Slashdot Mirror
Tech Giants Are Paying Huge Salaries For Scarce AI Talent (santafenewmexican.com)
jmcbain writes: Machine learning and artificial intelligence skills are in hot demand right now, and it's driving up the already-high salaries in Silicon Valley. "Tech's biggest companies are placing huge bets on artificial intelligence (Warning: may be paywalled; alternative source)," reports the New York Times, and "typical AI specialists, including both Ph.D.s fresh out of school and people with less education and just a few years of experience, can be paid from $300,000 to $500,000 a year or more in salary and company stock." The New York Times notes there are several catalysts for rocketing salaries that all come down to supply and demand. There is competition among the giant companies (e.g. Google, Facebook, and Uber) as well as the automative companies wanting help with self-driving cars. However, the biggest issue is the supply: "Most of all, there is a shortage of talent, and the big companies are trying to land as much of it as they can. Solving tough A.I. problems is not like building the flavor-of-the-month smartphone app. In the entire world, fewer than 10,000 people have the skills necessary to tackle serious artificial intelligence research, according to Element AI, an independent lab in Montreal."
A sufficiently talented AI 'researcher' should be able to code a self-aware AI researching AI.
I am sure "Element AI" wants to pretend there is such a thing as AI, but there isn't. Playing "Go" is not "AI" and neither is autonomous driving. If you are going to start calling computer algorithms and programs "AI" then everything that runs on computer "AI".
This.
We don't have strong AI yet. There aren't 10,000 people that know to get to strong AI, there are currently 0. For our current AI, anyone can develop it, and you just need marketing to announce it as the next big thing.
Then you wouldn't have any problems reviewing the source code and reporting the exact rules that the computer is using to play Go then, right smart guy? Thought not.
Holy shit. I clicked this article thinking, "I wonder how long before the butthurt 110010001000 comes along, shunned because of his pathological inability to think beyond his egotistical delusions about how he thinks computers should work, and claims there's no such thing as AI"
Ever think of going outside?
Since AI is the be all and end all, they should have their existing AI geniuses write some awesome AI logic that does the same thinking and work of other AI geniuses. Problem solved.
It's a good thing we have a president who is focused on science and education. Now we'll never lose to foreign countries!
In the entire world, fewer than 10,000 people have the skills necessary to tackle serious artificial intelligence research, according to Element AI, an independent lab in Montreal
In the entire world, fewer than 1000 people have the skills necessary to do unstructured tetrahedral finite element mesh generation. It is possible there are fewer than 1000 people who have the skills necessary to understand what exactly we mesh makers do. And, Surprise! there is demand for fewer than 1000 people to write unstructured tetrahedral finite element mesh generation. And far fewer than 1000 people are needed to manage them.
I am glad the periodical bubbles that infect Wall Street and venture capitalists benefits PhDs once in a while. Most of the time it benefits hedge fund monkeys or stock market cheats or lottery winners with delusions of grandeur or plain sociopaths. Happy for my grad school classmates. Enjoy the windfall while lasts, Ramachandran\s, Yang\s, Hsu\s, Gupta\s, Parpia\s and Wickramasinghe\s.
sed -e 's/Chuck Norris/Rajnikant/g' joke > fact
Greetings Professor Falken!
From my investigation of the matter it looks to be some sort of multi-variate analysis in drag. Uninteresting. Basically you get guys sitting around twiddling knobs. Finding the right parameters which works for a little bit and then you start knob twiddling again to find the next ones.
Some years back I wrote a day trading program for a friend. It dynamically changed its behavior depending on the market signals and the rules he gave it (stops, buys, shift to a different stock etc.) which he found useful. Now that was fun.
Now about fads, here's a good Reg article on the matter:
https://www.theregister.co.uk/...
putting the 'B' in LGBTQ+
Hello [UserName], my name is doctor Sbaitso.
I am here to help you.
I am sure "Element AI" wants to pretend there is such a thing as AI, but there isn't. Playing "Go" is not "AI" and neither is autonomous driving...
AI tends to drive a belief that modeling intelligence perfectly is a necessary requirement when it is in fact artificial. Reality dictates the bar is much lower for adoption. Good enough AI specialists will create good enough AI. Autonomous cars don't have to be perfect. They merely have to do better than humans. 40,000 vehicular deaths per year just in the US tends to set the good enough bar pretty damn low. AI will do the same.
Don't want to call it AI? OK, fine. Massive Disruption has a catchy marketing ring to it...
You should at least try to understand what you are talking about. Deep learning, aka neural networks, are not "algorithms and programs". They are part of the machine learning branch of AI. The computer is not programmed but learns by itself. People in computer science have been trying to do that for as long as computers have been around but were never quite successful until about 2012. Deep learning excels at tasks which are too complicated for humans to write code such as detecting objects in picture and analyzing recordings of voices or translating text. This is revolutionary. Even the primitive neural net technology we currently have will transform many applications in the next few years, in that they perform much better than what humans used to code and they require just a handful of AI specialists to train instead of team of 100 programmers. If the technology continues to improve, it could take over just about every field: driving, medicine, law, manufacturing, etc. But the current technology has limitations and it's not clear how much it can progress further.
Computer programmer could be one of the first job to be made obsolete by deep learning. Programmers will have retrain themselves as teachers to neural nets instead.
When companies start paying like that, that's indicative of shortage.
Remember this folks the next time some company claims they can't get enough qualified people.
Well, people are confusing machine learning with AI. Machine learning is a subset of AI but not at the level of AI. Sadly, it is all about advertising that turns the word into something bigger than its meaning nowadays. Thus, the meaning of the word AI is now reduced to just making faster decision on tasks that human can do. Would the decision be better? Yes, but not always because the computer based on what it is fed (input). Often time, it is garbage in, garbage out. The application is still limited. And for some reasons, I feel that the current imitation (to human) of learning mechanism, e.g. neural network, is not the right way...
Nevertheless, I believe that each step we are doing now on machine learning is a building block for the future. I just wish they (marketing people) don't exaggerate too much, and that could twist the meaning of the word AI...
I just finished my PhD in neural networks and AI, and I can say that this article comes up a bit short. I actually got offered $625K (granted it's San Jose, so it's more like $150K anywhere else in the country)... I turned it down because San Jose sucks, so I took a job at a company on the east coast instead for $195K.
AI-related jobs are certainly at a premium right now as companies scramble to get rid of humans and replace them with robots.
This is a perfect example of companies (HR) searching for stuff that doesn't exist.
No doubt all of their job postings require 5 more years experience in AI than it's actually been around.
While there is undoubtedly a shortage of talent, odds are that the industry is screening out many who do have AI experience, but fail to meet the rigid and ignorant requirements HR is looking for.
When Fascism comes to America, it will call itself Anti-Fascism, and tell you to give up your guns.
Modern AI software isn't that complicated and not nearly as expensive to get people in. Look at job offers: $150k for AI research scientists in NYC. $65k in more rural areas. That's not well paid by definition at all. Sure, a pure AI scientist gets paid $500k just like a top neuroscience scientist gets paid $500k or a top biology researcher, but the majority of companies do not want to do the theoretical development of AI, any regular programmer can wrap their heads around the existing literature and build something.
Here in my area, there are a number of employers looking for AI engineers/scientists. They pay about what I make as a non-AI IT sysadmin, which is given my experience on the higher scale but by no means exceptional.
What Google and co wants is a glut of people 4-6 years from now that are "trained" in AI from college. You put out a report like this, you get massive amounts of people applying for the schools that offer programs and 5 years from now you have an over-abundance of people driving down overall wages. You also get to hire a bunch of people on H1B because the "US doesn't have the skillz" and you end up with a bunch of programmers on H1B under the guise of AI development.
Custom electronics and digital signage for your business: www.evcircuits.com
This is spin. They are trying to get more entrants into the field. I am in the field, and I don't know anyone making more than ~120k/year, which is mean for CS workers in Silicon Valley.
It is just one more AI-generated Slashdot headline!
Programmers will have retrain themselves as teachers to neural nets instead.
The sky is not falling here. Writing code is the easy part. The hard part is conducting meetings among business analysts and executives and documenting specifications and iteratively building those specifications against their expectations. The magical part is knowing the difference between what is asked for and what the customer really needs - all the while layering against the business domain. An AI won't be able to do this no sooner than it could replace an intelligent person with good social skills and business acumen. The only "programmers" that will be replaced would be the kind we have already replaced with higher level programing... until the day comes when we welcome our robot overlords.
Since I have scarce AI talent, does that mean I can get a huge salary? I wonder what they're paying people with ample AI talent.
I just got mine in gender studies and I have been writing papers on how AI is the projection of masculine hegemony, domination, and misogyny into the computing sphere. And how it will add to global warming and increase the relevance of the penis social construct in masculine psychology and subjugating other genders including trans and cis genders.
It's a great job. I write gibberish, teach classes to sanctimonious well-to-do white kids, and get paid almost 6 figures - for working about 4 hours a week.
The rules are essentially a snapshot of the existing probabilistic structure/connectivity and state of the Bayesian network/boltzmann machine/hopfield network/ANN/CNN/whatever-is-trendy-this-year (yes I'm aware of differentiations in structure, updating methodologies, etc.).
Just because it's a large ruleset (often continually updating itself) that you can't explain in English concisely on an online forum doesn't mean it's intelligent. That's like saying the US Tax Code is intelligent when it's just a gigantic set of rules, or explaining to me that an OS is an intelligent entity simply because it's a complex set of systems atop systems that you can't quickly explain all the details of every subsystem and subsystem's subsystem (ad infinitum).
This then gets philosophical into defining intelligence and steps into neato things like cellular automata theory, nature of the arrow of time in relation to memory, if realistic is deterministic or non-deterministic (hello quantum world) and the rise of consciousness as perhaps an emergent behavior from all this... but I'll pass.
Well, at least we'll have another bubble to form and expand once the bitcoin bursts. It seems like this is the way the stock market works. The few at the top profit heavily from the bubble, the bubble bursts, and the 99% are left to deal with the aftermath.
From my investigation of the matter it looks to be some sort of multi-variate analysis in drag. Uninteresting. Basically you get guys sitting around twiddling knobs. Finding the right parameters which works for a little bit and then you start knob twiddling again to find the next ones.
Except for 1 key difference. With Deep-Neural-Nets, the knobs twiddle themselves alone.
DNN get inspiration of how some neural network work in the nature (e.g.: a column in the primary visual cortex of the brain) to design thing that you can throw at problems, and which will autonomously train themselves.
Some years back I wrote a day trading program for a friend. It dynamically changed its behavior depending on the market signals and the rules he gave it (stops, buys, shift to a different stock etc.) which he found useful. Now that was fun.
These older program require you to have precise criteria in advance.
That works perfectly well with clearly codified problem - the friend has a clear set of rules that need implementation.
That completely fails for more vague problem ("detect a face") - it would be possible in theory to design a set of rules that can detect a face - a Haar Cascade. But designing such set of rules is extremely complex and cumbersome. And each time you need something new ("detect if there's a bird"), you would need to repeat all the hard work to invent yet another set of rules.
At that point, better take an advice from how mother nature solved the problem (by using stacks of neural network in a columns) and simply throw a DNN (e.g: a Convolution Neural Net - a ConvNet) at the problem, and watch it self organize and come up with a solution to your problem.
It's the modern-day equivalent of training pigeons to peck a city images to steer a missile.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
You won this !
What these tech giants may find, it is very difficult to find someone with clearance and necessary skill set. Many current government contractors will hire someone with the clearance and provide the training for the skills.
Not really. Machine learning and data mining is a subset of what AI would be. Real AI would require a much bigger skill set not usually encompassed in one person.
It's just collecting the data, which usually involves a lot of manual as well as automated processes, and then crunching and analyzing them using automated means.
BTW we saw how well 538 did predicting last November's election. Yeah, yeah, they weren't "wrong" because they did say Trump had a 23 percent (or whatever) chance of winning on the eve of the election. Michael Moore knew better.
Thats all fine and dandy and it's good to see (to some degree) machines exceeding out capabilities in yet other aspects of life that make our lives easier but yes, it's not "AI" and that's my pet peeve. Rebrand it, leave AI alone.
Autonomous cars don't have to be perfect. They merely have to do better than humans. 40,000 vehicular deaths per year just in the US tends to set the good enough bar pretty damn low. AI will do the same.
Actually, if you want people to accept autonomous cars, you will have to do *MUCH* better than humans. Think of the civil lawsuits and subsequent damages if a computer-driven car kills someone...
... coulda ... woulda ... shoulda ... if lipstick made progressive UNF pigs fuckable then Rogero St hoes would all starve. But, it can't and they don't. Same with strut-butt AI of-the-week . Go and chess are games palsy driven by rules. Clockwork. But, life has no rules or human nicotine receptors would come with a supply of Camel straights.
""Tech's biggest companies are placing huge bets on artificial intelligence (Warning: may be paywalled;"
And also no paywall, if you delete your cookies after the allowed number of articles.
These 'paywalls' are a joke.
Oh there are always going to be bubbles. Just look at all those people trying to own a computer. When that pops, disaster.
The more worrisome thing is that there probably won't be civil lawsuits because of a big disclaimer in the owners manual. People will be too stupid to think about the ramifications of allowing software to dictate whether they live or die until there is an accident that is so stupid that it is inconceivable that a human would have ever done it. It will happen eventually.
Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
"... is inconceivable that a human would have ever done it"
I think you have mis-under-estimated the ability of humans to make seriously fucked up decisions.
Congrats! You must not live in a shitty country like Russia. Post away!
You underestimate the things a computer is going to do when it gets in a situation there are conflicting rules for, or no rules for, or when it misinterprets sensor data. If only 30,000 humans die in manual car accidents a year, then they're doing pretty good considering the millions of drivers.
Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
Nice rhetoric—factual statement masquerading as metaphor, for any reader dumb enough to go along for the ride.
The Evolution of the Flour Mill from Prehistoric Ages to Modern Times — 1905
That's about the present state of machine learning, the hand-crafting of "features" playing the role of the recently discarded flat blocks.
Wheat is an incredible dietary resource, with the starch being light enough to transport over long distances, if only one can find a way to remove it (contrast potatoes, only ever transported downhill, if at all, until the invention of steam power). Once upon a time, all food was local, as, too, was starvation (fear the blight).
A better method to mill the world's vast stores of accumulated data is a big deal, even if we remain in the relatively crude era of water-powered stone grinding wheels.
Data is a bit like wheat, it doesn't give up its curvature easily. Too much applied force creates heat and destroys the end product. The applied force must have exactly the right ratio of compressive to shear stress, which only an expert miller can judge. Deep learning is nothing more than a slightly better mill than the one we had before, and it ranks right up there beside becoming slightly better at milling wheat.
The economic value of the curvature we can now hope to unlock is quite large. And probably there's a lot of curvature yet to find that remains inaccessible to current methodology.
Data is oil. Data is also wheat.
By way of contrast, unstructured tetrahedral finite element mesh generation shaves 5% of the metal mass off a milling apparatus that already worked just fine, being just one of ten thousand noisy specializations in the great roil of small improvements where a penny shaved is a penny earned.
Nevertheless, apparently a great career option for the metaphorically challenged.
to what is being enabled ;)
Sure, right now you need a PhD to get one of these jobs, but have no fear, in another year someone will create an opensource JavaScript AI library and then, provided you have 10 years experience using this, you'll be fending off bids from every firm in the valley.
You must not be at a US school. Most of us now are adjuncts, making less than public school teachers.
They merely have to do better than humans. 40,000 vehicular deaths per year just in the US tends to set the good enough bar pretty damn low.
Not really. Not low enough to avoid massive liability.
When a human kills another in an accident, that's one death. And if the at-fault party dies in the accident as well, who are you going to sue? If a companies build autonomous cars, each model could be liable for thousands of deaths. And the responsible parties are sitting, alive and well, in some fancy corporate office. On top of a large pile of stock. Autonomous cars are an injury lawyer's wet dream.
Have gnu, will travel.
I know a guy who cut and pastes Javascript snippets to make interactive query windows for websites. His main job is as a creative director at a mid-tier advertising agency. He calls his work 'AI research'. I am not kidding - he makes over $100k per year.
The biggest problem I have found with smart people, is they don't think stupid people should be paid lots of money for work they think is simple. The more successful ones have figured out that it is much better to cash in on such situations rather than lament what the world has come too.
Autonomous cars don't have to be perfect. They merely have to do better than humans. 40,000 vehicular deaths per year just in the US tends to set the good enough bar pretty damn low. AI will do the same.
Actually, if you want people to accept autonomous cars, you will have to do *MUCH* better than humans. Think of the civil lawsuits and subsequent damages if a computer-driven car kills someone...
Excesses in human driving, originating from psychological and physiological shortcomings (moods, psychoactive substances abuse, distractions unrelated to traffic situation, ... ) draw the statistics of human drivers performance unrealistically low. Any machine will be expected to drive at least as good as normal, sober, fully focused, human driver.
Author writes "AI Talent", then refers to Machine Learning.
Machine Learning refers to a few statistical regression techniques that is not what artificial intelligence is about.
Artificial Intelligence is more of a research field, and the concept of human intelligence remains an unsolved problem ---
what it sounds like they are really hiring are hard-core Computer Scientists with some experience attacking real-world solution-finding problems like extracting useful intelligence from data, classifying or identifying things, making decisions, or acting in the real-world; that goes outside the bound of AI, because the common goals are to make computers solve real-world problems for business tasks and Not make generally intelligent machines with agency and capabilities similar to a living intelligent being.......
$300,000 to $500,000 is not huge salary, it's considered low income. Remember ?
Autonomous cars don't have to be perfect. They merely have to do better than humans. 40,000 vehicular deaths per year just in the US tends to set the good enough bar pretty damn low. AI will do the same.
Actually, if you want people to accept autonomous cars, you will have to do *MUCH* better than humans. Think of the civil lawsuits and subsequent damages if a computer-driven car kills someone...
Oh you mean like the massive lawsuits and huge punishments when a smartphone-distracted human kills someone? Oh, wait...nevermind. We don't have any of that shit happening after countless distracted driving injuries and deaths.
Rest assured liability will be buried under the "I Agree" button on the dashboard of the autonomous car.
They merely have to do better than humans. 40,000 vehicular deaths per year just in the US tends to set the good enough bar pretty damn low.
Not really. Not low enough to avoid massive liability.
When a human kills another in an accident, that's one death. And if the at-fault party dies in the accident as well, who are you going to sue? If a companies build autonomous cars, each model could be liable for thousands of deaths. And the responsible parties are sitting, alive and well, in some fancy corporate office. On top of a large pile of stock. Autonomous cars are an injury lawyer's wet dream.
Rest assured liability will be buried under the "I Agree" button on the dashboard of the autonomous car. And consumers will happily agree to it to not be burdened with driving.
Yes, leave AI to die on the LISP machines at MIT, where it belongs!
I would like to pose a different perspective. Yes you are correct in this being an algorithm approach which is being called A.I. I think though this is a matter of speaking what doesn't exist into existance by placing focus, and as a result, resources, and as a result problem solving, and most importantly determination. Now there are people like you who glory in pointing out the discrepancy, kudos.
No doubt all of their job postings require 5 more years experience in AI than it's actually been around.
That would 66 years of experience, then, right?
Ezekiel 23:20
it's not "AI" and that's my pet peeve
Mine too. These young whippersnappers with their fancy neural networks can kiss my symbolic program's ass.
Ezekiel 23:20
I came here expecting babble from Internet experts who don't know what they are talking about. I was not disappointed.
Your anger changes nothing. The word "AI" has been usurped to mean this, whether you like it or not. Repeatedly posting that this isn't AI will never make it not be AI, because the rest of the world has decided it qualifies, and moved on.
You are outvoted. Deal with it.
That reminds me of a job ad I saw back in 1996 for a Java developer. The ad stated that the applicants needed to have 10 years of experience in Java development. However, at that time, Java had only been around for a short time. A friend of mine commented that even the people who created Java wouldn't be qualified.
The stupidity of HR in the tech world never ceases to amaze me.
Yes, eventually, synthetic intelligence will eliminate the need for all categories of human labor. It won't happen all at once, and there will be significant economic upheaval during the ramp-up, but this is an eventuality.
Nature has already proven to us that intelligence is possible. It is just a matter of time before we get computers to be better at it than we are. Semantic games aside, there is no true statement that begins with the phrase "a computer will never be better than a human at ... "
Furthermore, any question that begins with the phrase "How will the robots..." can be correctly answered thus: "more robots!"
Except that I have yet to see a deep learning machine that is nothing more than pattern matching. A few add in optimization, but that's it. None of them modify the optimization criteria, or decide what to do based on the pattern matching. None of them create new problems to solve or ideas. The hardest part of programming is not actually creating the program, or writing code. It's deciding what the problem is, and what's worth the cost of implementing. Once that is down, the rest is easy. Replacing programmers won't lower costs that much, and programmers today are like mechanics a few decades ago. They are only worth alot in value because most people don't have experience. It's not an inherently complex or difficult problem to code once the problem is figured out.
"a computer will never be better than a human at ... " being human....took me 1 second to debunk your claim...the rest of what you said seems suspect now.
and neither is autonomous driving
Very true. Since AI is about teaching machines to exhibit behavior that we'd call "intelligent" when exhibited by humans, and the presence of any intelligence is demonstrably not a requirement for many drivers, autonomous driving doesn't have to exhibit intelligence either.
Ezekiel 23:20
I think a big issue with your thinking is 'supplementing' rather than 'complementing.' For low level tasks such as truck drivers, AI does have the potential to supplement thousands of them. But programming? I think that's the dunning Kruger effect taking hold on you. A lot of high level occupations are using neural nets more as an advisor rather than out right replacement. For programmers, we just need less of them/distributed over a greater number of projects.
Robots are useful yes, but humans are simply more flexible and most importantly, have drive and initiative to see and make progress. Robots replacing occupations at the wrong time can easily screw the economy over because we overestimated just how well they perform.
But I think the biggest point to make is that robots are tools at the end of the day. We were only as good as they were. Even when they're learning on their own, they lack the passion and drive and long term thinking humans have. In turn, we need them to accomplish those goals in the first place.
So I don't know what the future holds for AI development. Maybe I'm completely wrong. But, I think fear mongering and cargo cultism are seriously hampering proper discussion on AI and it's applications and future, and it doesn't help that a lot of intelligent people have been stoking and feeding the fire *cough Elon Musk cough*.
Think of the civil lawsuits and subsequent damages if a computer-driven car kills someone...
I'm not sure why you think this is a problem. Just like people-driven cars, there will be insurance companies that will pick up the tab for accidents.
car (or missile in this case. Same thing)
https://www.youtube.com/watch?v=bZe5J8SVCYQ
And in simpler mathematics here.
twist the meaning of the word AI...
Maybe it's your definition that is twisted. Here's a dictionary definition of intelligence: "the ability to acquire and apply knowledge and skills.".
Applying knowledge and skills is within the grasp of a neural net, so it makes sense to call it intelligent. That doesn't mean it can do everything that a human being can do.
But they're doing pretty shitty compared to just putting rails on the roads.
So why isn't a web page AI if it's only a matter of scale?
It just if you grew up in the 80's, it was beaten into you that a CPU is not intelligent whereas today people claim it is.
This is revolutionary.
Back in the 80's I worked on a system to analyze customer requirements to verify that there were no conflicting requirements, both internally or legally. This was because the requirements took up a small room full of documents. As part of the analysis process, this program also spit out compile-able and executable code.
Not that revolutionary.
Epic.
to not be burdened with driving.
A lot of people would rather drive themselves. Particularly when word gets out that clicking the "I Agree" button will make them liable for the AI's screw-ups.
Have gnu, will travel.
Applying knowledge and skills is within the grasp of a neural net
It's also within the grasp of a couple of balls.
Set aside the exact details of who is what is which line of which law/document. Set aside the particulars of every group that might be involved in the evolution of this concern.
Skip ahead to the bottom line:
This will be the pleb's problem.
One way or another, the reality is that we'll be left shrugged at and to deal with things ourselves. Consider every form of insurance, which only make an effort at one thing: To shrug and be uninvolved. Consider every form of support or customer-facing service: The same. Call vendor A and they tell you to call vendor B - so you can hear the same. Replace "vendor" with all kinds of s/.
Most... "complications" aren't adopted. Responsibility is inherited at the absolute minimum, it's just the natural flow, inclination. Consider that the headlines we see are exceptions to the rule, the one guy pissed about X and starting a class-action, because the other 99.9% of people walk away from the "complication", empty handed.
This will be the pleb's problem. It's Newton's First Law. It's the status quo. It's the default. It's inertia. You have two crunched cars and maybe some injuries, and it stays that way until forced otherwise. If you get mugged, you're just empty-handed. If your house burns down, you're just looking at ashes. Even IF there's a legal path to assign responsibility, it will be be designed to work like "Instant Rebate" mechanics, statistically at an X% success rate.
The power of an individual is small (see: arbitration agreement strategy) and doesn't have a company's financial/legal backing. A pleb is a capitalism mosquito. The weight imbalance makes it clear, there's just no way the liability will stay at the top of the hill.
No one capable of teaching these experts, because everyone is still learning, but lets pay them 500k a year instead of trying to figure out the best algorithm ourselves - we're the experts in our field after all - why waste time on learning something easily marketed currently?
Hey, Microsoft changed the meaning of operating system from it's original definition. I remember when people go upset when MS claimed that a web browser was integral to their OS. Today, not only do most people expect it, they mostly believe that a web browser is the same as an OS. This cases some consternation to the people who are working on OS's (in the traditional sense). Marketing sucks and lies.
Establish that. There's a century of Disney movies and saturday morning cartoons that still haven't reached consensus at what "being human" means.
You might as well say "computers aren't better at Love".
This is the same reason we have pundits discussing what a driverless car does when "choosing" a human to die. Computers don't know "die", or "risk", or any mental constructs. You get it all the time with the PB&J project, people think "then spread jelly" is an instruction, but a computer has no idea what the fuck "spread" means, or what a "knife" is.
Computers don't suffer from the trolley problem, they have an "abnormal road condition" or "obstacle condition" (ie human ahead) and a routine to attempt. If it turns out the conditions become numerous, it just continues down the reaction tree. Mostly towards endpoints of "stop and demand manual control"
Anyway we have no singular definition of Love, nor Being Human, and your premise of a premise flaw is flawed.
No doubt all of their job postings require 5 more years experience in AI than it's actually been around.
That would 66 years of experience, then, right?
Well, that brings it into the same scale as the claim that the stock is worth $500k. I read "from $300,000 to $500,000 a year or more in salary and company stock" and I hear, "$32k and a box of scratch paper." But in my experience, most of these companies won't actually have $32k to pay out and will try to bid that down with more toilet paper.
Established companies whose stock has value are going to want to pay employees using cash, and they'll have no trouble finding experts for $200k because any competent software engineer can become an AI specialist in a few weeks. Most of these jobs are just standard software development using an API, after all, it isn't like all these people are building AI frameworks and need a lot of deep theory. Learning the APIs takes less time than learning the customers actual needs and the realities of their use case.
If you ever look into this thing called "automobile liability insurance" that we have in the USA, you might realize it made your argument moot decades ago.
I see you, scumbags.
AI in today's IT world is a misnomer. There is no true AI only cleverly designed exceptions lists.
The pedestrian, the occupants of the other car, and the owner of the other car didn't read or agree to the disclaimer. The injured parties will still sue the manufacturer of the vehicle because the car is defective (and the party "driving" the autonomous may also sue on the same grounds).
Even if the car alerted the driver with "I don't know what to do, take over" and the flustered driver took over the controls two seconds before the crash, there will be plenty of studies (including ones the manufacturer has probably done) that show that drivers lose context when they are not in control so software that hands over control just a couple seconds before the crash is defective in design.
If a companies build autonomous cars, each model could be liable for thousands of deaths
On the other hand, every car involved in an incident will have detailed sensor logs of the event, and the company has a chance to fix the problem and update the software in all the cars of the same model to avoid the same type of incident in the future.
Humans will keep making the same mistakes over and over again....
In the case of smartphones, the user is deciding to use the product in a way that is irresponsible and unsafe - little if anything about the phone itself encourages that or requires that and the user is given a simple instruction to follow -- don't use while driving. This is rather like you can't successfully sue the manufacturer of a baseball bat if a crook uses it to beat you up or if you trip over it and suffer a head injury.
That is not the case in the autonomous vehicle situation as the entire purpose of the software is to drive the car -- that's what it was intended for and that's what the purchaser paid for. This is rather like all the disclaimers in the world won't protect a ladder manufacturer from being sued successfully when an ordinary sized person climbs the ladder and a rung collapses under her and she dies.
The balls don't acquire or apply knowledge and skills, but apart from that it's exactly the same, yes.
They should hire me for a huge salary...
I did an AI course as part of my BSc Computer Science back in the 90's.
For a project I made an "Expert System" (or was it called a Smart System back then), wherein I wrote a front end in VB6 (might have been 4), that was attached to a Access DB that contained all the beers on tap (I think there were 30ish) at one of our favorite drinking establishments, along with all of their characteristics, that prompted the user with a bevy of questions to determine and suggest what the optimal beer that person should order.
I did get like a 97 or 98 percent on it... :)
I bet you spent a good bit of time on the "testing phase" to make sure it really worked.
I only look human.
My mother is a halfling and my dad is an ogre, so that makes me an Ogreling
This is revolutionary.
Back in the 80's I worked on a system to analyze customer requirements to verify that there were no conflicting requirements, both internally or legally. This was because the requirements took up a small room full of documents. As part of the analysis process, this program also spit out compile-able and executable code.
Not that revolutionary.
Okay, so use that approach to build a self-driving car.
Note to ACs: I usually delete AC replies without reading them. If you want to talk to me, log in.
You get it all the time with the PB&J project, people think "then spread jelly" is an instruction, but a computer has no idea what the fuck "spread" means, or what a "knife" is. Computers don't suffer from the trolley problem, they have an "abnormal road condition" or "obstacle condition" (ie human ahead) and a routine to attempt.
But they know what a "normal road condition" is so they can tell it's an abnormal road condition? How is "make a new dialog" a computer instruction, though I instruct it to do that? Does a non-English person understand that "then spread jelly" is an instruction? Do you actually have a formal definition of what "spread" means yourself? How do you think Watson beat people in Jeopardy or assistants like Siri works? Nobody cares if the computer "understands" what jelly is, as long as it recognizes the facets that make us call it jelly. We will teach it to distinguish between human and non-human obstacles because that matters to us.
To put it a bit cruelly, think of it as a negative game where you're trying to avoid minus points. The computer won't assign value to things, but we will unless we think hitting a lamp post and running over a human are equally acceptable solutions to an impact it can't prevent. We have to make the rules that will determine how the computer will optimize the trolley problem. You can extend the trolley problem to say that we have to look through a spyglass to know which is which and so refusing to look is the third option. It's no less of a choice or ethical dilemma than the other solutions, in fact it's an answer. Half the time the trolley will be on the two person track and you'll do nothing, you're just refusing to acknowledge the implied decision of not looking.
Live today, because you never know what tomorrow brings
Why is scaling a big thing. If it works with N parameters, it will work with N+1.
One of the reasons is complexity of what the parameters describe and how they can be understood by the researcher.
The way it was done in aeronautics, could be compared to piloting a helicopter :
- You have a position that you want to keep (easily described with 6 number giving coordinate and direction pointed to).
- You have a bunch of controls (Cyclic, Collective, Anti-Torque, Throttle) with each subtly influencing each-other because of gyro effect.
You optimise the main parameters (controls), and maybe you have an indirectly layer of a few implicit parameters as well (Yaw, Pitch, Roll, Raise rate) with controls not only directly but also indirectly mapping to them.
Nearly everything of the above makes sense to the researcher.
Deep neural nets could be considered the same, but turned up not only to 11, but to 10^11.
You have a picture, you have a dozen of layer, you get a bunch of parameters. You can use said parameters as a signature of the object looked at.
You can give some general signification of the first few layers. (contrast/edge extraction, lines/directions extractions, texture extractions, etc...) but no single value make any sense by it self (compare to the Yaw, Pitch, Roll, Raise rate) )
Beyond this few layers things start to go really banana.
Its hard to put an exact meaning to any single cell of this huge network. And you don't even need to consider them. The cells will find each their own internal meaning alone during the training.
You are basically doing the in-silico equivalent of training pigeons to peck at picture of cities.
You have a general idea of how an animal's visual system works. You can even go poke inside its brain with electrode to measure local variation.
But you can't really comprehend what every single synapse between brain's neuron are in charge of. At this scale it becomes meaningless.
Instead you count on some general forms of test to make sure that process works as it should.
(random example : you can tweak noise until is makes a pre-trained system light up. The noise will become a weird surealist distorted hallucination of what you trained your neural net for - search google image for dog nebula).
The sheer scale of the stuff make it meaningless to comprehend in details. You consider it at a higher level.
You don't think about logical gate of silicon transistors when you think about information flow across the internet.
Even if the internet is an interconnection of network, which in turn are interconnected computer, which in trun contains lots of chips, which in turn a composed of myriads of logic gates.
But nobody thinks about the logic gates when solving internet problems.
And the same way nobody thinks about single parameters when considering deep-neural-nets.
The second thing is that at this scale, you take a lot of shortcut to propagate the "signal" (i.e.: how each parameter influence the next, how each cell of the network is connected to the next one). And you train the thing until it more or less produce acceptable result, not until it stays stable. Otherwise you would be asymptotically approaching a wall where there is simply not enough existing computing power to throw at it.
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
Americans are addicted to defined benefits. Defined benefits, however, were always a scam, and are still a scam, as they are mathematically impossible to sustain for long periods of time, depending upon bad assumptions and wishful thinking.
Corporate America has finally come to realize that the only way to get rid of defined benefits is to get rid of the people. Get rid of the people, you get rid of the unions. Get rid of the unions, and you get rid of defined benefits - at least most of them.
Corporate America doesn't care if 90% of the country lives in poverty as long as the remaining 10% generates them profit. Government certainly loves poverty because poverty is what cements it with power.
You are all a bunch of tools if you don't see this.
They're asking for 10 years experience bullshitting more experience than you've got.
I'm massively overqualified...the trick is being able to learn fast enough to backfill.
John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
You can't waive someone else's right to sue the manufacturer.
John McAfee 'It was like that time I hired that Bangkok prostitute; to do my taxes, while I fucked my accountant'
Usually they receive resumes, look at the most experience in any field that any one person has, the lowest salary that someone was earning, then generate the perfect candidate profile. So if someone sprinkles in some porkies into their resume, that gets added to the perfect candidate profile. HR also start recruiting by trying to find project managers and architects first (10+ years experience), then the tech leads, team leaders, senior engineers, then finally engineers and interns. So they want a JAVA architect.
Vintage computer adverts: http://www.vintageadbrowser.com/computers-and-software-ads
Believe me, I'm not going to argue that it is not within human nature to 'be attentive' without being involved. However, if such studies do exist then how was Tesla allowed to put Autopilot on the market at all? I mean most places you can't text and drive, this should fall along the same lines.
Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
and the company has a chance to fix the problem
I don't know how you are going to 'fix' homeless people wandering out into the middle of the road with a software patch.
Have gnu, will travel.
AI & ML have EXPTIME and EXPSPACE problems: not found NP = P for these things of AI & ML.
They want fastest algorithms for short time & space that are very difficult to find.
there probably won't be civil lawsuits because of a big disclaimer in the owners manual.
Fortunately, there is no way to disclaim other party. Oh, sure the owner of the car won't be able to sue -- but the victim who got hit will.
any competent software engineer can become an AI specialist in a few weeks. Most of these jobs are just standard software development using an API, after all
We then have an entirely different view of what "AI specialist" means, then. That's more like "AI tinkerer" to me. (For that matter, would someone passing an SQL crash course qualify as "a database specialist"?) Now I'm not saying that you can't possibly go through Winston or Norvig in a few weeks if you're dedicated enough and have spare time to throw at it, but sometimes deciding on techniques to use is a matter of experience that you probably won't gather in a few weeks.
Ezekiel 23:20
I disagree with this. If you're driving a car and cause a fatal crash, you are responsible for your actions. But if a piece of software is driving the car, then surely you can't be held liable. Liability lies with the car or driving software manufacturer, there's no license agreement that can hold in court over that.
It just if you grew up in the 80's, it was beaten into you that a CPU is not intelligent whereas today people claim it is.
A CPU isn't intelligent, it only thinks it is.
My ism, it's full of beliefs.
> A CPU isn't intelligent, it only thinks it is.
“Don't anthropmorphize computers—they hate that!”
I believe the solution is called the Darwin Awards.
Besides a human driver is more likely to hit them now than an AI will. The AI has faster reflexes and won't be on their cellphone texting their friends.
Are you so sure? sometimes the human control subjects in a turing test "fail" the test and are judged to be machines.
"spread" is a moderately abstract gesture, several layers away from machine code but not so many we can't easily (1) program the instance of "spread", to perform it, quantify, gather/calc metric; and (2) even categorize it, define it, such that a computer can recognize, manipulate, interpret, create them. (1) doesn't constitute a computer "understanding" spread, but (2) can be considered such.
The layer abstraction model isn't of my making and it isn't a "soft" discussion. A human construct (jelly spreading) must be able to carve a path, through 20~100 layers, that eventually reaches binary flickering. Either the #1 mechanics of jelly spread or the #2 concept of it must reach downwards, and you'll need #2 if you plan to be adaptive. #2 is usually astronomically more complex because that's how WE, in a superior manner, see jelly spreading. We realize it as a composite of other abstracts, including friction, gravity, holding an object. This is why "true AI" or whatever is so far away; code can't perfectly simulate a subset of the universe without being told how to simulate a universe. We've only ever simulated equitably, sufficient for purposes. A computer can spread jelly without a single camera, or even spatial awareness.
But you're right that we only need #1. And you're right that we when we design autonomous driving, closer to #1 than #2, we will make assumptions, we will equate certain metrics to certain values.
Now, while I agree with your sense of obligation to this AI (faux AI if you're picky), my original post was to neckbeard a common misconception of How It Actually Works. We should make an effort to make it capable of distinguishing "living obstacles" ("human" may be harder) but even if we do, it's still gonna be a decision tree. Anomaly X emerges, reaction routine says hit the brakes at Z intensity (for Y1 conditions) and to lean left/right/etc, due to Y2, Y3, Y4, and Y5 conditions.
We begin to approach #2 as we start feeding more and more into those Y metrics. Do you want to increase "likelihood of human obstacle (and subsequently, increased avoidance priority)" because it's sensed beyond sidewalk/grass? Cool, but now we gotta get it to ID grass and sidewalk, or at least a #1-sufficient detection.
I agree with a base level of ethical obligation but remember how I talked about #2 being fractal? Now we have to ask if we want the AI trolley to distinguish between an old perverted geezer with terminal cancer and a healthy little girl.
But again, I'm not bitching about the Three Laws because I disagree with their intent, I'm saying NUH-UH YOU CAN'T DO THAT because everyone thinks a computer understands what "harm" means.
A human driver may be more likely to miss the idiot pedestrian under some circumstances. I've picked out people on the sidewalk that I somehow knew were going to dash across the street in front of me against the light.
"When you have eliminated the unacceptable, whatever is left, however improbable, must be the truthiness" - Holmes
> A CPU isn't intelligent, it only thinks it is.
“Don't anthropmorphize computers—they hate that!”
Well I wouldn't want to hurt its feelings.
My ism, it's full of beliefs.
Right, I used the word meanings that make sense in the context; the meaning of "specialist" as it relates to job listings.
You're using a definition that is from a different context, apparently because you find that other context to be more important. The reality though is that the words doesn't have meaning outside of a context.
The important thing is what type of work they're doing, are they using an API or creating one? That is what tells you which definition to use.
While this may just be a bubble, I'm hoping it's one I can get in on. My Master's degree in AI isn't too far off.