm
getting this from an S3 bucket I'm
getting this from here it's really
complicated but actually we are even
though that's a faster process we are
still humans who are figuring this out
and doing Transformations Etc and doing
these sort of ETL pipelines if I project
a little bit forward in the future back
to our earlier discussion where the
models are going to be smaller they're
going to be uh have latency they're
going to have faster tokens per second
you're then going to be able to train
these smaller models to start to do that
restructuring work for you and therefore
uh you I think you're going to be in
this world where agents are going to
help you get your data into a structured
format and once your data is into a
structured format you're going to be
able to train your model and then you're
going to loop around and you're going to
be in this nice virtuous circle
so will there be a Marketplace for this
absolutely because at the end of the day
people own data right so
the uh the publishing companies the
media companies the they're all sitting
on gold mines right um at the moment
because that's data that is highly
valuable highly creative there are
things that are probably can be
synthetically generated so things like
all the math data Etc you could probably
argue you know that will just be
commoditized over time because that will
just get generated and synthetically
created and that would be the same for
anything that our puzzle Games Etc so
there will be this push and pull of who
owns that data and and I think that
human data in especially the creative
spaces will still be highly valued um so
um I don't see the record companies
giving up their ownerships of or
songwriters of yeah exactly so I think
that's going to be the push and pull
that we have over over time but we are
going to be moving into this uh
marketplace where sort of that soft IP
is just going to be uh the big thing
that distinguishes companies
because one of the examples I like to
give is if you have got a model trained
and you have the data of all the kind of
Spanish legal texts and you've got that
structured Etc and your model can answer
uh Spanish legal queries better than any
general purpose model if I'm going into
court you know what I want the the model
that's really good at Spanish law as
opposed to the model that's got a vague
understanding of Spanish law because
that's the difference of me getting a
large fine or going to jail right so you
know so there's a huge value on that
locality and I and I think that will be
one of the biggest Trends as models are
going to get more and more specialized
and we're just going to be like like
we've been having with the general
purpose benchmarks MML use and all that
we're going to have a benchmark for
everything you can imagine Tim it's
going to be here's the Spanish legal
Benchmark here's the car parking
benchmark that you name it is going to
be benchmarks everywhere and we're just
going to be in this big massive
Marketplace of specialization I I love
the image of of of hiring an agent AI
agent attorney right to defend you in a
in in a case I mean I think I think that
that is that is a that is a feature I
can get behind I used it myself I did a
an insurance claim I looked at the
insurance document and I was like I have
no clue what any of this means so right
and it was a kind of medical condition
thing and I was like I run through the
llm it's like huh tell me gave me the
the key points brought to the insurance
company pay out and you're like you know
uh that is this could go somewhere
exactly that's what you want from from
these things so I'm I'm yeah I but we
are we are going to be in a wild ride we
are we are going to be having like the
kind of the Uber style marketplaces
where you're matching up AIS to people
AI to AIS it's going to be wild over the
next few years anded do you want to
final to close us out for the day uh
agents agents
uh absolutely I mean I I uh some of the
work we're doing uh at IBM with agents
is super exciting uh and and it really
is going to kind of I think it's going
to be a step function right in terms of
the the the complexity of the of the
workloads and the use cases uh the the
creativity right uh to to solving
problems uh um potentially Beyond even
even even our approaches uh the the
automation right the fact that you know
you're going to have so much work
happening 24/7 365 a lot of stuff
already works that way right but but
this is going to take it to the next
level um and uh and and I think it's
it's it's exciting it's productive um I
think it's going to level the playing
field for for uh for for for consumers
in some cases uh for smaller
institutions right so so uh we're we're
we're excited um uh to be part of this
future and and to really be co-creating
it with h with our clients and uh and
our community well gentlemen this is um
wonderful uh Chris you're always welcome
back on mixture of experts um and uh
Edward I hope to have you on uh some
point in the future listeners out there
if you enjoyed what you heard you can
get mixture of experts on Apple podcast
Spotify and podcast platforms everywhere
and we will see you next week for
another Roundup