Is AI gonna rip us apart ?
Hey guys, just saw an article from Bloomberg about Machine Learning facilities at MS.
https://www.bloomberg.com/news/articles/2017-05-3…
Hey guys, I just saw an article from the Economist about Machine-Learning. According to the Economist, it seems that AI is already affecting BO/MO, S&T, and credit rating. Here is the link:
https://www.economist.com/news/finance-and-econom…
Below is the original post.
I personally hold an opinion that the over-worship of AI is a little dangerous, because after all this is a world of humans. We use AIs to help us instead of replacing us.
However, in college, at least my college, computer science majors increase exponentially. There are tons of my friends who want to apply/applied for deep learning/machine learning/human interaction graduate programs, all AI-related. Even a psychology major and a physics major did.
So no one thinks that this is a little crazy? Are banks adopting AIs now? How are we going to handle AI?
I would agree - it seems a lot more people are studying the software engineering/computer science/etc.
I think it is just going to lead to other avenues of work being opened, I don't think it would be a good idea to not develop AI for fear of it changing the industry.
Objective, SB given
Well you're posting in the investment banking forum section, so no, I don't think it's crazy. Yeah, there's quantitative trading, big data, and more, but it's very doubtful that AI will replace the actual traders, the MD's, or the industry. Computer Science is not Finance. Although perhaps they find things in common through Data Analytics. Banks can use technology and applications to improve their work or financial services, but not one has yet to touch something like say, blockchain technology. So if it's still taking time to get self-driving cars off the ground, then I think it's doubtful that AI technology will take over an entire banking industry.
I saw a picture a few days ago: GS has replaced a lot of traders with algorithm trading programmes. Is that true? The picture showed an almost-empty trading floor with just a handful people there.
You could be right and my perception on technology and banking could be flawed and/or outdated. I can't say on personal experience, but I've spoken with a few FinTech CIO's a few months ago and yeah, FinTech is definitely making a mark on the structure of the banking industry. It's definitely affecting the back office, possibly the front office, but personally like I said, I still don't think it'll completely consume all the MD's and what not. Compensation in IB is down by a lot, and this compensation may be moving to other jobs somewhere else in the industry, but remaining relevant to the topic of this thread, I don't fear the great reign of AI anytime soon.
I agree. It seems least likely that AI would overtake IB as compared to other areas in finance, given the importance of client relations. One could imagine human speculation being completely overtaken by computers, however. The fact is that a machine learning algorithm can make better, more consistent, and faster predictions than a human operating under the same principles. However, prediction is speculation, and it is unlikely, though not impossible, that a machine will be able to find true value. That being said, I've used some of the machine learning techniques, such as neural networks, random forests, and boosted trees, and they are truly astounding. After working with them, you inevitably have that frightening thought that they could do anything.
yes i do believe it is dangerous because you never know if these AIs are going to start coming up with thoughts of turning against humans.
Yes - S&T people would say "it's still a people business" yet head counts have gone down drastically and a lot of other trading jobs are obsolete. Therefore, I think it's more important than ever to keep skills and networks robust.
Also, there's a lot of AI and technology currently being developed for investment banking. There's a lot of fintech money in modeling; so perhaps in the future headcounts will fall.
you better LaVar Ball beeeelieve it!
i'll try to find the source for what i'm about to type, but basically there was a study done by (oxford? - again, will find and edit this) and it said the most-secure jobs from AI (if that is something that can even be quantified, looking 30 years into the future) jobs that were a) creative in nature and b) client-facing in nature fared the best in being the least replaceable.
Yea, that's why I feel IB will be around . After all banking or even ER is client service.
https://willrobotstakemyjob.com
Will AI wipe out the human race? (Originally Posted: 06/05/2017)
Elon Musk has said, "The risk of something seriously dangerous happening is in the five year time frame. 10 years at most The risk of something seriously dangerous happening is in the five year time frame. 10 years at most. Please note that I am normally super pro technology and have never raised this issue until recent months. This is not a case of crying wolf about something I don't understand."
"The pace of progress in artificial intelligence (I'm not referring to narrow AI) is incredibly fast," said Musk. "Unless you have direct exposure to groups like Deepmind, you have no idea how fast — it is growing at a pace close to exponential."
Ray Kurzweil, chief proponent of the singularity theory, predicts that we'll see the technological event come to pass sometime around 2045 — decades later than Musk's prediction. Singularity is the idea that artificial intelligence effectively becomes more intelligent in every way than human beings.
A lot of people are worried about artificial intelligence taking their jobs. Anyone else concerned that it'll mean the end of the human race, or at least some sort of cataclysmic event? Once machines become more intelligent than humans, isn't it in their best interest to collude against us?
No, haven't you seen Terminator?
I don't think it means the end of the world as we know it or for jobs, I think it'll create space for new jobs for creating the AI, for managing automation and new ideas on using AI. You could call it creative destruction, in one way old jobs are lost but news ones come in different ways.
I work for a AI company called Plum Fintech who automate savings for people, and I think its helped and created new opportunities, not less!
In that sense, however, wouldn't it be a fair assumption that much more jobs would be lost than would be created? In the United States, I think around 60% of the working population is involved in blue-collar labor, which AI could effectively replace in the future. If singularity theory were true, wouldn't it be feasible for AI down the road to create and manage themselves?
In Black Mirror, a theory surrounding the S1E2 is that AI has replaced almost all existing human functionality. Humans are left to indulge in entertainment and help to power the AI systems through cycling (an energy generation method being applied today).
Agriculture killed humanity. Overfed, underproductive populations. No wonder we have to "invent" jobs, or quite frankly, people would have nothing to do.
It is sure to happen in the coming days.
This Robot is a Game Changer for Fast Food (Originally Posted: 07/05/2017)
Giving up on life and going to work at Mcdonalds is slowly not becoming an option anymore. This new robot called “Flippy” is a robot prototype that receives orders from the front counter, and grills your food to perfection with its thermo sensing technology.
It can grill foods ranging from burgers, to chicken, to buns, which it can identify, in order to use different cooking instructions. It will also be able to do even more things like cutting onions and making French fries. There will still need to be a worker there to supervise and to wrap the food it makes, but a burger joint will require a lot less man power to operate.
The first commercially sold Flippy is predicted to be sold in the first quarter of 2018 to Caliburger in Los Angeles.
What does this mean for the millions of certified burger flippers working now? How long do you think until fast food restaurants will need almost no humans working? Are we becoming the Jetsons?
Article and video of Flippy
This just sounds like a long line of unemployment. How can somebody replace the burger flipping artists for a robot??
Nobody can replace the adrenaline rush I get when i check my burger to see if it's under cooked or burnt Lmao. But slowly all of the minimum wage jobs are being replaced because of how simple the they are. What does this mean for restaurants now that their service costs are slowly becoming fixed(With cashiers also being replaced by robots)
Article to end all the "Ermagerd Artificial Intelligence will take err jerbz!!!" posts (Originally Posted: 04/10/2017)
An article on Stratechery to end the ever-perpetual discussions on WSO (yes, I'm guilty of one too). Ben Thompson explains it spot on - nobody could have done it better.
https://stratechery.com/2017/the-arrival-of-artificial-intelligence/
I wouldn't be surprised if we worked 'harder' with AI in a sense. We might be doing our own farming, for example, using the technology. I certainly would want to if I had efficient tech to control what I eat and wear.
Nah this is the real problem.
https://en.wikipedia.org/wiki/P_versus_NP_problem
Never really thought that way....ermagerd!!!
What will you do when an AI replaces you? (Originally Posted: 04/05/2014)
For those of you who are still watching Ashton Kutcher put on a terrible act on Two and a Half Men, you might have seen an episode where the actor/investor was developing a software that spit out human thoughts. This part of the show might have inspired Kutcher, along with Tesla’s Elon Musk and Facebook’s Mark Zuckerberg to invest a total of $40 million in a covert artificial-intelligence start-up, Vicarious, whose mission is to “build a software that thinks and learns like a human.”
Little is known about the firm’s technology and business model, let alone its exact address, but it has been in development since 2010 after getting some seed money from PayPal’s Peter Thiel and a few other angel investors. The company claims that its software, Recursive Cortical Network, is able to interpret contents of photographs and videos the same way that humans do. This is perhaps nothing extraordinary but its long term goal is to make the technology cure diseases, create affordable renewable energy, and perform most human tasks. It is also supposedly working with the CIA on military projects.
It seems there’s a clear trend that tech firms are moving toward developing AI applications. Google is working on its robots, Amazon on drones delivery, and Facebook on producing a gigantic wisdom database. Some of these innovations obviously have huge potential and advantage but how would this affect human capital? I realize that it’s probably decades away until we are replaced, if possible, by computers. However, the large investment funding and geniuses behind these developments make me think if the time frame is even that long.
Any thoughts?
http://www.wired.com/wired/archive/8.04/joy.html
Sounds like a pipe dream for at least the next 20 years. Machine learning has been big in the AI and trading industry for some time now and I'm sure it can only continue to get better.
The thing I'm not sure can ever be completely replicated is intuition. Consider a time 100 years in the future and a robot is able to move and process data like any other human, except a lot quicker and more efficient. The robot comes upon a situation where it must perform an action, it collects as much possible data about its surroundings, processes it, performs the statistical analysis of every scenario to do such task based on past situations, and then executes the action with the highest probable odds of successfully completing the task. I think humans are able to experience a "sixth sense" (intuition) at times and just know when a situation isn't the same as the past. They'll go against the odds to perform a seemingly normal task.
Once robots are able to collect enough data of real life, maybe they'll be able to get that same sensation, but I doubt it.
/End hippie trippiness
Of course, the ability to "know" based on "intuition" that a scenario "isn't the same as the past" must be based on some apprehensible data about the situation which, putatively, the robot has not collected, violating the previous premise that the computer "collects as much data as possible".
Other than that ripple, I read this paragraph to mean "humans possess the competitive advantage of making wrong decisions ex ante". Which, I submit, is a dubious proposition.
I thought this was going to be another thread about S&T dying out.
AI will be able to do a lot of things humans do right now, but I would not be concerned about it from an overall employment perspective. The effect will be to push wages up as humans are freed from those tasks and can work on ever more productive areas.
That being said however, from an individual point of view you might be screwed if your only skills are related to a field that can get automated.
On another lightly related note, I really can't understand how people who do not study engineering can go about their lives totally ignorant of how the technology they use every day works. Must feel like living in a magical world.
"The effect will be to push wages up as humans are freed from those tasks and can work on ever more productive areas."
This is wrong. What productive areas ? The robots are already able to reap grappes more efficiently than about 90% of people working in vineyards, the last 10% percent working in extremly qualitative vineyards that will continue to employ workers instead of robots. This is the same in the automotive industry, this will be the case in textile when robots will be cheaper than Vietnamese... With IA the same thing will happen with lawyers, bankers, traders and so many other jobs that people think are protected today. Actualy robots and IA could progressively cut the jobs of the bottom 50%(or even more ?) of any given profession. What will we do with these people ? This is the question that we have to answer.
//www.youtube.com/embed/JtWcBP_l3m8
At around the 34 min mark, some prominent VCs debate machine learning as one of the top tech trends. I tend to agree with Jurvetson on this but that's probably biased by the techno-optimist in me.
TheSquale,
I stand by my comment about wages and human resources reallocation.
Humans never seem to stop trying to improve their standard of living and AI or robots are no different than any other technological advancement. When a task people used to do can be done more efficiently with capital equipment, that is great because it frees the time of the people to do something else, something perhaps no one has even thought about yet. We have seen that happening since the beginning of time and technology.
"What will we do with these people", that is the same everyone asked themselves when new farming technology was making jobs redundant in the western economies a few hundred years ago. USA is a very clear example, the economy went from having basically everyone working on agriculture to industry to services today all of that with a relatively low and constant unemployment rate. It was technological improvements together with the savings and investing needed to implement them which enabled the increase in productivity that allowed the human labour shifts. The net effect is ever increasing production and real wages.
Just because it is impossible for any individual to foresee the jobs that will appear in the future does not mean they will not come. In a free economy, future entrepreneurs will work on finding solutions to the new more refined consumer demands and new jobs will come with it.
tl;dr: No one could have predicted agriculture, the auto industry or silicon valley, but the fact is as technology improves productivity and makes jobs redundant, people find new ways to work to improve their living. The effect is more production and real wages.
People scared about "AI will take our jobs" will sound in a few years as stupid as these guys sound to us now: http://en.wikipedia.org/wiki/Luddite
+1
I get your point but no one can be sure that this will be the case. You're optimistic and I am not. The transition from agriculture to industry manufacturing resulted in an increasing number of jobs for the less qualified. With IA & robots this is different, we are basicaly suppressing the need for low skilled workers. I am sure we'll still need some workers that will work hand in hand with robots but not as many as we have today. Moreover I am based in Europe and I have somehow a different point of view about full employment and all. The problem we face is that there isn't work for everybody, even if no one want to say it out loud. And except in Germany if you look at our factories we are MUCH less automated than the USA. What will happen when managers will increasingly use robots because employees costs are too high ? I would also tone down a little the rethoric about the "full employment" situation in the USA, the labour participation rate in the USA is near 1978 level. Even if it's rebounded a little last month this is a very ugly picture and you can't honestly say that you have "low unemployment rate". Just look at this chart: http://www.zerohedge.com/sites/default/files/images/user5/imageroot/201…
"except it doesn't have to eat or sleep." right. or have sex. or do anything, for that matter, to survive. furthermore, there are no degrees of comfort for a robot's survival. It's on or it's off.
Any attempt to create a robot that can experience any genuine emotion at all will be a saccharine fallacy. Robots will continue to be the best and most complacent slaves humanity could ever hope for. As they become more intelligent and powerful, humans will compete to amass their ownership of these value-generating assets.
A purely rational being, not tethered to a transient, physical survival, is not a threat to humanity's quality of life.
This is how you know startups are in a bubble. Celebrities with no investment experience are getting in and funding wacky projects that have little chance of success in the next decade or so.
The secrecy may pertain to the CIA work or it may be that they don't have anything.
I am going to ask Reese Witherspoon to fund a time machine. We are going to go back to 1970 and sell modern PCs for $5000 apiece. We will also invent Disco and pet rocks. We will then invest $2 Billion of the proceeds in Apple Computers in 1978 (pre IPO) and generate $2 Trillion in wealth for our stockholders.
"Accidentally" delete the source code.
The automobile's invention temporarily put horse related workers out of a job. Car factory automation shifted skilled labor into service sector jobs.
Technology/innovation historically has created new industries and career paths. I'm wondering if today's innovation is inherently harmful to employment..or if it is more of the same.
Robert shiller said on a YouTube video that he is worried about the future. I'm excited to see how it turns out
Link to video?
Robert Shiller talks to Charlie Rose about inequa…:
Somewhere in the full interview he said he came back from a economic convention with "sweaty palms"
Ray Dalio recommends reading Lessons in History by William Durant which tells the reader that the wage gap widens under capitalism until some breaking point(revolts,war, wealth redistribution)
AI may expedite this process
Bob Shiller's a smart guy. To the point where I can't put an upper bound on how smart he is. There are things he can understand and connections he can form that I can't. And as a quant who worked on the trading floor of a major bank for several years, I'd like to think I'm a pretty smart guy.
I haven't seen the video, but my concern has always been that capitalism is always trying to squeeze more efficiency out of labor and is always trying to replace labor with capital. And over the past 15 years, capitalism has been doing an amazing job at it- perhaps too good of a job. And a lot of smart liberal leaning economists have raised a lot concern about that. Actually, Schumpeter, more of a conservative economist, raised it 70 years ago in Capitalism, Socialism, and Democracy ("Can Capitalism Survive?")
http://www.technologyreview.com/featuredstory/515926/how-technology-is-…
I was just thinking about the difference between my cell phone from 5 years ago vs today.
Amazing to say the least.
Advances in artificial intelligence now threaten white-collar jobs: Rise of the replicants ft (Originally Posted: 03/08/2014)
http://www.ft.com/cms/s/2/dc895d54-a2bf-11e3-9685-00144feab7de.html
Thoughts?
Might have to look into changing my belief structure to socialism...
http://en.wikipedia.org/wiki/Technological_singularity
http://www.technologyreview.com/view/425733/paul-allen-the-singularity-…
http://motherboard.vice.com/blog/for-one-second-a-supercomputer-mimicke…
Not likely in our prime working life. The complexities that arise from a plethora of human insanity induced problems will make developing an AI that can rationalize out problems created by humans incredibly difficult. Think about all of the cultural differences that exist in the US alone. Now task a geek with creating a program that sufficient self correct-ability for international cultural differences. The size and complexity this would take decades if not a couple hundred years for the programming ability to meet the real world demands.
On the other hand we can conceivably develop a AI that can scan financial documents for important information and present it for a human to correlate and interpret.
http://www.nbcnews.com/id/20676037/ns/technology_and_science-innovation…
@"IlliniProgrammer" Are these advances in artificial intelligence the first steps towards a socialist dictatorship administered by a supercomputer? Your time frame was 100 years IIRC.
@"NorthSider" What do you make of all this?
It is, perhaps, worth noting that I have consistently argued that people are underestimating, not overestimating the technical capabilities and functions that may be performed by computers in the near future. Many, restricted by their compartmentalization of computers as "strictly rational" and human actors as "capable of creativity", are unable to conceive of worlds where computers may perform what they arbitrarily classify as "human faculties". I do not share these reservations.
That said, it is yet another error of cognition to conceive of wealth as measured by "the number of jobs" or "level of employment". Wealth is measured by the abundance of resources, not the number of jobs. The two are, of course, correlated concepts in a loose sense; however, the reason why we are wealthier today than we were 100 years ago is not because we have "more labor" today, instead it is because we have astronomically greater productivity that has produced a commensurate decrease in the cost of goods. Today a half-second of labor on the average wage suffices to purchase an hour of reading light with CF bulbs; in 1800, with candles, that same hour cost six hours of labor.
It is no accident that the modern citizen works far less, yet consumes far more than his self-sufficient counterpart from the days of yore. If computers should substitute human labor in a particular trade, it is rightfully understood as an increase in the abundance of goods and thus an increase in wealth, not an ominous sign of obsolescence.
idk, having studied some machine learning and artificial intelligence (not an expert by any means) i am somewhat skeptical. i do think a lot of jobs can be replaced (doctor diagnostics, lawyers scanning documents)
however it's hard for me to imagine how a computer could create new mathematics or physics or come up with interesting ideas in finance.
everything i've read about machine learning seems to just be throwing in tons of data into an algorithm with the algorithm making updates to parameters based on some loss function to be optimized, seems way too constrained to do everything imo.
If anyone is truly f*cked, it is auditors.
http://web.mit.edu/cmagee/www/documents/29-singularitysdarticle.pdf http://research.caseyburkhardt.com/downloads/The%20Trajectory%20to%20th… https://www-rohan.sdsu.edu/faculty/vinge/misc/singularity.html
bump.
http://berglas.org/Articles/AIKillGrandchildren/AIKillGrandchildren.html
.
I created a thread a while ago regarding something about a company finding a way how to take out analysts from the equation and from what it seems people were optimistic that it will take a while before machines will be able to do that.
AI & the law (Originally Posted: 01/20/2018)
What happened
In a recent publication entitled “The Future Computed”, Microsoft said that a whole new field of laws and regulations will be needed to govern the application of Artificial Intelligence in a vast range of industries. This echo many opinion leader’s views, including that of Elon Musk, who said the government must regulate Artificial Intelligence “before it’s too late”, and that AI represents an “existential threat to humanity”.
However, with the many current applications of AI, as well as the huge potential in AI development; how visible is it to build a set of rules in order to ensure that the development of AI would not eventually lead to movie plots such as “Terminator” and “I, Robot”?
Leave a comment
Do you think there is a need for regulation in the development and use of AI? With companies of various industries and sizes pour money into AI investments, how do you think AI will affect our day to day life in 3 years? Should the manufacturer or user be held accountable for accidents that arise from the use of products with embedded Artificial Intelligence? Are laws only in place to govern human behaviors, or can it govern robots as well?
We're still decades away from general AI (the holy grail), and current applications of "artificial intelligence", even the impressive outcomes obtainable now with deep learning and neural networks, are only applicable in extremely niche situations (like playing board games or driving). Current commercialized applications of AI (machine learning) are mostly just sophisticated optimization algorithms trained with massive sets of data.
That doesn't mean that one day we won't get there. But if we start hamstringing our AI scientists now with useless and burdensome regulations, they'll just move away to innovate in a place with pretty much no regulatory oversight in tech and tons of capital (China) and we'll lose our technological edge.
There will be a day when regulation of general AI is necessary. But that day is years and years down the road.
AI Scientists shoot themselves - Program that programs themselves learns about SkyNet! [MIT Tech] (Originally Posted: 01/25/2017)
Programming is a grind for those who program or still program much more so than punching cells in Excel.
Object oriented programming is even worse! This is why programmers, especially those with AI and Machine Learning background, are in dire supply and mega demand even in the Fintech arena (Quants).
That's soon to change as these clever geeks code themselves out of jobs and thin out their own margins!
Check out OpenAI and the article linked below. Move in the right direction or are we building out the foundation to SkyNet?
Reference: AI Software Learns to Make AI Software
Vel ipsa est et dignissimos odit. Et quae velit et deserunt assumenda ducimus incidunt. Nostrum nihil dignissimos enim praesentium optio ea.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...
Perferendis dignissimos doloribus ad. Et quia atque facilis delectus illum est incidunt. Temporibus voluptatem sapiente voluptatem ut est quo. Enim dolores in voluptatem ullam in nihil est ab. Exercitationem itaque quia voluptatem quas.
Qui harum aliquam minima itaque eos. Quis dignissimos eos dolorum rerum rerum.
Debitis non omnis facilis placeat. Corporis nostrum tenetur harum sequi. Et voluptatem sed et quis reprehenderit quo rerum delectus. Sed iure eos voluptatem exercitationem quam. Inventore ex natus fugiat nesciunt. Sunt consequatur iure neque omnis eveniet sit et. Qui quam eligendi minus accusantium.
Maiores ex repellendus eos laudantium officiis voluptatem. Voluptatem ut deleniti voluptas et placeat accusamus quasi. Aut asperiores hic sunt sed voluptates et. Quam perferendis doloribus quae expedita esse eum.
Maxime dolores saepe facere voluptatibus quis et quia. Quas rerum omnis nemo ipsa. Aut velit rerum ipsa est ut. Ducimus nesciunt autem est.
Ea alias tempora laborum aut. Excepturi tempora earum non quis exercitationem. Earum similique corporis nihil itaque aut. Saepe debitis eum facere quibusdam recusandae.
Totam porro sint eum qui ducimus ab. Aut fugiat doloremque consequuntur autem non quam cumque. Sit quos sunt voluptas. Quia accusamus quo libero in consequatur.
Modi qui dignissimos corrupti qui. Consequuntur laborum hic nostrum omnis eos eius tempore numquam. Et aperiam quia repellendus et quis similique. Voluptatem recusandae sed praesentium numquam blanditiis.