Skip to main content

tv   Charlie Rose  PBS  April 4, 2017 4:00pm-5:01pm PDT

4:00 pm
. >> rose: welcome to the plam. we begin this evening with gillian tett from "the financial times," one of the three people who interviewed president trump on friday. >> a lot of people would say what could the u.s. go it alone. could the u.s. take military action to contain a nuclear threat from north korea. the answer is at the moment not easily or no without a huge loss of life on the korean peninsula. could the u.s. be sanctions to -- in north korea. it has been trying to do that for some time, it has not worked. so there aren't any easy options for the u.s. to go it alone. >> rose: we continue this evening with dr. siddhartha meuk hedgy who rose a fascinating article in the musical-- new york magazine for ai and m dks. >> for me it raises the whole question of diagnosis. you know, where are we going
4:01 pm
with diagnosis, are we going to overdiagnose, are we going to start invading in the body in a way we hadn't expected to. >> rose: we conclude with barbara finamore talking about the united states, chienar and climate control. >> today in the united states the solar industry alone employs roughly twice as many people as oil, kohlhepp and natural gas industries combined. so it's not really going to grow jobs. in fact, i think the main impact of what trump has announced is that he is going to put the u.s. in daker of-- danger of losing out on the biggest new global energy-- the big new global industry of any kind which is the transition to clean energy. >> rose: ft interview, medicine and-- all of that when we continue. >> fuming for charlie rose is provided by the following.
4:02 pm
>> and by bloomberg, a provider of multimedia news and information services worldwide. captioning sponsored by rose communications from our studios in new york city, this is charlie rose. >> rose: we begin this evening with the bomenting today in st. petersburg russia, at least 11 people were killed after a bomb exploded on a subway. a separate shrapnel filled explosive device was later found and disarmed at another subway station. there was no immediate claim of responsibility. russian president vladimir putin was in st. petersburg for the day and said that terrorism is
4:03 pm
being considered as a possible cause. for more on ta developing story we turn to coverage on the cbs evening news. >> the blast was strong enough to build out the subway doors. some passengers pulled others off the train. but some died on the spot. amateur video shows the chaos and shock underground as dazed commuters tried to figure out what had happened. the explosion took place between stations but the train carried on through the tunnel to the next stop where the wounded poured up and out on to the streets. investigators say they found a second bomb at another station which they diffused. it was hidden under a fire extinguisher and they say had it gone off the blast would have been even more powerful than the first. president vladimir putin speaking in his hometown of st. petersburg today said investigators would pursue all leads. islamic extremists from the north caucuses have attacked russian public transport before.
4:04 pm
their most recent subway attack was in 2010 when two suicide bomb oars killed at least 40 people in the moscow metro. but sis then russia has joined in syria's messy civil war and that may have put it squarely in the crosshairs of isis. >> rose: we begin this evening with politics, president trump will meet with chinese president xi jinping at mar-a-lago resort this thursday and friday. st t is their first meeting and a crucial test for the leaders of the world's two largest est economies. on sunday "the financial times" published an extensive interview with president trump. among the revelations was the willingness with the white house to act unilaterally with north caria in the absence of cooperation from beijing. joining me is gillian tett, the u.s. managing editor of "the financial times." she was one of three editors from the paper who sat down with president trump in ot val office on in the oval office on friday afternoon. this is the coverage they're giving it.
4:05 pm
look at this. >> well, it is the president of the the united states. you don't often get a chance to speak to him and we're grateful that we did. >> it made headlines around the world because of what he said first about north korea. a question you asked? >> i did ask the question about north korea because i think it's critically important. we spent much of the last two years worrying about syria, worrying about iran, worrying about russia. so understandable reasons. but what is happening right now in north korea is critical. and frankly, it needs to be talked about. >> rose: well, i mean, are they talking about it? do they seem to be not only talking about it but worrying about it. everyone you talk to, as you know, says the key to north korea is you have to go through china. >> the key to dealing with north korea is indeed china. because about 70% of the energy is thought to come from china. and if they turn off the taps and north korea will have a problem. unfortunately t does look as if some of china's leverage of north korea recently has diminished a little bit. but a key part-- . >> rose: and the reverse is true. they seem to be less willing to
4:06 pm
do everything north korea says or wants. >> certainly at the moment china is left willing to do everything north korea wants it does look as if relations are deteriorating a bit. the recent assassination that took place in north korea was quite symbolic. because it means most of the chinese-educatedded officials inside north korea are no longer within the innercircle. but nobody can have any doubt that the u.s. administration believes that china is critical to trying to reign in north korea and so it's going to be absolutely center stage of the meeting in mar-a-lago on thursday and friday. >> along with trade and south china. >> along with trade and south china. because the critical thing to understand is that increasingly people are saying that these issues shouldn't be looked at in isolation. on friday the white house released news about its big trade review. and certainly they're having tough conversations with china about the trade issue. and they've been beating the drum on that quite a bit.
4:07 pm
but the question people are asking is could the white house use the trade talks as part of a bargaining chip to try and get some grand deal over the korean peninsula. hard to see at the moment but certainly that is in discussion right now. >> back to north korea. so what did you ask the president? >> i asked the president whether he was keshed about north korea, whether he thought they could be some kind of discussion, meaningful discussion. and if the chinese were not willing to play ball, what exactly were they going to do? >> and he said? >> he indicated that he was-- . >> rose: he indicated or he said. >> he said that he was certainly going to raise it with the chinese. that he hoped they would play ball. and essentially be cooperative. and work together. he said that if that happens, that could be very good, dropping hints about the trade issue. and that if it doesn't, it was going to be bad. and he indicated, or he said that the u.s. would go it alone.
4:08 pm
later on he also had-- . >> rose: when he said it, did he say that's all i'm going to say? >> that was always a concern. he said they would go it alone. did he not indicate whether that meant militarily, sanctions or anything else because frankly, i don't think he knows. but he is certainly saber rattling. a lot of people would say could the u.s. go it alone. could the u.s. take military action to contain a nuclear threat from north korea. the answer is at the moment, not easily or not without a huge loss of life on the korean peninsula. could the u.s. use sanctions to reign in north korea, well, it's been trying to do that for some time, it's not worked. so there aren't any easy options for the u.s. to go it alone. >> rose: does this president believe that north korea, if it has icbm missiles that can deliver a nuclear warhead to the united states, is his biggest foreign policy challenge? does he know that? does he say that? does he believe that? >> well, i would hazard a guess
4:09 pm
that until recently president donald trump wouldn't have been able to locate pyongyang on a map. but i think he certainly does understand the gravity of the issue. and one very important detail about what is going on around the white house right now is that the -- working on a review of the problems of north korea and the options for a response, is actually accelerated it to make sure the review is in the president's hands before the meeting in mar-a-lago. and in fact i believe it was probably handled friday. and we also interviewed katie mcfarland, the deputy. and she made it clear that she now thinks, as well as the nse now think it is entirely possible that north korea will have an interkonl ballistic miss will nuclear warheads that could hit the west coast of america by the end of president trump's first term if the current tra jectory goes on unchecked that is a scary thought. i was in san francisco and los
4:10 pm
angeles last week. >> rose: a very scary thought, that is why president obama said to him your biggest k458 eng will be north korea. >> the thing i find amazing i was in san francisco ands will ang less last week and i said to people in silicon valley, boom boom sunshine, everything is wonderful, i said to a couple of people, are you worried about north korea. and i might as well have said are you worried about planet pluto. i don't think people have understood the threat at all. >> rose: and did you talk about a one-china policy? >> we talked about not so much one china policy. we talked about the wider question of the asian region. but we didn't get into details about what his policy would be. the most intriguing comment he made during the interview was, and i quote, that it's possible there could be a very big deal, that there could be some kind of bigger deal. i would not be at all surprised if we did something that would be very dramatic and good for both countries and i hope so. >> rose: that sounds like a grand bargain that he's thinking about. >> you can dismiss that and say
4:11 pm
that's donald trump, off-the-cuff statement. or you can say that actually, if you want to be optimistic, the very severity of the north korean threat means that not only the u.s. is concerned right now but also china is pretty concerned too. maybe just maybe, this could be the issue that breaks a logjam and forces these two huge countries to recognize they need to work together. not just on the immediate issue of north korea, but on a wider resolution in the south china sea, in the korean pen insurance will and maybe on the trade front. >> rose: and also climate. >> and also climate. now it sounds hard to believe. and frankly i find it hard to believe. but there is a 7 point plan floating around which would try to bring about some kind of wider bargain across the korean peninsula, probably wouldn't be acceptable to the japanese, let alone south korea be-- koreans but it is floating around. so when donald trump says i would not be at all surprised if we didn't do something that would be very dramatic, maybe he
4:12 pm
has seen that plan. >> rose: what was his overall demeanor? >> did he-- was he calm? soft spoken. >> the president's overall demeanor seemed to be very positive. very cheerful. he was in a good mood. he was keen to be charming. but also firm. and you know-- . >> rose: keen to be charming, that is a british expression. >> actually, we are-- he has a spot for the button on the desk which looks rather uncanily. >> rose: friday afternoon. >> looked like something out of dr. strange love, box on the desk, we looked at the button and he pushes the button and he calls up a diet coke. he had made jokes sometimes about-- so we-- he doesn't dring diet cola. so we had diet cokes. he showed us a picture of jackson. he showed us around the office. he was very keen to, you know, be hospitable and gracious and share good wil. >> rose: what do you make of his presidency so far?
4:13 pm
>> his presidency has certainly taken the rest of the world aback. it has been unconventional. it's been often frustrating and baffling. it's been exhausting as a journalist to cover it it because we're dealing with tweets at all hours of day and night. however, as we wrote in our piece today, he is an imper yas president. he's an unconventional president. but there is a method in the madness, we think. >> rose: what is the method. >> the method in the madness is really two of three things. first to enable him to connect with the voters that he thinks brought him into office. secondly, it is to if you like destabilize or surprise the people who he regards as his rivals, i will say that includes the media, and thirdly, it's to show that he is going to be unconventional. he wants to be a different type of president. >> rose: so he believes in chaos and disruption? >> i don't know whether he had ever sat down and created a philosophy of chaos and philosophy but certainly his instincts are to be
4:14 pm
unconventional. we asked him, or rather i asked him whether he regretted any of these tweets that have been coming out. he said he doesn't do regrets. but he also said he's convinced that the tweets won him the election. the tweets are essentially whattive goes him the power of the presidency and he's not going to stop any time soon. >> rose: how did it win the election? >> basically-- . >> rose: being disruptived. >> it enabled him to connect with the public. at one point in the interview he turned around and he said tell me how many followers i've got now on twitter. and they came back and said a hundred million. then corrected him and said no, 101 million. and he was very pleased. >> rose: why do you think he's still sensitive to any questions about winning the election and constantly refers to how well he did and all of that? do you have any psychological insight on that. >> actually as we were writing the piece he opened the sper view saying i won, you didn't. and you know, it is an indication of his mentality. i would say he puts the media in the camp of people with whom he
4:15 pm
has been battling over the last couple of years. i think he still feels the need to assert his position. and again, many people would say that is part of the unconventional strategy. some people would say that shows a man who doesn't feel secure in the the job yet. others would say well, actually, that is just being human and part of what has enabled him to connect with voters, is the fact that he is a no-holds barred, tell it how it is, very, very human president who changes his mind, says what he is el tooing, in a pretty unfiltered way. >> rose: do you believe, because this is in your arena, do you believe that the economic progress that we have seen especially the market, will continue? >> that's a very interesting question, charlie. because during the course of the interview, we asked him about what history usms were, we did speak to a large a writer of people in the white house.
4:16 pm
and they said well, confidence is surging. stock mar kelts are up. basically the mood has changed dramically. and people in the white house regard that as a strong sign of what they have achieved. they could be right. maybe there's go to be a self-reinforcing upsurge of animal spirits that actually go get the economy growing again. but one thing is very interesting right now and it's something that is worrying economists. there is a big split between what they call the hard data and the soft data. the soft data is about sentedment, about consumer confidence and business confidence, and that is definitely rising. the hard data is about what is happening. >> rose: because they think there will be tax cuts and less regulation. >> people are fed up of too many regulations the last several years and worried about quantitative easing. the hard data is not nearly as optimistic. that may be a time laps, maybe the hard data will improve as confidence feeds lieu. >> rose: and the hard data is
4:17 pm
simply. >> things like auto sales we had today, gdp, retail send spending. none of these figures have been optimistic, that positive. they have been okay, but not great. so the question people are asking is is this just a time lag effect and when businesses and consumers see president trump's policies actually coming through, the actual underlying data will improve dramically too. or is it the case that people have drunk the kool-aid around the presidency, and there's going to be a very nasty bought of disappointment going footer. >> rose: there is this idea still, well, we got, we being everybody, got his election wrong and therefore you can't really say that anything is bad now because in the end, like the election, he'll prevail. >> at one point in the interview president trump pointed out that, you know, he predicted brexit, predicted all kinds of things. and he has been more correct in his prediption-- predictions than many other people including the media. and certainly the white house thinks that they are on the the
4:18 pm
right track, that many american voters with them. business confidence is rising, and their strategy is pork working. >> rose: there is also this which they obviously know. he does feel a strong connection to the people who elected him. or at least he gives expression to that. yet some of the things that he's doing, the proposed budget, and health care, help, i mean hinder, hurt, take away things that very constituency has considered important to their lives. >> i could not agree more, charlie. because there is a par a docks. he has been-- paradox, he has been elected as a man of the people. is he not a man of the people, he is a billionaire as he keeps telling us. is he not behaving as somebody without wants to put the people's interests first in terms of the poorest people, in the sense that, you know, the tax cuts will end up favoring the rich, not the poor. his decision about health-care reforms are things that actually would not help many of his
4:19 pm
voters. >> rose: people in his cabinet and the principal advisors are all very rich as we saw an article over the weekend. >> extraordinary paradox. you can be cynical and say that is one reason he keeps stressing the tweets. he keeps talking like somebody from the people. >> rose: how about the specific tweet about president obama wanting to, or having been part of wiretapping trurch tower. >> we didn't discuss that particular one. >> rose: you didn't raise that particular one. >> no. >> rose: was that on purpose or. >> no, it's because you go into an interview, and frankly, i mean we talked about the tweets. he said at one stage if you are going to make hundreds of tweets you are going to make a-- now and then we were about to talk about and started on north korea and frankly right now issues like north korea and china are so important on the agenda, that as with any interview, you go where the conversation goes. >> rose: he faces a congress
4:20 pm
now in which he has said he's going to take on the freedom caucus. >> exactly. one the very interesting details we asked him about was what about health-care reform. is it dead and buried. everybody in the white house was keen to stress it's noted dead and buried. they claim there is still a lot of negotiations still going on. president trump himself stressed that if the freedom caucus were not going to play ball, we turn to the democrats instead and try and get what he could. >> rose: he said that to you. >> he did, yeah. and get a bipartisan deal. if we don't get what we want, ie from the freedom caucus, we will make a deal with the democrats. and we'll have in my opinion not as good a form of health care but we're going to have a very good form of health care. it will be a bipartisan form of health care. >> rose: so he's prepared to take them on. >> judging from what he said. >> rose: so far there is no indication-- one of the things that came out of the problems with passing health care was that the freedom caucus members are not intimidated by him.
4:21 pm
>> yeah, that much is very clear. >> rose: exactly. >> an when he says, i think he's going to be ruthlessly pragmatic. and you know, we were told several times that negotiations are still under way. and they will continue. and it was clear, we were told that the reason they had not conducted the vote was not because they were scared of losing. it is because they wanted to keep the door open for more negotiations. like so much coming out of the white house right now, it's very hard to tell what is actually true or not. >> rose: based on the moment he walked into the white house and the moment you left the white house, what did you learn that you didn't know or what perceptions did you have that were not confirmed and in fact new ideas, new impressions were formed? >> well, i think they basically come down to two or three. the first one is the issue of north korea is center stage right now. and very important and very certificate yution. i knew it was certificate quus before hand but i came out, you know, aware that it was actually really serious.
4:22 pm
>> rose: and that they were on top of it. >> they certainly planned to make it a centerpiece of the summit in mar-a-lago. i think number two or three weeks ago, people were looking at this more in terms of the trade deal, the trade issue, the currency issue. so that is point one. point two is it was reaffirmed that he really does believe that tweeting, his unorthodox style is working. he's not going to change it. anybody-- . >> rose: even though. >> any journalist who hope these tweets stop, no chance of that. >> rose: even though it creates distractions that get in the way of his message. >> again, we asked him that, you know, i don't regret anything because there is nothing you can do about it. you know if you issue hundreds of tweets and once in a while you have a clinger, that's not so bad. that pretty much-- . >> rose: but it's bad if it distracts from the central message because that's what. >> he says i have over a hundred million foarl between facebook, twitter and instagram, over a hundred million. i don't have to go to the fake media. >> rose: what is the fake
4:23 pm
media? >> you and me rdz i know that. but broadly-- broadly it is anybody who writes things that he doesn't believe. >> anything that is basically dis-- him and the voters. you know n some ways it is a bit like franklin roosevelt with the radio back in the 1930s. >> rose: kennedy with television. >> exactly. he wants to go directly to 4eus base. >> rose: fdr radio, kennedy television. >> the fake media is anything that gets in the way of that. >> rose: and obama, i guess, with social media. >> exactly. we've had many presidents who have actually used new technological platforms to connect with the voters and get around. and every time it happened, people have complained. remember, roosevelt in the 1930s, it was a huge outkrie from the mainstream newspapers at the time about the idea of doing radio broadcasts. >> rose: there has been more military engagement. >> dpakly. >> rose: started by president obama but he has continued that. and his secretary of defense has put in more people, more people
4:24 pm
that are closer to the front lines and more people who are taking in iraqi soldiers and others. >> absolutely. i mean we have-- . >> rose: with the risk of being killed. >> absolutely. i mean we have an increasingly militarized government. however you want to put it. we have a number of military figures sitting at the core of the government. we have pledges to increase military spending, this came up a lot, that basically there is a need for more military spending, almost above all else. and there is clearly, you know, a focus on war. and there is a determination to talk about war in an abrasive way that we have not seen for quite a long time. so you know, striking. now he also says in the interview that he wants to have alliances. he goes out of his way when you ask him what surprised me from this et moog. he goes out of his way to stress that he says nice things about the chinese leadership. he says nice things about angela merkel.
4:25 pm
he tried to deny that he ever didn't shake hands with her and said he didn't hear the reporters questions. so in some ways what you are starting to see from president trump is a patent where he talks tough and aggressively and then often climbs back down trying to play-- . >> rose: i wouldn't significant that has been his strategy. if you look at art of the deal, that is what he-- that's what he lays out. >> he's not a details man. at one point he tossed out the figure talking about china and other countries. he said look where we are, we have an 800 billion trade deficit. we actually quickly fact checked him and it is actually just over 500 billion. >> rose: trade def-- deficit with china. >> with all the countries, with china it is 347 with china. so again, if are you looking for signs of how the man operates, strong posturing, strong words, not detail focused, not giving you policy prescriptions. i mean this was not a carefully crafted manifesto. it is about as far away as you
4:26 pm
can possibly get in style from president obama. >> rose: remains to be seen. >> he is the ultimate anti-wonk. >> rose: thank you. >> thank you. >> rose: gillian tett from the financial times with a big scoop and a friday interview. we'll be right back. stay with us. siddhartha mukherjee sheer, assistant professor of medicine from columbia and cancer physician and researcher. he is also a very good writer, a pulitzer prize winning author. his latest book the gene an intimate history comes out in paperback this month. he writes about the future of automated medicine in a new piece for "the new yorker" magazine. it is called the algorithm will see you now. i am pleased to have siddhartha mukherjee back at the stable. welcome. >> thank you, charlie. >> with some sense of fraternity, tell me how this piece came about. >> this piece came about because you know, it is a conference
4:27 pm
that you organized. >> rose: right. >> every fall, and i was talking at this conference to sebastian-- who works on ai. and sebastian told me, you know, i've been working on this thing in ai in medicine. and so my ears perked up. we were at the conference and my ears perked up. i said what are you doing? and sebastian than began to describe sort of an early version, this wasn't there yet. it certainly hadn't been published in nature as it eventually was. but it instantly caught my attention as a physician, what if we could use these powerful technologies, deep learning paradigms to start doing diagnosis. what would happen to radiology, what will happen to determine tolling. what will happen to doctors once we engage powerful computers to aid in diagnosis. so that is sort of the birth. >> that was-- so it began with a what it?
4:28 pm
>> yeah, it began with a what if, yeah. >> and what did you discover? >> one of the questions that was raised almost immediately for me was you know, once we enter this space, where do we stop? where do we start and where do we stop. as this conversation proceeded that afternoon, at this conference that you organized, the as the conversation preceded it became clear to me that the ambitions of these diagnostic technologies was wider or wider than hi imagined. so for instance, you know, sebastian talked about a mirror that would photograph you every day and using these technologies, map every growing mold in your body, or sitting in a bathtub, where you know you would have a scan performed at granular detail and it could figure out because it had learned, it could figure out what was growing and not growing. >> rose: so if there was a mall ig nant cell it could see it it grow.
4:29 pm
>> basically, to me it raised a whole series of questions. questions number one, what was the-- could this be accurate? could this-- could you accurately, could you train a machine to recognize a melanoma and distinning wsh a melanoma from a benign skin lesion, whatever it might be. that was one question. but the second question, i think is very, very important. the second question is what if you overdiagnosed. what if you were sitting in this hall of mirrors, constantly being surveyed, this kind of big data is watching you. what would happen? i mean would we start intervening on cancers that we would not otherwise intervene on. maybe those cancers are harmless. so for an oncologist, for me t raises the whole question of diagnosis. you know, where are we going with diagnosis. are we going to overdiagnose, are we going to start invading in the body in a way we hadn't expected to just because as machines learn, they'll keep telling us. >> rose: help me understand the difference. i will always believe that all
4:30 pm
knowledge is good. the more you knew, the better you were. and you are saying here, if the more you know causes you to take certain kind-- too many tests. >> that's exactly right. and in fact, let me turn it back to you and say all knowledge isn't good in medicine. there are things you don't need to know about. >> rose: that's one thing, say you don't need to know about, i'm not talking about need. i'm talking about what harm does it do. >> the harm might be overdiagnosis can cause severe harm. you can cut up pieces, the body can buy op see them, there is economic cost, there is consequently cost to it. you might be chasing cancers. we now know from autopsy series that many people will die of cancers that are incidental. so in other words a person dies of an automobile accident. you do an autopsy and you find all these cancers in their bodies, this has been documented, many of them are incidental. they will not die of these cancers but imagine if a machine was telling you i found a cancer
4:31 pm
there, here, i found it. but it may not be able to tell you, you know, is that cancer going to metastasize, is it going to become agressive. so you have the possibility of entering into this kind of hall of mer mirrors. so that is where diagnosis, it, you have to think about diagnosis very clearly. now the paper that tipped me off is the study in nature which is can you teach a machine to learn how to distinguish a benign skik about it in this article. one way is giving, knowing fact, people say you know, this fact is true, melanomas have these character statistics, so if you see these character statistics, they are asymmetric, they have funny borders, their die am ter is some big and their color, so
4:32 pm
abcd. so those are rules. that is a rule based learning. but if you really think about how doctors dying diagnose t makes-- doctors often make a transition from these rule-based learning to a pattern-based learning. so we begin to pick up patterns. we begin to figure out that melanomas, we can't tell you exactly but that looks like a melanoma, that doesn't look like a melanoma. that looks like cancer. that doesn't look like cancer. this patient looks like he or she has significant heart disease, this patient looks like she doesn't. what is amazing about these, what i found amazing about these deep learning algorithms is you don't teach them any rules. you just give examples, here say melanoma, here is not one. you figure it out. >> that is what they do, what one part of ai is about. >> you learn by, you say go from here to there. and you learn by the experience of going there. >> absolutely. and it is sort of t is a knowing how. the machine sort of just maybe,
4:33 pm
we don't know exactly how the brain works, of course. but perhaps this is how the brain, perhaps by adjusting weights and balances and along these categories, i give the example in the piece, about a child learning, this is a dog and that is a wolf. and the question is how does a child know that is a dog and that is a wolf, right? the child knows by, one thinks, by using a hundred categories of dog and a hundred pictures of a wolf and beginning to make their own judgements about what characterizes a dog and what characterizes a wolf. we don't say to a child here are the rules for dogs. and those are the rules for wolf. we say here is an example of a dog. here is an example of a wolf. you figure out what the difference between them is. >> so that is exactly in that part, doing this, that is exactly what happens and it offers the most far reaching some will say, because you are
4:34 pm
actually allowing the machine. >> yes. >> to learn. >> thanked is. >> are you not just feeding it a huge amount of data. >> right. >> you are feeding it a huge amount of data. are you not feeding it rules. >> exactly. >> so that is the crucial difference between the machine figures out basically that the algorithm figures out how to adjust its own weights and balances. and ultimately capture those categories with a great degree of accuracy. this is what as fonnished, going back to that conversation that i had, you know, with the ai groups. and then i began to see the real results. and the real results is that the machine outperforms seasoned dermatologists. and jeffrey hinton who i spoke to who is considered the father of a certain kind of deep learning, jeffrey hineton thinks that these machines will outperform radiol guests. and then after that gentic
4:35 pm
diagnosis. after, the idea that now we can look at genes and figure out, you know, this is truly a cancer that will likely metastasized. we've really entered oregon to enter a powerful diagnostic era. >> rose: it it will be able to determine early on whether a gene has the possibility of. >> becoming so, it will be able to determine whether a lesion, pathological lesion, a melanoma has the possibility of becoming aggressive and metastasize rapidly. it is very significant, for instance, as you know very well in prostate cancer. prostate cancer is a great example of some prostate cancers will never be aggressive. they will be indole ent. other prostate cancers with be extremely aggressive, metastasize to bones and cause terrible disease. it would be astonishing if a combination of deep learning plus genetic technologies plus real doctors, one would hope,
4:36 pm
plus real doctors would show be able to tell patients how to distinguish between one and the other. >> rose: my impression is some of this is already being used in the case of radiologists. >> yes, it is important, so the-- . >> rose: some kind of ai is being used. >> the kind of ai that is being used so far in-- as far as that, i went and sat through as part t through some of thesee this, i demonstrations. the ones that are in real clinical use, sort of really being workhorsed, would be hard it to call them intelligence. they are artificial but they are not intelligent. >> rose: what is it then they have? >> well, they have speed, they have mechanisms to spot, they can guide you to spot an area, for instance in mammography, they can guide you. what they cannot do, i make an important distinction here, that machine, the ones that, i would call it it first generation tools.
4:37 pm
when they-- when the cat-- even if that is seen 4,000 images trk is not smart than the one-- it is rule based, based on telling the machine some rules f it is this big, flag it. but the important distinction here, the radical advance here is that these machines, are not told any rules. they figure out the rules for themselves. and that is what is important. because we're not telling them what to think. they are thinking if you can call it thinking, they are doing it by themselves. that is what is astonishing. >> so what does algorithm teach them? >> it, as far as i understand it it, i'm more on the medical end than the computer end. but as far as i understand it, it allows them to-- if you give them, if you give the machine, if you say here is category one, and here is category two, the algorithm spot features in category one and category two, extracts those features from these two categories and then
4:38 pm
begins to build as it were, a mechanism to distinguish category one and category two so it extracts the features it says, you know, what is amazing about it, and here is what is amazing, this took me by surprise. it cannot tell you, it's not easy for you to go back and query the machine and say what are those features. you can do that to a human being. you can go back to a dermatologist and ask her or him. you can say well, what-- you learn now what are the features that you were picking up. and they will give you some answers. they may be right, they may be wrong. they can give you some answers, what is astonishing is it is difficult to do this to a machine. the machine knows but it cannot tell you. and that's what is amazing. it is a black box. >> rose: so where is all this going? are we going to now go to any doctor, and therefore is he going to be or she is going to be in partnership with whatever
4:39 pm
the latest developments are in artificial intelligence and therefore it becomes their partner in diagnose were diagnosis. >> most people think more than that. most people think it it will be a partner in diagnosis, a partner in treatment. so there will be-- you know, in a discipline which is extremely information rich, pattern rich and whether the costs of making mistakes are very high, medicine, most people think that we will sently begin to use computers as our expend-- extended arm. it will be an entirely sim biotic relationship. that we will have a-- and gene sequencing and you know other technologies. we will become symbiance in medicine with these kinds of technologies it is not just going to be relegated to diagnosis it will move right through. >> rose: when i tell you this now you will tell me it has no relevance at all. but i am constantly amazed about
4:40 pm
how the accumulation of huge amounts of data, is serving every end. there is a story in "the new york times" today or yesterday about how golfers are now able to accumulate data like they've never done before. >> right. >> in terms of understanding a, what they are doing wrong. and b, how 20 play the game better. >> that's absolutely correct. but i want to talk you towards the ends of the piece where i raise the question about big data. which is big data in this case is absolutely necessary to teach these algorithms how to distinguish between the categories. if you need it 2,000 examples of melanoma and 2,000 examples of nonmelanoma it will keep learning. but big data is not very good at, might be better at soon but is not very good at is it can't help us answer why. it it doesn't help us understand the patterns behind the patterns. and that is where, i think, human beings still have.
4:41 pm
>> is that likely to change? >> that's a big question. you know, as i gave you again the example of a child who distinguishes between a dog and a cat or a dog and a wolf. but then the child, our children, we perform additional functions. we ask the question why is a dog different from a cat? we open our own black boxes. and say what makes a dog different from a cat. i give the example in this, you know, of a baseball player who knows exactly where the ball will land because he or she has thrown the ball a million times. if you say landed exactly there, will fall right at the spot but that person may not know newton's laws. they don't know why the ball is landing at the spot. they have become using their mental and physical algorithm, they have become absolutely expert in landing the ball at the spot.
4:42 pm
and the question i raise in the piece is, you know, as we become more and more, as we depend more and more on the baseball player model of landing the ball in a spot, will we start losing the physicists, the person who takes the step back and says the reason that the ball landed there is was because there is a hidden law behind all of this. >> rose: and the reason i assume without knowing, the reason that is important is so you can therefore apply that physics to this-- to another area, another area, another area. >> absolutely. and the reason in medicine is you can apply 245 fundamental knowledge to figuring out, you know, how do you address the immune system against that melanoma that you just spotted. why does the heart behave in this day. how can we make new medicine. so there is a is he duction of big data, there are certainly deep solutions that come out of big data but at the end of the piece i raise a question about you know, there is a is he duction of big data too.
4:43 pm
online you can find the piece, it's called ai versus md. >> right. >> and you know, of course, it is a bit of a provocative joke because in the end it will be ai plus md. we are going to fuse. but there is a sense at least within the medical community that there might be a law in this. not only the problems of overdiagnosis. we talked about those already. but the possibility that you know, that we will lose a kind of way of thinking if we become very reliant on these devices. on the other hand, there are obviously great benefits. imagine being able to diagnose melanoma with a great degree of accuracy or some other cancer. >> where time is significant. >> where time is significant. and time costs money. >> but still i go back to the point, if you can teach a computer how to learn so the machine knows how to learn, and
4:44 pm
the machine learns on its own, won't it be able to learn about the laws of physics wnd well, so that is you are asking some of the deeper questions in deep learning, as it were. you know, i mean again, from the standpoint of medicine, which is sort of where i would come from, from the standpoint of medicine f it would remain a challenge to figure out deeper laws of physiology. but i would, it is a question that the deep learning people are working on. i know this from having spent significant amount of time talking to them about how machines learn about medicine. >> well, i will close with this. i did a piece for 60 minutes about artificial intelligence and watson which is the ibm variation of all of this. what is amazing to me is that you go to most companies in america that are the least bit sophisticated and they are now,
4:45 pm
have dedicated significant development funds to figuring out how ai influences their business. >> and medicine has gt to be one of them. >> medicine has to be that. you go to most investment firms and they're looking at places. and they are trying to put together funds that look exactly into ai. i mean st perhaps the most discussed single topic in terms of the fuse, there are other kinds of areas in which you can talk about. but it has enormous focus right now from a whole range of people. who understand its power. and those people who understand how far that power may go. >> genetic technology is ai. these two things are changing what human beings will be like in the future. there's no doubt about that. so something to know about. and particularly in medicine. >> thank you for coming. >> my pleasure, thank you so much. >> let me point out the gene, an int mass history, a remarkable book, now out in paperback.
4:46 pm
back in a moment. stay with us. >> last week president trump signed an executive order designed to unravel former president obama's climate change initiatives and also to revive the kohlhepp industry. president trump made it clear that his administration has no intention of living up to the commitments laid out in the 2015 paris agreement in what many consider a major role reversal, china appears poised to step into the leadership void created by the president. joining me now is barbara finamore a senior lawyer an asia director at the natural resources defense council. we're pleased to have her here. welcome. >> my pleasure. >> rose: tell me the impact of what the president has done with, as a consequence of these executive orders. >> trump has announced a slew of new executive orders designed to dismantle the obama administration's energy, climate and clean air actions. >> rose: lay out what he as
4:47 pm
dismantled. >> well, he has called for rolling back the clean power plant that sets limits on what power plants in the united states can emit in terms of co2 emissions. it is not clear that he will be able to do that any time soon, however. because he has to follow the law in the same way that it took, to develop those regulations, will have to put that out for public comment and give good reason for why this needs to be dismantled and then also defend that action in court. so that may take quite some time. during that time, it is expected that these renewable energy business in the united states will be the fastest growing in the electricity sector. trump has also called for-- . >> rose: that's good then, isn't it. >> that is a very good thing. that is a very good thing. so it is unclear what type of impact his actions will have. he's trying to prop up the
4:48 pm
kohlhepp industry and despite exhaustive analyses by many showing that it is not going to come back. the kohlhepp industry is not going to-- . >> rose: the thrust of that for him is simply to create jobs. >> yes, but he is ignoring the fact that today in the united states the solar industry alone employs roughly twice as many people as the oil, kohlhepp and natural gas industries combined. so it's not really going to grow jobs. in fact, i think the main impact of what trump has announced is that he is going to put the u.s. in danger of losing out on the biggest new global energy-- the big new global energy of any kind which is the transician to clean energy. >> but in fact tom freedman wrote a colume about that, about that very idea, that china will become a leader where the united states should be. and was poised to be. but china coming on so strong because of the huge size of its
4:49 pm
issues and its problems, you know, has therefore been forced to accelerate alternative forking sources. >> that's true. china is the leader in investment and wind and solar and other renewable energies. $88 billion last year compared to the united states, 58 billion. it plans to invest in other $361 billion dollars in renewable energy by 2020, creating 13 million new jobs. so that is the story that trump is missing, that the renewable energy industry is where to go if you want to create new jobs. >> has he, when you look at his decisions, is it likely that he will pull out of the paris agreement? >> well, there seems to be an internal debate. rex tillerson, secretary of state among others has said it's probably best if the u.s. keeps a seat at the table. >> not only that, exxon mobile the company that he headed has made that point clear. >> and a whole slew of u.s. companies have also made that
4:50 pm
point, and the insurance business and so forth. understand that the u.s. has to take a leading role in moving forward on climate. >> what happens to paris if the u.s. should pull out? >> if paris, if trump announces that he wants to step back from the paris agreement, it's going to take several years probably until the end of his administration. but it's clear to us that many many countries are going to move ahead despite what the u.s. does in fact, at least, i don't know, a dozen countries have joined the paris agreement since trump was elected. china has made it clear that they as a leading emitter of greenhouse gases are going to carry ahead with their commitments no matter what the u.s. does. >> will they move to the forefront of being the most rapid, developer of alternative sources of energy and at the same time being the leader of the sort of effort to do something about climate control on a global basis? >> that's right. there are two aspects to that. one is cutting down china's
4:51 pm
reliance on kohlhepp which is the leading sour of its co2 commission an its air pollution. at the national people's congress session earlier this month, china, the premier announced that cutting back on china's excess kohlhepp capacity is a national priority. so china is doing it in a whole range of ways including cutting back on 280 million tons of excess kohlhepp capacity and new permitting for koa plants dropped 85% last year. china has put a cap on kohlhepp. so that's 58% of its total energy mix by 2020. that is one half of it. >> so clearly they, because of the size of who they are and what they do, the idea of leadership is thrurs upon them. is it something that they want to own? do they want to be the leader? >> yes, they do. climate change there is a whole slew of reasons why they are acting on climate now. one is their own studies scientific studies show that china is one of the countries
4:52 pm
that is most vulnerable to the impact of climate change. sea level rise of its koasessal cities, very highly pop lated. it it feeds 22% of the world's population on 70% of it'sariable land and food production is already going down. and air pollution is a key problem, it's one of the worst in the world. 366 million people die every year from air pollution. and so china also sees acting on climate in its economic interest, it is in the midst of a fundamental transformation of its economic system, to one that is reliant on kohlhepp-fired heavy industry to one that is reliant more on the service sector. >> rose: it is also moving from an exporting economy to a dom ease particular consumption economy. >> yes, that's right. and it recognizes that it has to deal with its massive overcapacity in kohlhepp plants t was building two to three kohlhepp plants a week at one point. but now these planteds are not
4:53 pm
even running half of the time. if china doesn't get rid of that overcapacity it faces up to one trillion dollars in economic losses from stranded assets. >> rose: how much of the in in technology and how much of innovation in alternative fuels is taking place in the united states. >> the united states has always been a leader in the development of the technology, the brains behind the technology development. and that can still go on if the trump administration continues to fund research and development on those. what china has done is take that, and through economies of scale, take the lead in manufacturing. even though they are still, as i said before, many many jobs in the united states in solar industry. but that is where, you know, the u.s. and china have something called the clean energy research center where the two countries have put their best scientists together to jointly develop new technology on things like electric vehicles, carbon capture and sequestration, building energy efficiency and
4:54 pm
have sign add greements to share the intellectual property that is the way to move ahead. >> so you can argue the question is whether the united states is prepared under the presented administration to be a global leader in alternative fuels and other aspects of that, and therefore lead what is considered one of the new emerging industries of our time. >> you know, the u.s. can lead. and if trump wants to, he can use funding to retrain the workers in the kohlhepp industry. it's a dying industry. because not because of obama's environmental regulations but because of the increasingly cheap natural gas and renewable energy of the united states. he can use that funding to help retrain these workers to fund r & d for new technologies of the future, and that way continue to assert its leadership.
4:55 pm
we are competitive in all of these industries at this moment. >> rose: thank you for coming. >> thank you so much. >> rose: pleasure to have you here. thank you for joining us. see you next time. >> for more about this program line at pbs.org and charlieus rose.com. captioning sponsored by rose communications captioned by media access group at wgbh access.wgbh.org
4:56 pm
>> rose: funding for "charlie rose" has been provided by: >> and by bloomberg, a provider of multimedia news and information services worldwide.
4:57 pm
4:58 pm
4:59 pm
5:00 pm
this is "nightly business report" with tyler mathieson and sue herrera. >> tensions rise. president trump will meet with china's president in just a few days and now we know the controversial topics the two will talk about. >> sudden resignation, a long-time fed officially stepping down abruptly over his involvement in a 2012 case where confidential fed information may have been leaked. tax items with two day weeks left to file, those stories tonight on "nightly business report" for tuesday, april 4th. >> good evening, everyone, welcome, investors know that there the a lot riding on the economic relationship between the united states and china.

14 Views

info Stream Only

Uploaded by TV Archive on