The Singulairty is nearer than you think...

[Mallette]

LOL Dude!!!!!!!!!!!!!!!!!!!!!!

Glad you took that as intended! Simply couldn't help myself...[6]

Dave

[Mallette]

Just got back from the Society for Applied Learning Technology confab in Orlando. The session chair for the track in which I presented was Dr. Peter Rizza, head of the Princeton Education Center. I asked him about Watson, and he lit up and said "Wowie. Paradigm shift. NASA is all over it..."

Maybe more telling was Dr. Robert Ogilvie, Professor Emeritus of the Medical University of South Carolina who said "This is going to be a complete revolution in diagnostics. These algorithms as they stand right now are equal to the finest diagnostic minds on the planet, and only a privileged few at the moment have access to these. It won't be long until there is a app for that."

There was more, especially from the DOD crowd.

Seems I am not the only one rather impressed with this "parlor trick."

Dave

[Don Richard]

I just saw in the news that congressman Rush Holt of New Hampshire beat Watson (What, Son?) Monday night in a match of Congressmen vs machine in a Washington hotel. Holt was a five-time champion on the original Jeopardy show 35 years ago.

I had a cell phone that would dial a number when I spoke someone's name.

I just got a renewed credit card that I computer-activated by speaking my number into the phone.

My vehicle speaks to me when I do not fasten my seat belt.

My GPS speaks and tells me how to get where I want to go.

Hardware and software which make computers appear to be intelligent have been around for quite a while. However, no computer can think or even come close to abstract thought. A chimpanzee thinks on a higher level than the best computer that we are likely to have for another 60 years or so, if ever.

[fini]

Someone should train a chimp to give driving directions. That would make road trips FUN again!

[Mallette]

A chimpanzee thinks on a higher level than the best computer that we are likely to have for another 60 years or so, if ever.

I understand your train of thought, Don, but it is on the wrong track. Watson is completely unlike any "intelligence" program to date. I would trust it before all but a handful of doctors on diagnostics and many lives will be saved and a lot of money as well when it is on line as a diagnostics tool...and that is only one application.

I am not going to try to prognosticate when a computer will equal or suppass human intelligence...but it will be a lot sooner than 60 years and I expect it in my life time.

Dave

[Islander]

This high-falutin' talk is all well and good, but could you please correct the spelling of "singularity"? It's been sticking out like a sore thumb since the first post.

[Marvel]

This high-falutin' talk is all well and good, but could you please correct the spelling of "singularity"? It's been sticking out like a sore thumb since the first post.

You can change it when you reply...

[Don Richard]

Watson is completely unlike any "intelligence" program to date. I would trust it before all but a handful of doctors on diagnostics and many lives will be saved and a lot of money as well when it is on line as a diagnostics tool

You do realize that "a handful of doctors" or more will be required to program any computer medical diagnostic application. The computer, therefore, can know no more than what it was programmed to know. By doctors.

All that can be expected is that a faster diagnosis of medical problems will occur as a result of doctors using a computer algorithm. That's it. Not more accurate, just faster.

Medical diagnostic programs will be used by doctors like other computerized medical information is being used by doctors today. Doctors will access the info, review it, and make decisions based on that information. Nothing new there.

[Islander]

This high-falutin' talk is all well and good, but could you please correct the spelling of "singularity"? It's been sticking out like a sore thumb since the first post.

You can change it when you reply...

Thanks, but when I've done that in the past, reactions suggested I'd violated Netiquette and done something rude.

[Mallette]

This high-falutin' talk is all well and good, but could you please correct the spelling of "singularity"? It's been sticking out like a sore thumb since the first post.

Watson checked it...

Dave

[Mallette]

Not more accurate, just faster.

Both, with "faster" being an incredible leap. Increased speed to accurate diagnosis is the single greatest need in medicine.

Doctors will access the info, review it, and make decisions based on that information. Nothing new there.

Very new. Diagnostic programs based on traditional "Deep Blue" type algorithms have been very disappointing to date, according to Prof. Ogilvie whom I referenced earlier. The Watson programs more human-like ability to synthesize missing pieces (House's lament about patients: "They lie.") is a major breakthrough.

Dave

[Mallette]

Forgot this:

You do realize that "a handful of doctors" or more will be required to program any computer medical diagnostic application. The computer, therefore, can know no more than what it was programmed to know. By doctors.

Not true at all. Bear in mind that the Watson diagnostics program will be partly in the "cloud" and have access to all the accumulated knowledge of the world. Once the basic programming is complete it will learn on its own and do so the moment new information is stored. No latency as with the current system, which can take months or years for new findings to filter down to Marcus Welby's office.

Dave

[Don Richard]

You do realize that "a handful of doctors" or more will be required to program any computer medical diagnostic application. The computer, therefore, can know no more than what it was programmed to know. By doctors.

Not true at all. Bear in mind that the Watson diagnostics program will be partly in the "cloud" and have access to all the accumulated knowledge of the world. Once the basic programming is complete it will learn on its own and do so the moment new information is stored.

So Watson is going to be able to do medical research and generate its own data? No. Doctors and research scientists, humans all, will experiment, collect data, and analyze those data before it is "stored". Computer medical diagnostics will always just be providing information for human doctors to consider in deciding a course of treatment. The lawyers won't let it happen any other way.

[Mallette]

So Watson is going to be able to do medical research and generate its own data?

Yes and no. Watson will make extrapolations not heitherto made based on existing data. Further, as I mentioned, new data will be availble to Watson's successor long before your local MD ever hears about it.

As to the lawyers, as this system will be far more reliable than your local hydrocodone-impaired physician the litigators will hate it, and the defense lawyers will love it.

It is not going to be the final word. As you said, a doctor will access and evaluate the output. However, it is likely to be right more often and a lot faster than the vast majority of primary care physicians available at the moment.

Your MD is far better equiped to understand and endorse a diagnosis than to make it in the first place.

As I have a daughter with a very complex syndrome of issues from a congenital heart defect who is treated at Texas Children's Hospital by some of the worlds' finest physicians...whov'e often taken too long to figure some things out...I cannot wait for them to have this tool at their disposal. Each of them is a specialist who looks for issues in their specialty and often find them, even though they are not there. The Watson diagnostic will be a generalists with complete knowledge of all these specialties and provide an interdisciplinary tool that does not now exist.

Dave

[Don Richard]

So Watson is going to be able to do medical research and generate its own data?

Yes and no. Watson will make extrapolations not heitherto made based on existing data.

Watson can only make either/or decisions. It can, however, make many of them very fast. That does not mean that it can think. Interpretational analysis requires three dimensional thought. Watson has clearly demonstrated that abstraction is beyond its capability. Several humans have now beaten Watson at its own game, that game being the only thing that it "knows" how to do. This in spite of it having an advantage in button pressing speed.

As to the lawyers, as this system will be far more reliable than your local hydrocodone-impaired physician

Uh, Dave, of course House is fiction, I hope you understand.

[Jay481985]

I have not read this post but.... My state senator is beat watson []

http://www.nj.com/news/index.ssf/2011/03/rep_rush_holt_defeats_ibm_supe.html

[Mallette]

Watson can only make either/or decisions.

You are thinking Deep Blue. Your statement is a common misconception about Watson. This program contains a mass of new machine learning algorithms allowing human-like deductive, inductive, and cross-relational "reasoning." As to occasionally being beaten by humans, bear in mind this one is free-standing with only the information contain in its memory and that it is Rev 1.0. Further, the humans that have beaten it are the best we have. Few patients get the best we have, and the bulk don't even get the better ones. I'd go with Watson over my personal MD, who is very good but probably wouldn't last long in a match with Watson.

As to "House," he fictionally represents the best we have and medical professionals don't miss that show. Due to my daughter's health issues I am exposed to many first class medical pros and they are always discussing who made the diagnosis first in that show.

Dave

[Don Richard]

Watson can only make either/or decisions.

You are thinking Deep Blue. Your statement is a common misconception about Watson. This program contains a mass of new machine learning algorithms allowing human-like deductive, inductive, and cross-relational "reasoning."

Algorithms made up from either/or decisions. Simple logic tied together in such a manner that the machine appears to think.

Logical ops are one of the four things a digital computer can do. Let's look at it's math capability. It can add binary numbers. At the hardware level that is all it can do. There is no algorithm that can devise the abstractions of which the human mind is capable. I'll say I'm wrong when Watson can come up with the equivalent of Einstein's E=MC^2.

Look at what it takes to get Watson to work - a room full of equipment, pushing our technology to the limit. I just read that we humans use only 5% of our mental capability. And some of us can already beat Watson.

As to occasionally being beaten by humans, bear in mind this one is free-standing with only the information contain in its memory and that it is Rev 1.0.

Humans contestants are also "free-standing with only the information contained in our memory". Sorry, but I do not know what Rev # we humans are on now.[]

[Mallette]

Look at what it takes to get Watson to work - a room full of equipment, pushing our technology to the limit. I just read that we humans use only 5% of our mental capability.

That's the free standing part. The code itself requires only a tiny fraction. Look at your brain to body size ratio. Without the body, your brain is neither useful nor can it survive. As to the old "5%" thing, that is an old saw that basically tells you how little we understand about how the brain works. I assure you that if they remove 95% of your brain you are done with this conversation.

Sorry, but I do not know what Rev # we humans are on now.

Homo Sapiens is at Rev. 1. Homo Sapiens 2.0 will be a human/computer hybrid providing direct brain connection to the web and other resources. First steps have been demonstrated and progress is pretty rapid...though quite a way to go. We've constantly thought about this as "us" and "them" both in science fact and science fiction.

I've come up with an alternative, as Pogo the Possum once said in the long gone comic strip: "We have met the enemy and he is US." I suspect that the path to "computer intelligence" and intelligent robotics is not through independent, free standing machines, but through hybridization directly with humans.

Consider the successful implants allowing a paralytic to directly manipulate a mouse cursor with the mind. As that is gradually extended, one will eventually simply do this interface without thought. Your "memory" will seamlessly include everything on the WWW. You will not know where your brain ends and the computer begins. Further, the Future Warrior Project of DOD with its exoskeletons and computer assists suggests a different path to robots. Your household robot need not ever be able to "think," but simply act as an extension of your mind and no more capable of independant action or "revolt" than your hand.

While I've gotten a lot of pushback on this idea (though mainly just "I don't like it" without any science based evidence) it's my belief there will never be an "us" against "them" showdown, but an evolution towards a homo sapiens 2.0 and more based on these concepts.

Dave

[Mallette]

Reminds me of the early comments about TV: "Just radio with pictures."

Our potential remains as homo sapiens 1.0 until our ability to interact with the computer "oversoul" is seamless. Once we cross that line, the way we access and process information today will remind us of the patent office guy in the late 19th century who suggested closing that office as everything of importance had already been invented.

All I've suggested here is that we are on the threshold of an incredible advance. I am completely incapable of prognisticating the impact of that advance as it's far more than a simple paradigm shift. I get to thinking about things I have mixed emotions about. All here know how critical I feel the release of humans from this prison planet is. But, given the "hybridization" above and the potential it offers to be in sensory and exo control of anywhere the network extends I can see many saying "Why the cost and risk of going to (planet X in space) when we can be there directly in effect?" And that is a good question. An "Asimo" with all mechanics and sensory directly inhabited by a human would be the same as being there.

Anyway, as I said, attempting to predict when the road changes direction as completely as we are on the brink of doing is almost always completely futile.

Dave