January 31, 2025 • 8 mins
Chinese AI startup DeepSeek shot to global fame after the launch of its R1 chatbot, seen as a rival to the likes of ChatGPT. But how disruptive will its apparently cheaper business model be to the global tech industry? Bloomberg's Tom Mackenzie joins us to explain.
See omnystudio.com/listener for privacy information.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:02):
Bloomberg Audio Studios, podcasts, radio news.
Speaker 2 (00:09):
I'm Stephen Carol and this is here is Why, where
we take one news story and explain it in just
a few minutes with our experts here at Bloomberg.
Speaker 1 (00:19):
It was done very cheaply. So why the crazy amounts
of spending that are happening from the US big tech companies.
Speaker 2 (00:25):
What's happening here is the collapse in the cost of innovation.
Speaker 1 (00:30):
What this development has shown is the hardware that these
companies have procured, they weren't using it efficiently. A lot
of question marks around deep Seek, in particular, how much
did it really cost for them to develop this new
quote unquote cheap model that they've self reported.
Speaker 2 (00:46):
The implication is, of course that things got cheap, that
things got more simple, but we'll have to see that
manifest itself in the real world. Shocking, a game changer,
a Spotnik moment, just a few of the things that
have been said about the Chinese AI startup deep Seek.
The company shot to global fame over a weekend after
the launch of it's R one chatbart, seen as arrival
(01:07):
to the likes of Chat GPT, but crucially, the company
says it used less expensive chips. So here's why Deep
Seek is a wake up call for the AI Titans.
Joining me Iwo to discuss Boomberg's TV anchor. Tom mackenzie,
a former China correspondent and our resident tech watcher. Thanks
for being with us. Tom, First of all, tell me
(01:27):
about R one, the product that Deep Seek makes. Is
it something that's really comparable to.
Speaker 1 (01:32):
Chat GPT or to Gemini so on a number of
metrics measured by experts, and there are platforms online where
you can see how these measurements are displayed and how
they categorize these models. Then yes, R one competes with
chat gbt's most sophisticated model, the latest model to zero one,
but also the models from Gemini and Anthropic. It is
(01:53):
up there top of the list alongside those kind of players.
It is sophisticated. It is a text based model and
chat bop you can download it onto your phone. It
is not multimodal, so it doesn't produce video, it doesn't
produce pictures, but what it does do is reasoning. Like
one from chat GBT, it goes through how it comes
(02:14):
to its responses, and unlike O one, it actually displays
those for you, so you can see the chat bot
processing your question, going through how it's going to get
to the answer, and that level of transparency has led
to a lot of tech enthusiasts out there who are
getting their hands on this thing responding very favorably, and
that's led to a lot of optimism and some very
(02:35):
positive feedback about what this chatbot based on deep seeks
model can actually do. Now, one thing it can't do
is answer an honest question about Hi Jinping and his
leadership or what happened in Chaneman Square in nineteen eighty nine,
because it is not allowed to full foul of Chinese
government censorship. So there is that important caveat.
Speaker 2 (02:53):
Okay, all very interesting. What do we know about how
it was developed?
Speaker 1 (02:56):
What we know is what is claimed by deep Seek.
So they claim that it was developed at a fraction
of the cost of some of their competitors, around six
million US dollars. They claim the model was trained on
much older chips, not the most cutting edge in video chips,
because those are restricted from the Chinese markets. We know
(03:17):
that they have very innovative and sophisticated engineers. They've been
recruiting talent for the top universities domestically in China systems engineers.
We also know that they put in place an infrastructure
and a method of building models called a mixture of
experts procedure, which basically has lots of mini models much
easier to put together, and if you align them, you
can create these efficiencies. So the engineers, the mixture of
(03:39):
experts method that they've used, they say has led them
to creating these efficiencies, building a model much more cheaply,
using less efistigated chips, and much more quickly. They say
they built this model, design this model within about two months.
Now there are question marks. Microsoft and open ai are
scrutinizing whether or not deep Seak actually lent on open
AI's own model to learn from the outputs from that
(04:00):
model that they then fed into the training of the
deep Seat model that potentially went over and above what
was allowed. That is being scrutinized, And we have to
take them on face value when they talk about the
chips they're using, because we don't know exactly how they
built this. But if we take them on face value,
the more cheaply, more cost effectively, in a shorter time
and with a slightly different method, that created a very
(04:23):
efficient and very capable model.
Speaker 2 (04:25):
Okay, got questions being asked, as you know, more broadly,
when we talk about AI open on now we've mainly
been talking about the models developed by American companies. What
is this revealing something that we didn't already know about
China's AI industry? How does it compare to what we
know out of the US.
Speaker 1 (04:43):
There is an argument that we've had a blind spot
when it comes to the innovation that's coming out of China.
There's been a lot of talk about the slowdown in
the economy rightly, so, there's been a lot of concern
about the real estate market. There's been a lot of
concern about a crackdown on technology in recent years out
of China. That has allowed some to overlook the real
innovation that is happening in that Chinese market. One of
(05:03):
the most competitive places to build and test technology is
in the Chinese markets because you have so many people
who pile in, they test their products, they fight to
the death, and then the survivors come out on top.
And if they can compete in the Chinese market, my goodness,
they can compete globally. And we've seen that, whether it
is with drone makers like DGI, whether it's like TikTok,
social media companies like TikTok, or whether it's indeed the
solar panel makers of China. Across all those different areas,
(05:26):
they compete in the domestic market, they win, and then
they go to compete internationally. I haven't even mentioned the
electric vehicle makers that are posing a challenge to Europe's
EV model now as well. They have the engineers, They
have almost double the number of engineers of the US
in terms of AI engineers. They have the data in
vast quantities what they don't have, and the most sophisticated chips.
If Deep Sea really is an example of how to
(05:47):
circumvent that by using older chips, then China has all
the ingredients it needs to compete on the global level. Interestingly,
we also saw another Chinese company, Ali Baba, coming out
with model this week that is also as capable as
the most sophisticated models coming out of the US. So
every indication suggests that China is a serious player, and
(06:09):
if it hasn't caught up with the US yet, then
it's very very close to doing that.
Speaker 2 (06:13):
It does seem like all of a sudden everyone has
an opinion on deep Seek and its breakthrough. Even the
US President.
Speaker 1 (06:21):
The release of deep Seek AI from a Chinese company
should be a wake up call for our industries that
we need to be laser focused on competing to win.
Speaker 2 (06:31):
So, Tom, what do the AIA companies in the US
need to be woken up to with this deep Seek story?
Speaker 1 (06:38):
Well, interestingly, it's earnings week as well, so we've been
hearing from the CEOs of some of those major players.
We've been hearing from Mark Zuckerberg from Meta, We've been
hearing from sati In adell At, CEO of Microsoft. What
you're not hearing from them is any walk back on
the spend around CAPEX, the spend on the chips, the
spend on the data centers, the spend on the servers,
and the spend on the energy that's needed to power
this AI revolution. Because one of the major questions that
(07:00):
Deep Seek has posed this week is is all of
that spending worth it? Do we need to spend tens
of billions of dollars and all of that AI infrastructure?
If a bootstrap company in China with six million dollars
can produce a model that competes with Chat GBT at
a fraction of the price. But the CEOs are not
walking back from these investment commitments. In fact, Mark Zuckerberg,
(07:22):
who praised deep Seat and said there was real innovation there.
He's committed to sixty five billion dollars of spending this year.
He says they're going to be building their own AI agents,
and he hopes the Meta is going to be getting
those agents to a billion people by the end of
this year, and that they'll be leading in that space.
Microsoft Satiy and Nadella also praising deep Seek, but his
take was this is going to lead to greater adoption
(07:42):
of artificial intelligence. It will drive down prices, and longer
term that will be good news for Microsoft. Microsoft committing
to spend eighty billion dollars this fiscal year on AI infrastructure.
Speaker 2 (07:53):
Okay, a rude awakening for some, perhaps not for everyone.
Tom McKenzie, our Blimberg TV anchor, thank you very much
for joining us for more xp like this from our
team of twenty nine hundred journalists and analysts around the world.
Search for Quick Take on the Bloomberg website or Bloomberg
Business app. I'm Stephen Carroll. This is Here's why I'll
be back next week with more thanks for listening
Bloomberg Audio Studios, podcasts, radio news.
Speaker 2 (00:09):
I'm Stephen Carol and this is here is Why, where
we take one news story and explain it in just
a few minutes with our experts here at Bloomberg.
Speaker 1 (00:19):
It was done very cheaply. So why the crazy amounts
of spending that are happening from the US big tech companies.
Speaker 2 (00:25):
What's happening here is the collapse in the cost of innovation.
Speaker 1 (00:30):
What this development has shown is the hardware that these
companies have procured, they weren't using it efficiently. A lot
of question marks around deep Seek, in particular, how much
did it really cost for them to develop this new
quote unquote cheap model that they've self reported.
Speaker 2 (00:46):
The implication is, of course that things got cheap, that
things got more simple, but we'll have to see that
manifest itself in the real world. Shocking, a game changer,
a Spotnik moment, just a few of the things that
have been said about the Chinese AI startup deep Seek.
The company shot to global fame over a weekend after
the launch of it's R one chatbart, seen as arrival
(01:07):
to the likes of Chat GPT, but crucially, the company
says it used less expensive chips. So here's why Deep
Seek is a wake up call for the AI Titans.
Joining me Iwo to discuss Boomberg's TV anchor. Tom mackenzie,
a former China correspondent and our resident tech watcher. Thanks
for being with us. Tom, First of all, tell me
(01:27):
about R one, the product that Deep Seek makes. Is
it something that's really comparable to.
Speaker 1 (01:32):
Chat GPT or to Gemini so on a number of
metrics measured by experts, and there are platforms online where
you can see how these measurements are displayed and how
they categorize these models. Then yes, R one competes with
chat gbt's most sophisticated model, the latest model to zero one,
but also the models from Gemini and Anthropic. It is
(01:53):
up there top of the list alongside those kind of players.
It is sophisticated. It is a text based model and
chat bop you can download it onto your phone. It
is not multimodal, so it doesn't produce video, it doesn't
produce pictures, but what it does do is reasoning. Like
one from chat GBT, it goes through how it comes
(02:14):
to its responses, and unlike O one, it actually displays
those for you, so you can see the chat bot
processing your question, going through how it's going to get
to the answer, and that level of transparency has led
to a lot of tech enthusiasts out there who are
getting their hands on this thing responding very favorably, and
that's led to a lot of optimism and some very
(02:35):
positive feedback about what this chatbot based on deep seeks
model can actually do. Now, one thing it can't do
is answer an honest question about Hi Jinping and his
leadership or what happened in Chaneman Square in nineteen eighty nine,
because it is not allowed to full foul of Chinese
government censorship. So there is that important caveat.
Speaker 2 (02:53):
Okay, all very interesting. What do we know about how
it was developed?
Speaker 1 (02:56):
What we know is what is claimed by deep Seek.
So they claim that it was developed at a fraction
of the cost of some of their competitors, around six
million US dollars. They claim the model was trained on
much older chips, not the most cutting edge in video chips,
because those are restricted from the Chinese markets. We know
(03:17):
that they have very innovative and sophisticated engineers. They've been
recruiting talent for the top universities domestically in China systems engineers.
We also know that they put in place an infrastructure
and a method of building models called a mixture of
experts procedure, which basically has lots of mini models much
easier to put together, and if you align them, you
can create these efficiencies. So the engineers, the mixture of
(03:39):
experts method that they've used, they say has led them
to creating these efficiencies, building a model much more cheaply,
using less efistigated chips, and much more quickly. They say
they built this model, design this model within about two months.
Now there are question marks. Microsoft and open ai are
scrutinizing whether or not deep Seak actually lent on open
AI's own model to learn from the outputs from that
(04:00):
model that they then fed into the training of the
deep Seat model that potentially went over and above what
was allowed. That is being scrutinized, And we have to
take them on face value when they talk about the
chips they're using, because we don't know exactly how they
built this. But if we take them on face value,
the more cheaply, more cost effectively, in a shorter time
and with a slightly different method, that created a very
(04:23):
efficient and very capable model.
Speaker 2 (04:25):
Okay, got questions being asked, as you know, more broadly,
when we talk about AI open on now we've mainly
been talking about the models developed by American companies. What
is this revealing something that we didn't already know about
China's AI industry? How does it compare to what we
know out of the US.
Speaker 1 (04:43):
There is an argument that we've had a blind spot
when it comes to the innovation that's coming out of China.
There's been a lot of talk about the slowdown in
the economy rightly, so, there's been a lot of concern
about the real estate market. There's been a lot of
concern about a crackdown on technology in recent years out
of China. That has allowed some to overlook the real
innovation that is happening in that Chinese market. One of
(05:03):
the most competitive places to build and test technology is
in the Chinese markets because you have so many people
who pile in, they test their products, they fight to
the death, and then the survivors come out on top.
And if they can compete in the Chinese market, my goodness,
they can compete globally. And we've seen that, whether it
is with drone makers like DGI, whether it's like TikTok,
social media companies like TikTok, or whether it's indeed the
solar panel makers of China. Across all those different areas,
(05:26):
they compete in the domestic market, they win, and then
they go to compete internationally. I haven't even mentioned the
electric vehicle makers that are posing a challenge to Europe's
EV model now as well. They have the engineers, They
have almost double the number of engineers of the US
in terms of AI engineers. They have the data in
vast quantities what they don't have, and the most sophisticated chips.
If Deep Sea really is an example of how to
(05:47):
circumvent that by using older chips, then China has all
the ingredients it needs to compete on the global level. Interestingly,
we also saw another Chinese company, Ali Baba, coming out
with model this week that is also as capable as
the most sophisticated models coming out of the US. So
every indication suggests that China is a serious player, and
(06:09):
if it hasn't caught up with the US yet, then
it's very very close to doing that.
Speaker 2 (06:13):
It does seem like all of a sudden everyone has
an opinion on deep Seek and its breakthrough. Even the
US President.
Speaker 1 (06:21):
The release of deep Seek AI from a Chinese company
should be a wake up call for our industries that
we need to be laser focused on competing to win.
Speaker 2 (06:31):
So, Tom, what do the AIA companies in the US
need to be woken up to with this deep Seek story?
Speaker 1 (06:38):
Well, interestingly, it's earnings week as well, so we've been
hearing from the CEOs of some of those major players.
We've been hearing from Mark Zuckerberg from Meta, We've been
hearing from sati In adell At, CEO of Microsoft. What
you're not hearing from them is any walk back on
the spend around CAPEX, the spend on the chips, the
spend on the data centers, the spend on the servers,
and the spend on the energy that's needed to power
this AI revolution. Because one of the major questions that
(07:00):
Deep Seek has posed this week is is all of
that spending worth it? Do we need to spend tens
of billions of dollars and all of that AI infrastructure?
If a bootstrap company in China with six million dollars
can produce a model that competes with Chat GBT at
a fraction of the price. But the CEOs are not
walking back from these investment commitments. In fact, Mark Zuckerberg,
(07:22):
who praised deep Seat and said there was real innovation there.
He's committed to sixty five billion dollars of spending this year.
He says they're going to be building their own AI agents,
and he hopes the Meta is going to be getting
those agents to a billion people by the end of
this year, and that they'll be leading in that space.
Microsoft Satiy and Nadella also praising deep Seek, but his
take was this is going to lead to greater adoption
(07:42):
of artificial intelligence. It will drive down prices, and longer
term that will be good news for Microsoft. Microsoft committing
to spend eighty billion dollars this fiscal year on AI infrastructure.
Speaker 2 (07:53):
Okay, a rude awakening for some, perhaps not for everyone.
Tom McKenzie, our Blimberg TV anchor, thank you very much
for joining us for more xp like this from our
team of twenty nine hundred journalists and analysts around the world.
Search for Quick Take on the Bloomberg website or Bloomberg
Business app. I'm Stephen Carroll. This is Here's why I'll
be back next week with more thanks for listening
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Boysober
Have you ever wondered what life might be like if you stopped worrying about being wanted, and focused on understanding what you actually want? That was the question Hope Woodard asked herself after a string of situationships inspired her to take a break from sex and dating. She went "boysober," a personal concept that sparked a global movement among women looking to prioritize themselves over men. Now, Hope is looking to expand the ways we explore our relationship to relationships. Taking a bold, unfiltered look into modern love, romance, and self-discovery, Boysober will dive into messy stories about dating, sex, love, friendship, and breaking generational patterns—all with humor, vulnerability, and a fresh perspective.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com