Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

AlphaGo's next move

deepmind.com

Post reply on HN
Print view View original

171–179 of 179 posts

Re: AlphaGo's next move

#171

Post by cocktailpeanuts »

YeGoblynQueenne wrote: post #169
cocktailpeanuts wrote: post #130

Earlier quoted context omitted.

> For example, when learning a new language That's a good point, because I think this "ability to speak multiple languages" will become commoditized too through technology. You already see pieces of technology that enable you to communicate in realtime (although clunky and not accessible enough at the moment) I agree that "realtime" aspect would be the last wall that will stand to distinguish humans from machines, th…

See my post above. We have very good translation dictionaries, that list all the meanings of words in two languages you might ever need. Why isn't that multilingual ability being commoditized? And why is it still not possible for most people to speak many languages as well as their native one?

I am going to assume that this was a question you asked out of true curiosity and not a rhetorical one.

Technology has commoditized a lot of things historically. Just because you don't see something happening, doesn't mean it won't happen in the future.

Also did you read the comment you just replied to? Because that's exactly the answer to your question. Currently the language translation has not been commoditized because we still have the last mile problem. There is always certain threshold technology needs to cross before it becomes widely spread. Currently using those dictionaries and google translate are still too slow and you can't interact in realtime that way. Once you can directly hook circuits into your brain (or something with a similar effect), this this will change.

Re: AlphaGo's next move

#172

Post by lloydde »

cocktailpeanuts wrote: post #166
yladiz wrote: post #157

Earlier quoted context omitted.

It is not a "first world problem" to want to search for accurate and unbiased information.

I guess you didn't get the metaphor "first world problem". Go back a couple of decades and you'll find central authorities and organizations with agenda to manipulate the public could easily do so because the rest 99.99% didn't have access to information. People who wanted truth had absolutely 0 way of attaining the truth back then. Now you're complaining about how it's "hard" to search for accurate and unbiased info…

Context matters and the context of my assertion was "Nowadays information is just a Google search away, so knowing a lot doesn't really mean as much as it used to."

Knowing a lot still means a lot ;)

I do generally agree with you though I'm hesitant to use the term "truth". There is a part where it is first world solution: a lot more information is now also available to citizens in the first world.

But you are wrong that it is not that hard or that it is an issue of googling a bit more. In many areas Google's improvements in these decades has been incremental. And if your search term matches hot commerce terms then likely no amount of googling will get you where you want to go.

Re: AlphaGo's next move

#173

Post by cocktailpeanuts »

lloydde wrote: post #172
cocktailpeanuts wrote: post #166

Earlier quoted context omitted.

I guess you didn't get the metaphor "first world problem". Go back a couple of decades and you'll find central authorities and organizations with agenda to manipulate the public could easily do so because the rest 99.99% didn't have access to information. People who wanted truth had absolutely 0 way of attaining the truth back then. Now you're complaining about how it's "hard" to search for accurate and unbiased info…

Context matters and the context of my assertion was "Nowadays information is just a Google search away, so knowing a lot doesn't really mean as much as it used to." Knowing a lot still means a lot ;) I do generally agree with you though I'm hesitant to use the term "truth". There is a part where it is first world solution: a lot more information is now also available to citizens in the first world. But you are wrong…

Yeah i take back my words about how it's not that hard. I think I'm relatively good at google myself, sometimes I wonder how people who're not good at searching can live their lives. For example, most of the times "being a good programmer" means being able to come up with the right search query to search for your problem on stackoverflow. Most of the times you don't even know what to search for.

That said, I've been comparing the current status with how it used to be before the Internet, so I guess we're pretty much on the same page.

Re: AlphaGo's next move

#174

Post by nickpsecurity »

computerex wrote: post #154
nickpsecurity wrote: post #131

Earlier quoted context omitted.

"The goal is to do things like self driving cars. " Replacing a human driver takes an AGI at least for the exceptional or new situations. It's why we're including it as a counter instead of supporting point.

Nope, not at all. Like I said, we have had autonomous vehicles operating in normal traffic for decades now. Autonomous vehicles do not require AGI, far from it. Waymo, Uber, Tesla and more are all competing to bring autonomous vehicles to the mass consumer market and indeed most estimates claim that we'll have autonomous trucks by 2027: https://arxiv.org/abs/1705.08807 I hate how everyone thinks they know enough to t…

Define normal traffic. Last time I read on it, the autonomous vehicles couldn't handle rain and some other weather conditions without a human operator. Then abnormal situations can still require a human operator. The AI has to correctly detect its own inadequacy and then hand ig over to human who must react in time starting out distracted by whatever they are doing. Safe, automated handling of these situations might take a broader AI thaf understands context.

Note: I'm all for getting data that shows the narrow AI's have been corrected to handle what I described.

Re: AlphaGo's next move

#175

Post by lawrenceyan »

nathan_f77 wrote: post #165
doktrin wrote: post #107

Earlier quoted context omitted.

> I'd love to be proven wrong but that hypothetical general AI is at least decades away in my humble opinion. I hope so. General AI is the source of dreams and nightmares, and I think it will take us decades just to adequately prepare for it from a safety & management perspective. > Now would be a great time to join and make contributions. Sincere question : short of embarking on a PhD, is AI research something the a…

I've skimmed some papers, and my impression was that you would need to have at least a bachelors degree in mathematics. And by "skimmed some papers", I mean they may as well have been written in a foreign language. This also helped me realize that I'm not super excited about getting into machine learning research, because it's just way over my head. I can play around with TensorFlow, and I enjoy writing all of the co…

To be honest, if you're smart enough to be doing software engineering, then you're smart enough to learn linear algebra, statistics, and maybe a little calculus but not even really. That's all the underlying mathematical foundation you need.

Re: AlphaGo's next move

#176

Post by YeGoblynQueenne »

cocktailpeanuts wrote: post #171
YeGoblynQueenne wrote: post #169

Earlier quoted context omitted.

See my post above. We have very good translation dictionaries, that list all the meanings of words in two languages you might ever need. Why isn't that multilingual ability being commoditized? And why is it still not possible for most people to speak many languages as well as their native one?

I am going to assume that this was a question you asked out of true curiosity and not a rhetorical one. Technology has commoditized a lot of things historically. Just because you don't see something happening, doesn't mean it won't happen in the future. Also did you read the comment you just replied to? Because that's exactly the answer to your question. Currently the language translation has not been commoditized be…

>> Also did you read the comment you just replied to?

Sure. Did you read the posting guidelines?

Re: AlphaGo's next move

#177

Post by yters »

reckoner2 wrote: post #146
yters wrote: post #123

The problem is an AI that is good at Go is not at all transferable to any other game. However, a human prodigy can apply their genius to many domains.

> However, a human prodigy can apply their genius to many domains I'm not sure how true this is. It's pretty rare to find someone who is a genius in more than one domain. Einstein was famously offered the presidency of Israel. Sure, he could probably do well in other sciences, but he was smart enough to know he could not apply his genius to unrelated domains.

Of course there is domain knowledge that needs to be learned, along with non intelligence related characteristics, such as personality. But general intelligence appears to be widely applicable.

Re: AlphaGo's next move

#178

Post by johnsmith21006 »

gwern wrote: post #115
rojobuffalo wrote: post #114

Earlier quoted context omitted.

DeepMind's mission is to build AGI. I think it's probably good if they have a buffered lead on all other efforts. That concern probably weighs on decisions about releasing code. The rationale for why a buffer would be good is described by Demis Hassabis here: https://youtu.be/h0962biiZa4?t=11m24s ...the main points are: there may be safety considerations along the way that are costly. More "capitalistic" organization…

That would require some radically inconsistent thinking on their part. DM does occasionally release source code and trained models for other things, and the arms race logic ( https://www.fhi.ox.ac.uk/wp-content/uploads/Racing-to-the-pr... ) would even more strongly argue for not releasing anything, even research (they're privately owned, they don't have to publish squat), and especially not running stunts like the Al…

I am constantly amazed what Google shares. They are a company with shareholders to be fair.

Re: AlphaGo's next move

#179

Post by Cyphase »

dmoy wrote: post #155
Cyphase wrote: post #134

Earlier quoted context omitted.

http://www.slate.com/articles/sports/sports_nut/2012/06/long... "There's no denying it—our kind started substituting brains for brawn long ago, and it shows: We can't begin to compete with animals when it comes to the raw ingredients of athletic prowess. Yet being the absurdly self-enthralled species we are, we crowd into arenas and stadiums to marvel at our pathetic physical abilities as if they were something speci…

RE long distance running: Actually sled dogs way outpace humans at long distance running. Like they can do a marathon a day for a week straight, while pulling a sled. Humans are second place tho I think.

Here's a relevant portion of the linked article, particularly the last bit about Huskies:

> But what most sets us apart as runners is that we’re really cool—we naked apes are champion sweaters and can dissipate body heat faster than any other large mammal. Our main rivals for the endurance-running crown fall into two groups: migratory ungulates, such as horses and wildebeest, and social carnivores, such as dogs and hyenas. They can easily out-sprint us by galloping. But none can gallop very far without overheating—they largely rely on panting to keep cool, and they can't pant when galloping, for panting involves taking very rapid, shallow breaths that would interfere with respiration when running. Dogs can gallop for only about 10 to 15 minutes before reverting to a trot, and so their distance-running speed tops out at about 3.8 meters per second. Horses' average distance-running speed is 5.8 meters per second—a canter. Wildebeests’ is 5.1 meters per second.

> Elite human runners, however, can sustain speeds up to 6.5 meters per second. Even run-of-the-mill joggers typically do between 3.2 and 4.2 meters per second, which means they can outrun dogs at distances greater than two kilometers.

> Our "sustainable distance" is also hard to beat. African hunting dogs typically travel an average of 10 kilometers a day. Wolves and hyenas tend to go about 14 and 19 kilometers, respectively. In repeated distance runs, horses can cover about 20 kilometers a day. Vast throngs of human runners, by comparison, routinely run 42.2-kilometer marathons in just a few hours, and each year tens of thousands of people complete ultra-marathons of 100 kilometers and longer. (A few animals can match that under special circumstances. Huskies can trot up to 100 kilometers in Arctic conditions when forced to by people. But in warmer climes—no way.)

Post reply on HN