Slashdot Mirror
Eric Schmidt and Bob Work: Our AI 'Sputnik Moment' Is Now (breakingdefense.com)
schwit1 shares a report from Breaking Defense: China's just announced an AI strategy designed to assure it will be dominant in the host of technologies by 2030. "If you believe this is important, as I believe, then we need to get our act together as a country," [Alphabet Exec Chairman Eric] Schmidt said this morning. In a Q and A session at the event organized by the Center for a New American Security, Schmidt said he thought the U.S. will maintain its lead over the People's Republic of China for the next five years, but he expects China to catch up about then and pass us "extremely quickly." How important does China think AI can be? Former Deputy Defense Secretary Bob Work told Breaking Defense the Chinese estimate they can boost economic growth with AI by 26 percent by 2030. "It's quite stunning," Work said. And, of course, the PRC's government has published a national strategy and released it to the world. What's the best response by the United States, I asked Work after Schmidt spoke. The federal government needs to answer this question at its highest levels, as happened after the Soviet Union stunned the world and launched the first satellite, Sputnik, Work said.
I was born 2 days after sputnik
and we got to the moon in '69
and haven't gone back there since '72
sometimes progress just stalls
Imagine if Sputnik never left the ground, and there was a huge campaign by businesses and media to convince everyone that Sputnik was flying around in orbit. That's AI.
Make a sales pitch and sell stock. Don't plunge future generations further in debt to fund your "right around the corner" artificial intelligence breakthroughs.
Sounds like them. Hurry up and make a stink else everyone will see all we are doing is throwing our shit around!
That Eric Schmidt and people like him who promote AI are not disinterested parties? They all stand to make lots of money if they're right and if they're not there's little or no downside. We've seen this movie before: AI Winter. Now they're arguing for more public funding of AI research so that if and when it does bear fruit they can snatch up the results, patent them and then sell them back to us at an obscene profit. Socializing your costs and privatizing your profits, it's the American way. As for the Chinese, they'll throw vast sums of money behind whatever we Americans think is hot at the moment. The last Chinese invention that was actually original might have been paper or maybe gunpowder. In other words, what have they done for us lately besides steal our ideas and tech?
Maybe it is time that they Americans let go of their superiority complex.
"The federal government needs to answer this question at its highest levels"
Schmidt is lobbying for free tax money. Their multi-billion advertising company apparently cannot pay the bill, so please, ordinary people, pay it for them, and then the profits will be theirs.
People have been saying that we're no closer to general AI now than we were 50 years ago, while others say that progress isn't linear and we might just stumble upon it one day accidentally while trying to do something else. I'm of the opinion that the specialist systems we're creating today are indirectly leading to the creation of general purpose AI. Eventually, someone will look at the dozens of sensory pattern-matching, deep learning and analysis specialist systems, wonder 'what would happen if I stitched all of these together into one system?' and accidentally create something capable of things that no one system does by itself, due to emergent behavior. This proto general AI will be buggy as hell, and you'll kind of have to squint in order to see the intelligence... but it'll resemble something we haven't seen before. After lots of finagling, it might even be stable (and mentally stable) enough to impress lay persons.
What I'm wondering is: how much will people (perhaps retroactively) see the moral status of these early alphas -- as beings deserving rights -- to be like gametes (or even early fetuses) of humans? Even if it gets into infinite loops and is less than sound in its output, would people demand it have rights just like the more advanced versions would be likely to have advocated for? How many running instances would it then be entitled to?
Corruption is convincing someone that the selfless ideal is the same as their selfish ideal.
This should be great. Then China will have technology secrets worth stealing and hackers from elsewhere in the world can pirate them. Or other countries can require the Chinese to manufacture their AI products locally, in cooperation with local companies, who can rip off trade secrets. China will be pulling their weight in technological advancement and balancing up the flow of stolen trade secrets. Where's the problem?
Let's bring coal back!
Sure, an AI arms race will compel both us and the Chinese to develop AI as rapidly as possible and turn it loose without first considering the consequences of the singularity and without implementing adequate controls first. What could possibly go wrong?
After all, the entire internet was designed around the idea of transparency and the free exchange of information. It wasn't that the designers of TCP/IP meant to force everyone to give up their privacy. It simply didn't occur to them that some people might not want the email spam, mass surveillance of web page accesses and clicks, the entire advertising industry, and the world-wide hacking that never stops that their communist utopia of a TCP/IP suite brought us all.
After all, the FAA is mandating an automated communications technology called ADS-B that is supposed to replace ATC radar and provide localized separation between pilots that cannot see each other in the air (like TCAS). It wasn't that the designers of ADS-B meant to leave their communications protocol completely unsecured so a script kiddie, a mischievous teenager, or the very rare actual terrorist can sit on the ground sending signals to aircraft in the air that makes them think they're about to crash into each other, forcing evasive maneuvers that can lead to real crashes. No, the government just claims that it has everything in hand and it's all under control. Don't worry, Mr. President, we can tell the signals from the ground apart from signals from the air. We've got this.
So, as the government runs at breakneck speed towards the cliff above a skynet-based armageddon just so they can beat the Chinese and commit to the final leap before anyone else, we should think about how we will react to a world filled with computers that don't obey their owners and masters. Just like the internet will give the singularity access to every computer on Earth, perhaps isolation will allow us to regain control.
After all, we are the ones who have to come in and clean up the messes the government makes. How many times have we bailed out our economy after the government failed to enforce the law? 2008? Bernie Madoff? Michael Milken? General Motors? Takata? BP and the Deepwater Horizon? The price of gas in the 70's? Microsoft? Exxon Valdez? The price of gas in the Naughts? Etc? Etc? Etc?
Perhaps the government constantly fails to do the right thing because they're too busy implementing their fascist police state in our nation's airports, court houses, and anywhere else they can because stripping us of our freedom and violating our Constitutional rights is more important to them.
This disaster will be no different than the others. The singularity will have its day and an unknown number of people will die. Maybe it will only be 20 million or so like in "Fast Forward." Maybe we'll get lucky and have almost no fatalities but we'll lose our internet and our electricity like in "Transcendence." Whatever the death toll, when it's all over you and I will pick up the pieces, clean up the mess, and pay for everything.
Maybe we should stop drooling over self driving cars now -- we've already had fatalities. Maybe Airbus should put the pilots in control of their airplanes -- we've already had fatalities there, too. Maybe we should stop AI research before the singularity happens -- before we have fatalities.
China seeing AlphaGo beat their top player was THEIR Sputnik moment and that motivated them to make the huge investment in AI. You can't have a Sputnik moment without a competitor pulling ahead. But Schmidt is right on one point, after it does look like they are ahead, it will be very hard to catch up.
I would like to see fewer announcements about AI based on other peoples' wild claims about AI's future, and more based on actual achievements. I know it can be difficult to quantify an "achievement" in this field, but, wow, the sheer volume of woo, fairy dust and unicorn farts we're getting is incredible . It's like taking the CGI in skin care advertisements seriously, and believing that we already have functioning nanotech.
You have to plan ahead and hedge your bets, as a country as much as a corporation. And don't pretend there are no actual achievements, either.
We've been to the moon loads of times (including the USA), since 1972.
https://en.wikipedia.org/wiki/List_of_missions_to_the_Moon
Progress is not having to send a man to do a robots job.
And soon we won't even need to communicate to tell the robot we send what to do. We'll *teach* its AI what to do and it will make the decisions as it goes and report back.
Welcome to progress. Progress by its nature isn't doing the same thing over and over again.
I guess the 26% figure is just some mean estimation from some modelling based on the typical benefits from ML-based resource optimization, projected development rate, etc. I guess there was no appropriate room to share the underlying assumptions, methodology, the margin of error, etc, because it was not a scientific conference.
I think it also makes sense for all countries to invest in this technology, just because it pays for itself. The US needs to pay attention and try to match or exceed China's commitment only as long as they want to continue seeing themselves as a major world power.
Actually 26% was offered by the AI as strictly an opinion of what it thought it could do.
So the last world war will be Chinese AI God Emperor vs Russian AI Military Dictator vs USA AI IP Monopolist. European AI Bureaucrat will be still compiling the european law so will not participate
And don't pretend there are no actual achievements, either.
And those would be?
A Sputnik moment implies that the US is in some kind of race. In fact, they are not competing in a race. They are on the field and walking around the track, but they are not aware that there are other runners on the starting line preparing to begin sprinting.
I do not believe that the US is capable in this moment to have any sort of "moon shot" program in any area. Back in the 50's and 60's, most Americans trusted their government and they trusted American business to "do the right thing". Those days are over. These days a deep mistrust exists between the population and their governmental and corporate masters, and rightfully so.
My personal opinion, any perhaps you disagree, is that in the current climate, it is simply unthinkable to pour treasure into massive national "science" type programs.
American only has enough money to support the war machine... not to increase the knowledge base of the betterment of all.
Hell, there are loads of Americans who actively oppose the government spending any money on research or science in general. I suppose that is not a surprise given the rise of people in the US who do not "believe" in global warming and the damn scientists are just after those fat research grants.
How much money to those morons think the average scientist makes? Because...it is not a lot.
So you're saying we should wait until China has passed us with actual achievements, like the Russians did with the Sputnik, before starting a program to catch up ?
Does no one remember history?
https://en.wikipedia.org/wiki/Fifth_generation_computer
Japanese claim they are taking over AI. US and EU panic and talk about DOOOOOM! Absolutely nothing happens.
And NOW the end game of all this AI hype is clear: grab some Federal Money. I knew after hearing all the recent "AI" hype there was money to made somewhere. There is no such thing as AI, but that doesn't stop people from grabbing taxpayers money.
A machine played "GO" and won! And my Phone can TALK to me. Blubber blubber blubber
A machine played "GO" and won!
It did more than winning. It took 40 days of self-learning, starting from zero, using just the rules of the game, to eclipse thousands of years of human study.
China: we will pioneer research in this field for the benefit of all
America: the rest of the world is our enemy, and we must effort to win this war
It didn't learn anything. It ran algorithms based on a strict ruleset. Computers are very good at that. That is all they are good at. A computer will win any game you program it to play. Chess, go, checkers, etc.
Just add it to munition list and let the war industrial complex create your sweet Skynet.
Computers are very good at that
No, before AlphaGo came along, they all sucked at it.
Just checked tertiary education attainment, still a very long way to go but steady progress.
BUT when you check (upper) secondary school percentage, you are below 14% of the population (Most occidental countries are at 90%). This is the limiting factor. Only a fraction of those will pursue in tertiary education, and secondary level labor is essential for the tertiary educated people being able to concentrate on their job properly. Even with the population size taken into account this is still for too low.
AND this is without accounting for the imbalance in male to female ratio which should have disastrous psychological/social effects on the population at large.
AND that with higher eduction comes comes scientific thinking which is very dangerous for the regime...
Quick! Everyone pour money into "leading" AI research monopolies **cough** Google/Alphabet **cough** so they can make super-Human AI to censor our opinions better than the hacks they currently have to PAY to do the job poorly. This is seriously life-or-death stuff, we NEED better AI, free speech is just horrible and so it having to pay people things to do things.
I have yet to meet any Americans who oppose spending money on science or research, ever.
I have met many that oppose spending money on political problems that people claim is "science" where the predetermined outcome is increasing their taxes greatly. When those people are questioned on those facts, terms like "denier" comes up and the discussion is shut down. When peer reviewed research is suggested, again the "denier" label comes out and it is prevented.
You are confusing political movements designed to oppress the middle class with "science". Its understandable because you probably agree with oppressing the middle class based on the fact that they don't vote the way you think they should and they need to be punished until they learn better. We have entered a climate where punishing people based on their political views is acceptable, and you don't want to fess up to it and shut down the discussion with terms like "denier" and "anti-science".
Like I said, I have not met a single American not wanting to see money going into ACTUAL scientific research. But you confuse political positions with predetermined outcomes with the term "science", and that is your problem.
For this goal, it is too late to start obscure government programs to support education. I mean that helps, but is not enough. By now, what would really help is a government-sponsored Manhattan project style lab with top scientists stolen from everywhere, with a virtually unlimited budget, and a firm goal to build the legendary self-improving thinking machine, while keeping it under control.
Schmidt probably hopes that such a project would be based on DeepMind, because they have some sort of a head start, and a great team. That may be the case, but if the government did it with full force, Alphabet would not keep control of the workgroup. At best, they might work out some deal for a share in the intellectual property.
Deep Blue would disagree with you.
Instead of worrying about some vague future threat (or pretending to, in order to get cheaper future labor), how about the folks at Google fix the rogue AI at YouTube which are behind the Ad-pocalypse first? They're killing the golden goose at a rapid rate. Their search engine isn't doing as well as it used to, either.... I've heard that Bing is now the best way to google something.
Comment removed based on user account deletion
Anything the guy says is guaranteed to be crap even the most credulous "smart creative" couldn't believe. For those who haven't read his shitty book, "smart creative" is a term that is repeated about every other page, and is supposed to describe the kind of worker Google claims to want to hire. I guess "smart creative" sounds better to investors than "visa slave".
And before some Linux hobbyiest tries to give me shit for using a single "=", yes, I am assigning the value Full Of Shit to Schmidt not doing a comparison, you stupid fuck.
...if you want to chuckle about AI never happening. For Mr. Schmidt's warning message, Sputnik works better.
By the way, sustainable fusion could be one of the first problems that we might want to throw at an advanced AI. A lot of it is concerned with plasma field configuration, containment control, and materials design. All problems close to the automation, simulation and optimization techniques where computers already play a big role.
- reliable text and speech recognition
- auto tagging of photos and videos
- AlphaGo victory, then a stronger version learning with no human game examples
- DeepMind AI mastering many video games with raw screen input and score only
- self-driving cars
- chatbots moving from toy projects into sales and marketing
- generative models making art and music which start getting enjoyable
Really, you have to be a committed denialist to not notice these, or brush them off as "not AI".
Maybe if it doubles as a pork barrel project distributed across multiple states... Actually, as computer work is easy to distribute remotely, this just might work.
Here are some achievements from my memory, mind you there might be some small errors in the list. I have to say that I'm starting to see a lot of AI applications nowadays compared to e.g. 5 years ago.
- We have a chatbot that fooled humans (by pretending to be a foreign kid)
- We have AI that learned by itself how to beat best humans in Go
- We have self driving cars
- We have open source library for AI that needs a few hundred images for training in order to work pretty well, takes about an hour to take into use.
- We are starting to see AI being used in farms for picking fruits and sorting them, several cases.
- We have AI that can make recipes
- We have AI that can do diagnoses better than human doctors or small areas.
- We have AI that has figured out 4 different indicators for specific decease when textbooks only knew about 3 indicators.
- We have AI pretending to being student assist and answering student questions. Some students even nominated the AI as best student assist.
- We have AI that scored better than 2/3 of humans in Tokio University entrance exam (the AI did the same paper test as other students, reading and writing to the paper).
- We have AI that can draw realistic images of birds based on text description of what the bird looks like.
- We have AI that came pretty close in winning essay writing competition.
- We have AI that does some trivial journalist work by reporting sport and election news.
- We have AI that has learned how to walk
- Image recognition is now better than humans.
If Google wants to make America great again, why don't they stop fucking around with advertising technology and focus on things that will improve the world? I mean, training AI to better monetize their users is so last decade, and you can only squeeze more money out of people if they have more money, and those people still need to get money somewhere.
Why?
1. More people.
2. They value education. Most kids in the US want to have fun and find all them fereigners a pain because they push up the curve. I know a girl in the math program at UT who is now in her 2nd year and has yet to complete a math class. She might get thru one this semester. She dropped the first 2 she took.
3. They want it.
4. US is spending most of its resources on sports and political fighting. I am not even surprised anymore when I hear about some new left/right wing "think tank". That is pretty much where everything goes these days. Trying to convince others we should spend more/less on ??? or cut/raise taxes.
The AI book that everyone should get is available for pre-order (no release date yet). "Artificial Intelligence For Dummies" by John Paul Mueller and Luca Massaron.
The concern here is not about exclusivity, but the massive economic and military advantage in case of success, which would be difficult to match in good time, without a timely matching research commitment.
Schmidt is old enough, as I am, to remember the Japanese fifth generation computer initiative.
Japan at this time was not just an Asian tiger, but an Asian tiger of terrifying, almost mythical dimension. Yes children, America was once so terrified of Japan, we practically threw our lunch money at their cozy, indoor slippers.
This was 1982. In my view, the world started to get serious about SMP around 2005 (as SMP rounded the bend from luxury to necessity due to bumping up against the GHz wall). Does anyone else now recall what Japan brought to this party? I certainly don't.
Lost Decade (Japan)
Has Japan been heard from since? Oh, right, the Cell chip.
There's just a bit of a difference between a) launching a napkin plan to launch a Sputnik, b) launching a Sputnik, or c) launching a Sputnik and then following through with decades of incremental refinement that no other country can match.
How about we wait for B before beginning to worry about C?
The thing is, strong advances in AI can bring such a massive economic and military advantage that it would be hard to catch up after your (B) moment. So it might make sense to closely match their research commitment. Also, this is not something like the A-Bomb, where profitable spin-offs were possible but strictly side effects. Pretty much all AI advances have numerous immediate obvious applications, and basically pay for themselves pretty fast. So there is little reason to resist this trend, just some friction about balancing the funding. A few dramatic appeals like this one from Mr Schmidt may actually help.
Exactly. What if China's AI program fails, and we've developed all this great technology for nothing ?
Smart AI is when it extracts money from people with money.
Junk ads that fail to make me reach for my pocket - have failed.
AI is a failure because they know I am tight as - but bite on loss leader promos.
I can't spend my money - because Google can't or won't introduce me low margin goods, nor tell paid sponsors they can send a different price to known cheapskates. Nor react when I buy elsewhere, or note seasonal predictability.
Here is the secret. It is not AI but SQL data cube shit, the same as Walmart has been using forever. Only the Democrats/Obama used AI to target voters by classes, even better than Amazon.
The second problem for AI is stagnant or declining real disposable income. Forget AI, tailored marketing can be done with rules.AI comes AFTER the easy basics have been done.
... actually still believe that information can be "owned".
Believe. Because reality disagrees, and you keep the mindset anyway. Like a religion.
And all, because you can't produce a legitimate business model, not based on stealing and robbing (aka profit, imaginary property, anything that banks and traders do, etc) like everybody else since the dawn of humanity!
OH SNAP.
Someone just got told!
No. Even Deep Blue sucked at Go.
There won't be any 'sputnik moment' because the current approach to what they're inaccurately referring to as 'artificial intelligence' is a dead end, it will never yield anything even so smart as your dog or cat, it will always fall short, because it will never be self-aware or capable of true cognition. The correct approach to AI will only be possible once we understand how a biological brain is capable of those things, and we're nowhere near beginning to figure that out, no matter what the undereducated fanbois on the Internet say about it.
Oh and by the way: The real hazard that so-called 'AI' poses is not it taking over, but people expecting way too much of it, and things getting mucked up when it fails to do as advertised.
The reason why you see so much of this nonsense is because it seems like 99% of everyone thinks that this thing they keep calling 'AI' is like what they see in movies and TV shows (conscious, self-aware, actually 'thinks', has a personality, etc) when in reality the family dog or cat is smarter than even the best of what they inaccurately call 'AI'. The current approach is a dead end technology; it won't yield a true mind, only a pale imitation that falls way short of the mark. We won't have real AI until we understand how a biological brain is capable of doing the things is does, and we're nowhere near that and won't be for quite some time yet.
Planes flew well before we understood the exact air dynamics of bird and insect wings.
Also, an AI does not have to do all the same things as humans, or in exactly the same ways, to be useful. In fact, one of the things that will make it useful, is the ease of integrating with our current non-human computers, networks, sensors and actuators.
Google, which laughably says donâ(TM)t be evil whilst being a tremendous force for global evil, teaming up with the Pentagon which wrecks more havoc on the globe than every other military on Earth combined. Clearly this is a recipe for disaster.
Stop enabling Google. Ditch Gmail and all Google services. Use DuckDuckGo for search. Bankrupt Schmidt.
Wake me up when AlphaGo or Deep Blue create something practical on their own that works, with no input from humans.
Then and only then we can start a discussion on AI.
The current approach is a dead end technology
Well said. I, for one, don't believe that General AI will come from the field of Computer Sciences. I'm betting that it will be born in a physics lab, with plenty of neuroscientists/biologists around.
So you're saying we should wait until China has passed us with actual achievements, like the Russians did with the Sputnik, before starting a program to catch up ?
Why does AI have to be American?
Nobody's imagination is going to be "electrified" like how being in a Cold War and having Sputnik flying overhead with anyone with a shortwave radio being able to hear it did. Honestly Schmidt is an idiot if he thinks the analogy is remotely similar.
Wow, you Chinese people have gotten *significantly* better at your English since we first started seeing you on Slashdot many years ago. Not so good at your propaganda and ies... but *significantly* better at your English.