Elementary, my dear IBM: When will Watson make money? IBM Watson has taken heat from Wall Street for not adding to Big Blue's revenue as the company reported a 19th successive quarter of decline. The Watson brand won worldwide attention back in 2011 when the supercomputer defeated human contestants during a special series of television quiz show Jeopardy. Now the brand lives on …
Good money after bad. Machine Learning is still crap. The majority of current applications are essentially consumer trinkets where "good enough for free" is all that counts (think Facebook or Google's respective creepiness). And much of what is being described as AI/ML is simply a clever algorithm with a minor bit of mathematical feedback, but then crunching through a big data set via parallel processing. Even these systems need a very narrow problem to solve, so the word "intelligence" is very questionable.
You have to laugh at the words "firmly, firmly established". Sounds like a politician's "complete confidence". Having said that, Watson is a brand, not a technology. It is unlikely to disappear, it'll just get more and more debased as every trivial IBM program that addresses a large data set gets branded as Watson, and their marketing dweebs prattle on about machine learning.
Here's the acid test: If Watson is so fucking clever, why haven't IBM used it to chart a strategy of profitable growth?
>Here's the acid test: If Watson is so fucking clever, why haven't IBM used it to chart a strategy of profitable growth?
IBM exec: "Watson, how do we turn around a decade of decline?"
beep beep boop boop whirr whirrr ding.
Watson: "Get rid of the CEO and her Wall Street cronies"
(silence)
IBM exec: "I think he means we need to move more resources to developing countries"
There's entirely too much hype around 'AI' these days, investors shouldn't be worrying about IBM turning a profit on that division. They're investing for the future, and if they're the first to have truly useful AI the billions (probably trillions) will roll in for them.
Imagine where we'd all be today if AT&T investors had insisted on a profit while they were doing R&D that led to the development of the laser, the transistor, and so forth?
Now the investors could argue that its a dead end, or they don't want to wait that long, or whatever, but then don't complain when they don't have any major new markets in a decade. Oh, I forgot, investors only care about the next quarter these days... Maybe they need a new class of stock without voting rights, and can use the proceeds from selling shares of the new class to buy back most of the shares that can vote. Then the people who care only about quarterly earnings can buy the non-voting stock, and the people who want to actually invest rather than speculate can buy the stock with voting rights and take a longer term view.
I think you don't understand history at all. AT&T's investors DID insist on profit, and as a monopoly, AT&T delivered massive profits. AT&T didn't develop the transistor out of a sense of altruism or out of any notion of basic research. The transistor was developed partly to solve a very specific problem in the phone system. Undersea cables needed amplifiers and repeaters. Before the transistor, those components were vacuum tube based. You see the problem with respect to durability when you drop the things a mile beneath the surface of the ocean? So, AT&T had a very specific application in mind for the transistor while it was under development. The story is similar with improvements on the laser (which was not invented at Bell Labs, by the way). The ability to replace heavy, expensive, bulky copper with thin, light glass fiber drove that work.
The point is that the work you cite was driven by compelling business cases -- the R&D was going to pay for itself many times over, and that was known in advance. Now, contrast that with Watson.
It's the cost of all those pesky humans, experts and programmers, that are curating and tuning the datasets. What they need to cut costs is replace them with AI.
After all it's working for the Japanese purchasers of IBM kit?
I'd love to see a break down of HW costs vs ongoing "development costs" for each Watson sale. Or are Watson systems like a sale of a Adobe Creative Cloud or Oracle DB*, a trojan to ensure annual income from support, so it's losing money as it's still at the infiltrate stage (like inkjet printers in supermarkets which have to have only partially filled cartridges as they are so cheap).
(*Though those have an almost zero additional cost per sale unlike Watson sales.)
The single biggest market for AI, IBM and Google don't own. And Microsoft does not have a good AI. I'm talking about managing IP space and switches. And managing Microsoft domains. There is actually a lot less to think about and program into an AI for that. If Watson was coded into the software layer, instead of using shell script and GUI like we do, it would be far simpler for it to think and act. And if it was beefy enough to oversee and manage everyone's remote thin client session, it's hard to see needing a help desk either. Watson would be watching your every click and head you off before you broke things. You'd have two Watson's. One to watch the other. And a service contract with IBM gives you someone to call and unstuck Watson by going through his think log like backing out of a bad transaction in a database.
More PHB bloviating and saying nothing. My take is Itsy Bitsy Morons haven't a clue why the profits are declining and are throwing money at projects and acquisitions in the hopes that something sticks. Unless they get lucky or fire the incompetents in charge they will eventually join Wang and DEC as remanent owned by someone else stupid enough to buy the carcass.
"Do you have visibility as to when we should expect an inflection in revenue recognition from Watson, or should we just not think about this as a contributing factor or moving the needle in our models over the next couple of years,"
When Watson can translate this sentence into intelligible English, it will be ready for prime time.
Attended an IBM event early last year where they famously marketed Watson was used to "design" the lunch menu. Lunch arrives and I get a look at the lunch they are serving up. I walked out of the event to the corner pub for a counter lunch and a couple of beers. Quite a few others joined me.
I attended another IBM event last year at Hursley where they demoed Watson to answer questions about the history of the building.
Even given such a limited set of source data, which consisted mostly of one guidebook only; asking rudimentary questions like "When was the building completed?" were non-functional.
Now, despite this shining recommendation I am aware of other Watson projects being trialled to look at, for example, technical standards a whole lot more diverse and poorly referenced than a guidebook with contents and index pages. Can't wait to see how dreadful the results are.
Incidentally the same demo session also included IBM attempting sell ideas back to us that, not only did we tell them about, but already have in production.
Big Blue has some fantastic technical experts, but in my experience those that can are being poached off elsewhere leaving a shell of a company and a lot of managers that talk the talk but haven't the techie expertise anymore.
IBM came to a recent security event, and spouted a lot of hot air about what Watson could do for security, but when pressed, confessed its all conjecture, because Watson doesn't know security yet.
I suspect this is the problem to monetise it, they need to train several versions for several applications, and unfortunately Jeopardy contestants can't afford to hire Watson.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
