IBM: Playing Jeopardy with tax dollars

2/14/2011 7:44 PM EST

At least this is tax money best spent. Let us have more others like the type IBM provides.

2/14/2011 8:04 PM EST

Of course our tax dollar helped fund IBM's Watson efforts. IBM as a result provides, as you noted, a slew of very advanced computing platforms to the U.S. taxpayer. Most successful businesses run that way, be they here in the U.S. or someplace else in the world.

Watson was clearly a publicity stunt. As was the China supercomputer mentioned. Both were built for the same reason: PR and prestige with the intent of selling their products to a wider audience.

2/14/2011 9:13 PM EST

@Duane: I would not call the Tianjin supercomputer a publicity stunt so much as an exercise in national pride. A shade of difference, I think.

2/14/2011 11:46 PM EST

It doesn't take much ingenuity to crank up the number of processors to gain a performance feat like the Chinese have done; it takes real software breakthroughs like what IBM is trying to show to make supercomputers useful. Playing against humans on Jeopardy shows how computesr could become more aware of their environments and calculate problems in context.

2/15/2011 2:59 PM EST

Your premise would make much more sense if IBM was using a supercomputer, such as Blue Gene, to power Watson. But Watson is powered by IBM's commercial Power7 system -- the same one in use by thousands of commercial customers. Nothing exotic. Standard -- but leadership -- IBM hardware. Clients can buy them today.

Scott Brooks
IBM

2/15/2011 3:02 PM EST

Journalism must be easy if you just speculate and don't bother finding any sources.

Lots of companies contact to the government.

2/15/2011 3:35 PM EST

Tax dollars??? I don't think so. Cash is fungible so money that IBM received from a government could find its way to the Watson Program but after it leaves the government's hands it becomes IBM's money not tax dollars.

The author or the headline writer at EE Times is using a red herring to spike readership.

Tsk, tsk.

2/15/2011 5:18 PM EST

I'd rather our computing tax dollars go to IBM than a foreign firm.

2/15/2011 6:43 PM EST

This is a totally biased article that made obvious from its headline. I can't think of ONE tech firm in the US that has not been helped by the US government's funding one way or the other, which is a fact the author conveniently forgot to mention.

2/15/2011 11:31 PM EST

It wouldn't be a waste of bandwidth and my time, if you could come up with a better way of spending tax dollars.

2/16/2011 7:35 AM EST

@Rick, this is the nature of the game I am afraid. Some good has, does and will come out of such initiatives. As for the question whether it would have been better to spend tax payer's money on other HPC providers, who knows? There is merit in going with big players such as IBM (less risk involved) but there is also merit in spreading the money around to encourage innovation and competition.

All in all, I would rather see tax payer's money spent on projects like these (with a plethora of application that could benefit us all) than see it spent on meaningless and destructive wars...

2/16/2011 7:38 AM EST

This is the second insultingly ill informed column I've read about Supercomputing this week. The first was a piece that proclaimed a new machine going into Argonne was "no cheap date" when, as far as I can tell from the information available it is the least expensive per flop machine out there that isn't using GPUs. In this article on Watson, in addition to the complaints outlined above, it is clear that Mr. Merritt hasn't got a clue about the procurement procedures around COTS based supercomputers. The Petaflops for Picobucks procurement process that has dominated the industry for the past 15 years or so has resulted in SGI going out of business a couple of times, Cray going out of business and being re-invented, HP publicly walking away from HPC until recently, IBM stating that they would not be bidding traditional HPC opportunities, Sun (after the Oracle purchase) walking away from the business, Linux Networks Rest In Peace, Kendall Square Research Rest in Peace, nCUBE Rest in Peace, NCR -- out of the business, Fujitsu -- out of the business, Amdahl -- out of the business, etc.. The list of dead, dying and fiscally self defending companies is long and distinguished.

2/16/2011 7:39 AM EST

(continued)

This crushing push toward Supercomputing based on cheap (and I don't just mean inexpensive) components with no margin for funding innovation has left us with a very sad bone yard of companies who were known for innovation.

There are many reasons for our industry winding up in this sad state: Top500? Partially. COTS mandate by DARPA? Certainly. There are many others. The achievement by my esteemed competitors at IBM should be lauded for the revolutionary approach it is. For once, the industry isn't continuing in the soul sucking innovation death spiral. I for one, hope this coupled with the Peter Koggee piece in IEEE in the last couple of weeks about the serious issues that are waiting for us as we head toward the elusive exaflop wakes some people up.

As for eetimes, I hope the editors take a bit more care in researching their "journalist's" posts. This piece is full of cheap half truths, tired cliche's and mis-statements.

2/16/2011 11:14 AM EST

Having watched the second episode where Watson crushed his human opponents in skill and speed, a strange and scary thought crossed my mind. I don't think the average TV Jeopardy! watcher has a clue as to the significance of IBM's and partners' achievements. The supercomputer's real achievements never came across no matter what the IBM commercial breaks had to say about the IBM feats and its applications in the real world. The problem lies in the sophistication of explaining content within context to the average Watson watcher. In fact, the final answer showed how hard that can be in machine understanding. Category: Cities. Final answer: Which city has an airport named after a WWII hero and another after a WWII battle. Anybody flying thru the midwest and knowing their history can collect on this: Chicago, with O'Hare and Midway. Watson wagered a very small amount because the program algorithm knew that its answer,-- Toronto with six questions marks--, was not a sure bet. Of course, the machine had to calculate millions of variables among cities, and on WWII data points without really knowing the context of all three. We have a long way to go for Watson being reliable in business and scientific feats, and arguing about procurement procedures ain't going to cut it. As an industry we need a political action committee to sway our legislators to the side of nerds and geeks who can solve real problems with powerful machines. But Watson deserves credit for opening the debate.

2/16/2011 6:58 PM EST

The real strength in Watson will be its ability to listen in on every cell phone call in every language and serve to identify and prevent terrorist activity before it occurs.

2/17/2011 4:58 PM EST

BM has contracts for the supercomputers at national labs Los Alamos, Lawrence Livermore and Argonne. Some of these supercomputer contracts are at least partially managed by the National Security Agency. IBM system S was developed substantially by money from the NSA, called distillery internally. Note that IBM was founded on a government contract for the tabulating machines to compute the 1890 census.

2/17/2011 5:59 PM EST

I guess I looked at it from the perspective of advancement of science. Challenges of doing something that is not currently possible are usually the best incentive for great minds to focus their efforts towards a specific task. Kennedy placed that challenge in front of NASA decades ago to put a man on the moon and safely return him before the end of the decade, and it was achieved. The science and technology that was developed as a result has touched all of our lives. I for one congratulate the IBM engineers for their accomplishments. Remember, there are always two sides to a coin. If Watson was successful in beating the humans the engineers are praised, but on the other side of the coin this could have be a complete disaster and the engineers would be selling parts at their local Radio Shack. They stuck their necks out and were successful, and we will all undoubtedly benefit from the new technology developed from their efforts.