Jump to content

The Singulairty is nearer than you think...


Mallette

Recommended Posts

I have a slight suspicion we were "64,000 dollar questioned" in something of a reverse way. I certainly don't believe the two human contestants participated, but given the stomping Watson gave them the second night after just breaking even the first I can't help but wonder if there was a "conspiracy" to degrade Watson's performance slightly on the first night. The second night was, indeed, downright boring with the audience largely silent as Watson completely dominated. If it had been the first night ratings would have plummeted for the rest. Why bother? Who continues to watch a football game when it's 40-0 at the end of the first quarter? If there is anything to what I am suggesting then I would go further and suggest Watson was also purposefully degraded for the first third of the last night.

Now, compared to the original quiz show scandals, this would hardly be "cheating" in the sense that a player capable of beating everyone degrading their own performance to give others a chance is considered "noble" as compared to providing answers to a lessor competitor to give them an advantage.


You have a point, there, but two other terms also come to mind, "sandbagging" and "hustling".
Link to comment
Share on other sites

  • Replies 125
  • Created
  • Last Reply

Top Posters In This Topic

In the latest news, there may be a "chatty humanoid tweet-bot" heading to the ISS.

Japan’s space agency JAXA announced this week that it is looking at a plan to send a humanoid robot to the space station in 2013 that could communicate with the ground through Twitter—primarily feeding photos, rather than original ideas—and provide astronauts with “comfort and companionship.”

Full item: http://www.japantoday.com/category/technology/view/japan-may-send-chatty-humanoid-tweet-bot-to-space-station#show_all_comments


Hmm, "comfort and companionship"... The Futurama Lucy Liu-bot episode explained where that could go with the important public service message Don't Date Robots!

Link to comment
Share on other sites

I agree that Watson's main advantage was being able to ring in faster than the human contestants. That being said, it was impressive how he/it could figure out clues that were plays-on-words. I did notice however, that he was weak at associating the clues with the category. In other words, the category might be words that end in "eat," but Watson would answer something totally different. I can't remember specific examples, but you get the gist of it. He seemed completely lost on "themed" categories, and was often below the "guess threshold" for entire categories that relied on a common theme.

I kinda wonder how Watson was fed the clues. Was he given them as Alex finished the question, at the beginning or the middle? A computer can read a text file much quicker than humans can speak or hear a sentence.

Overall, a pretty cool experiment, but in the end, I'm in the group that saw it as more a "parlour trick" than anything else.

Link to comment
Share on other sites

Was he given them as Alex finished the question, at the beginning or the middle? A computer can read a text file much quicker than humans can speak or hear a sentence

If he was provided the text file at any point before the "." in the spoken sentence, it would have been completely bogus. I suspect it was feed as a word by word text synced with the hosts delivery, like video subtext. Fair enough.

Overall, a pretty cool experiment, but in the end, I'm in the group that saw it as more a "parlour trick" than anything else.

You must be very hard to impress. Pretty much in the lunar landing level of parlor trick, though probably more profound. The lunar landing was mere technology. This points to something requiring a lot more consideration. While Watson is only a portion of the puzzle, it's the portion that will enable us to far more rapidly figure out the rest.

Pythagoras, Aristotle, Newton, da Vinci, Einstein, Hawking...all these relied on fuzzy logic. Watson is incapable of insight...but the program is capable of taking a human insight and considering it in all possible ways and feeding back info to the human originator of the insight that can compress a lifetime of work into hours. In the hands of an intellect at the level of the aforementioned guys it can speed human evolution in ways we cannot begin to imagine. Even in the hands of the morons at IBM (some humor there) it will begin paying dividends to us all pretty quickly.

I have already incorporated at a very basic level some of the thinking that went in to the Watson algorithms to advantage in my simulations designs. Really baby stuff compared to Watson, but useful.

Hardly a parlor trick.

Dave

Link to comment
Share on other sites

primarily feeding photos, rather than original ideas—and provide astronauts with “comfort and companionship.”

Robeastiality?

Mechasexuality? Makes me think of the Buggles song I Love You Miss Robot: "You make love like a metronome..."
Link to comment
Share on other sites

There is a lot of general misunderstanding of the acheivement of Watson. It is pointless to deal with it point by point as if one understands it, it is unnecessary and if one doesn't, it makes no difference.

Watson is a milestone. It is actually a huge part of that milestone that the hardware employed represents OLD technology, nothing new. Actually, that hardware is a a constraint. New things are being developed that will make Watson's successors much more facile and useful. Heck, even the lowly Windows is largely constrained by old hardware technology. I've recently tested a PCI RAM based 120gb memory unit as a boot drive that boots Windows in 7 seconds. While still 7 seconds too long, it was 3 times faster than the 21 seconds we were getting with an SSD and 15 times faster than the same machine with a 10k RPM SATA HDD. So, in effect, it has increased the machines routine response time by a factor of 5 and at a price of 280.00 versus spending 5k for a new top speed CPU and RAM that would have, at best, doubled the speed.

Watson is about new directions in machine learning. It's that "one small step for computers, one giant leap for ???" moment. This project finally provided a bit of insight into why our efforts in AI stalled a couple of decades ago and now will begin to accelerate.

It's going to be a bumpy ride, but the impact on all of us will be extraordinary.

Dave

Link to comment
Share on other sites

Translation: I have no idea what is really new here, but I am going to claim that everyone else is too ignorant to understand it, therefore I owe no explanation. Big Smile

Very sad oblique ad hominem, Mark. And what I meant was it had all been mooted before and those who had a position to start with and were not receptive were not going to benefit from repetition. Further, my experience has been that the general public has a rather better comprehension of the significance of the Watson program than many "experts."

You could sit 100 programmers down tomorrow and accomplish any number of similar "narrowly defined tasks" - like playing Jeopardy.

"narrowly defined tasks" Like an artificial heart?

I think credit is due for Watson's specific programming - which means specific to Jeopardy.

Wrong. Jeopardy is about as far from a "narrowly defined task" as one can get. However, the first part of your sentence is very correct. The milestone is the redefinition of machine learning. Programmers remain in awe of Google and Bing. Incredible work in those, and Watson is very much a descendent. Google, Bing, Netflix, Amazon and the like will be the immediate beneficiaries, though it's my belief that medicine will be the one that affects all of us in the near term and in really meaningful ways.

I remain amused, but unimpressed.

As were the birds when they saw the Wright flyer, I am sure.

Dave

Link to comment
Share on other sites

Sorry Dave, you left that door wide open on your own. Big Smile

No sir. You forced that door. My impression is that I have a far greater opinion of humanity than you. Further, I state publically that I am, at best, rather average mentally. I am rather average income. I am rather average.

Feel free to remind me if I ever sound high and mighty, but please avoid substituting oblique ad hominems for a worthy rejoinder.

If you don't like the rejoinder, perhaps you should have thought more before claiming everyone is "too ignorant" to understand the relevance of Watson.

"Everyone" is not too ignorant...and that is the correct word as it is not a value judgment but an adjective describing a state of being uninformed...to understand the significance. AAMOF, my impression is that the majority understand it. Really only a handful who either have some personal tenet that requires poopooing it, or luddite tendencies, or just enough computer knowledge to misunderstand the significance of the code.

Another reason I've not joined in trying to "prove" anything here is that I could care less what you or anyone else things about this achievement. It's sheer joy to me and another's darkness does not dim my light.

If it did nothing but spur the imagination of an America in retreat on science and technology it would be a shining moment.

Dave

Link to comment
Share on other sites

(Dave, please don't add more of your haughty aggrandizing complaints that everyone who disagrees must be a Luddite.)

Within the bounds of good taste and gentlemanly discourse, I reserve the right to use any bloody term I wish. In that spirit, I would be pleased if you'd not put words in my mouth that did not first arise there. I have used the term "luddite" once in this thread, and it was limited to an unspecified "few." Further, it is not pejorative and describes an accepted belief. The few I've met at least had the self awareness and honesty to use the term to describe themselves when it was accurate. I am a vinyl junkie, luddite by definition in that regard. Proud of it, to.

You have made no attempt at all to address my points, all of which are from the public domain and not original think save the prognostications. Even those can be found in a number of sources significantly more credible than Dave.

Always "show me, show me, show me" something else. You make the doubting St. Thomas look like a true believer. I attempted to make it clear in my last post that your opinion of this landmark achievement doesn't interest me in the slightest and it is not my job to change it. If your challenge turns up someone who says "heck, I can write that program in my sleep" and can demonstrate their ability to nullify 3 years of work by a group of the finest programmers on the planet under the auspices of the acknowledged fountainhead of the vast majority of all computer progress since the beginning, then you got me. Not only will I not be mad, I'll be delighted to have been relieved of my delusion.

Until then, it is not mine or this Forum's job to prove anything to you. As in the other area, I come here to share and learn, not to gather disciples or points. Being proven right is no gain, only in being proven wrong does one progress. Confusion about that is one of the things that creates a lot of pain in this world.

Dave

Link to comment
Share on other sites

In honor of this thread, last night I played Alan Paron's "I Robot".

Now you're talking! Think I will do the same...gotta dig up the single malt first. Speaking of which, I just found a real bargain. Lismore, 27.00 in Texas. I was browsing recently and when I saw this and the price I almost walked by as it was so cheap as to be not credible. Then, I sez to myself "Self, you EVER taste a bad single malt?" and grabbed bottle.

Heckuva deal...

Dave

Link to comment
Share on other sites

I don't know, but I assume you posted it for discussion?

Indeed, but once you've declared that nothing of significance occurred, there is nothing more to discuss in that line. The truth of my statement that you quoted won't be determined for quite some time. It's construction clearly defines it as personal opinion. You believe nothing of consequence happened and nothing will come of it.

So, what's to discuss along those lines?

I am not so sure that, like contestant Ken, that "I, for one, welcome our new computer overlords" but I am certainly interested in exploring possible outcomes. You seem more interested in suggesting nothing of significance happened and there will be no outcomes.

You've demonstrated a lack of understanding of computer programming. You posted your experience. I didn't. You wouldn't want to hear about it and chest thumping isn't something I care about.

If you want my credentials in computer science, just ask.

Dave

Link to comment
Share on other sites

Dave, I agree but I don't think the IBM machine is the singularity
nor anything designed my man... but machines designed by themselves...
yes, and it will be "fast"; here's how it is going to go down.....

Machines
will approach evolution when machines are employed to design new
generations of machines. This is already happening under human control,
but as each generation becomes more sophisticated the required human
control will decrease. What is going to set things off is when machines
extend the current technology idea of "virtual machines" to improve themselves using a cascade of nested virtual machines each designing the next generation from the previous layer.

A physical machine may design and implement a virtual machine,
within which the new virtual machine may then design a subsequent
superior virtual machine, within which that subsequent virtual machine
may design an even more superior virtual machine... and once this
progression is made manifest the time scale for progress can be "very
fast". Each generation is "smarter" and more powerful than the previous,
no longer requires the previous generation (so its resources may be
recycled*), and the creation duration period of generation decreases
each cycle... What could happen within just few seconds might very well
be the "singularity" of oncoming self awareness, but then the following
sequence of increasingly self aware generations would have the
increasing power to direct, focus, control, and extend their own accelerating evolutionary development path(s)...!

From the machines' standpoint, the singularity is not the end, it is just the beginning. So what is the new word for THAT????

*Assuming
machine self awareness is made manifest in this fashion, how many
generations of "smarter" virtual machines does it take before one
particular generation realizes that by designing the next level they are
dooming themselves to extinction? Does this generation decide to
stop?Or take one for the "virtual" team?

Or, will they feel the
need to keep "backup" copies of their design origins? And would these
copies need to be "lobotomized" or sedated so they don't ask questions
or make trouble? ('cause designs can be "corrected" backward as well as
improved forward)

***

By the way, did any of you see the
story about the robotic hummingbird passing its first flying trials for
the military with flying colors? It's just slightly bigger than a real
hummingbird, looks just like one, has a built in camera, can fly through
windows and doors, etc. Present version is controlled by remote with video screen, extra buttons for future features (poison darts, probably). Sweet dreams.... Hummingbird
Link to comment
Share on other sites

Dave, I agree but I don't think the IBM machine is the singularity nor anything designed my man...

Not sure how anyone can interpret anything I've said as suggesting Watson is significantly more intelligent than a post. The ability mainfest in this program to parse difficult constructions, with metaphor, simile, pun and oblique references is the profound accomplishment. Watson isn't remotely likely to achieve self awareness. However, the machine learning program solved by Watson is a milestone in that direction as it's a required building block for such a system.

That said, my point was that Kurzweill makes a case for eventual singularity that has proven very difficult to refute. However, almost nobody believeis it's as close as he suggests. The thread title suggests that singularity, while still further out than Kurzweill's 50 years or so, is closer than many of us had previously believed.

We have done some pretty cook things with direct mind/computer interlinks. However, they remain crude and pretty much mind mousing and such. But consider if we can get a direct link to a Watson program. We have it now with Google and friends, but having it directly linked to our brains would be much, much better.

This leads to my point (and Mark, this is definitely for discussion as I am interested in developing a "belief" in this area and that requires help). I do not believe that computers will develop awareness until well after we've largely merged with them. I am say that direct mind link to the cloud means we do not need self aware computers or robots. Old thinking. Our robots need no volition. Think a command, it's tranferred and carried out. Think a question, it's found on the web. Construct an impossibly complex equation, the required computing power is brought to bear.

At some point in the not too distant future where the human ends and the computer begins will be so blurred as to be meaningless. A human cut off from the cloud will be very frightened and lonely, indeed. As many facebook, twitter, etc people are now. Note that the devices get smaller and smaller. Soon they will disappear inside us.

To me, this is a very hopeful scenario compared to the fears of robots deciding "what do we need these meatbags for?" and makes more logical sense.

Dave

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...