Slashdot Mirror
IT Snake Oil — Six Tech Cure-Alls That Went Bunk
snydeq writes "InfoWorld's Dan Tynan surveys six 'transformational' tech-panacea sales pitches that have left egg on at least some IT department faces. Billed with legendary promises, each of the six technologies — five old, one new — has earned the dubious distinction of being the hype king of its respective era, falling far short of legendary promises. Consultant greed, analyst oversight, dirty vendor tricks — 'the one thing you can count on in the land of IT is a slick vendor presentation and a whole lot of hype. Eras shift, technologies change, but the sales pitch always sounds eerily familiar. In virtually every decade there's at least one transformational technology that promises to revolutionize the enterprise, slash operational costs, reduce capital expenditures, align your IT initiatives with your core business practices, boost employee productivity, and leave your breath clean and minty fresh.' Today, cloud computing, virtualization, and tablet PCs are vying for the hype crown." What other horrible hype stories do some of our seasoned vets have?
The bad news is that artificial intelligence has yet to fully deliver on its promises.
Only idiots, marketers, businessmen and outsiders ever thought we would be completely replaced by artificially intelligent machines. The people actually putting artificial intelligence into practice knew that AI, like so many other things, would benefit us in small steps. So many forms of automation are technically basic artificial intelligence, it's just very simple artificial intelligence. While you might want to argue that the things we benefit from are heuristics, statistics and messes of if/then decision trees, successful AI is nothing more than that. Everyone reading this enjoys benefits of AI but you probably don't know it. For instance, your hand written mail is most likely read by a machine that uses optical character recognition to decide where it goes with a pretty good success rate and confidence factor to fail over to humans. Recommendation systems are often based on AI algorithms. I mean, the article even says this:
The ability of your bank's financial software to detect potentially fraudulent activity on your accounts or alter your credit score when you miss a mortgage payment are just two of many common examples of AI at work, says Mow. Speech and handwriting recognition, business process management, data mining, and medical diagnostics -- they all owe a debt to AI.
Having taken several courses on AI, I never found a contributor to the field that promised it to be the silver bullet -- or even remotely comparable to the human mind. I don't ever recall reading anything other than fiction claiming that humans would soon be replaced completely by thinking machines.
In short, I don't think it's fair to put it in this list as it has had success. It's easy to dismiss AI if the only person you hear talking about it is the cult-like Ray Kurzweil but I assure you the field is a valid one (unlike CASE or ERP). In short, AI will never die because the list of applications -- though small -- slowly but surely grows. It has not gone 'bunk' (whatever the hell that means). You can say expert systems have failed to keep their promises but not AI on the whole. The only thing that's left a sour taste in your mouth is salesmen and businessmen promising you something they simply cannot deliver on. And that's nothing new nor anything specific to AI.
My work here is dung.
Not sure why virtualization made it into the potential snake-oil of the future. It's demonstrating real benefits today...practically all of the companies I deal with have virtualized big chunks of their infrastructure.
I'd vote for cloud computing, previously known as utility computing. It's a lot more work than expected to offload processing outside your organization.
Bonus points if the salesman admits that he doesn't need to know your problems before selling it to you.
That went over real well once they saw user visits drop by almost half...
There's a pattern here. Many of the hyped technologies eventually find a nice little niche. It's good to experiment with new things to find out where they might fit in or teach us new options. The problem comes when they are touted as a general solution to most IT ills. Treat them like the religious dudes who knock on your door: go ahead and talk to them for a while on the porch, but don't let them into the house.
Table-ized A.I.
Most of the technologies in the article were overhyped but almost all have had real value in the marketplace.
For example, AI works and is a very strong technology, but only the SF authors and idiots expect their computer to have a conversation with them. Expert systems (a better name) or technologies that are part of them are in place in thousands of back-office systems.
But, if you're looking for HAL, you have another 2001 years to wait. Nobody seriously is working toward that, except as a dream goal. Everybody wants a better prediction model for the stock market first.
We used to play buzzword bingo when vendors would come in for a show. Some of my personal favorites:
IT Best Practices - Has anyone seen my big book of best practices? I seem to have misplaced it. But that never stopped vendors from pretending there was an IT bible out there that spelled out the procedures for running an IT shop. And always it was their product at the core of IT best practices.
Agile Computing - I never did figure that one out. This is your PC, this is your PC in spin class.
Lean IT - Cut half your staff and spend 3x what you were paying them to pay us for doing the exact same thing only with worse service.
Web 2.0 - Javascript by any other name is still var rose.
SOA - What a gold mine that one was. Calling it "web services" didn't command a very high premium. But tack on a great acronym like SOA and you can charge lots more!
All those are just ways for vendors and contractors to make management feel stupid and out of touch. Many management teams don't need any help in that arena, most of them are already out of touch before the vendor walks in. Exactly why they're not running back to their internal IT people to inquire why installing Siebel is a really BAD idea. You can't fix bad business practices with technology. Fix your business practices first, then find the solution that best fits what you're already doing.
And whoever has my IT Best Practices book, please bring it back. Thanks.
That's our life, the big wheel of shit. - The Fat Man, Blue Tango Salvage
The most obvious counterexample to the "AI" nonsense is to consider that, back around 1800 or any time earlier, it was obvious to anyone that the ability to count and do arithmetic was a sign of intelligence. Not even smart animals like dogs or monkeys could add or subtract; only we smart humans could do that. Then those engineer types invented the adding machine. Were people amazed by the advent of intelligent machines? No; they simply reclassified adding and subtracting as "mechanical" actions that required no intelligence at all.
Fast forward to the computer age, and you see the same process over and over. As soon as something becomes routinely doable by a computer, it is no longer considered a sign of intelligence; it's a mere mechanical activity. Back in the 1960s, when the widely-used programming languages were Fortran and Cobol, the AI researchers were developing languages like LISP that could actually process free-form, variable-length lists. This promised to be the start of truly intelligent computers. By the early 1970s, however, list processing was taught in low-level programming courses and had become a routine part of the software developers toolkits. So it was just a "software engineering" tool, a mechanical activity that didn't require any machine intelligence.
Meanwhile, the AI researchers were developing more sophisticated "intelligent" data structures, such as tables that could associate arbitrary strings with each other. Did these lead to development of intelligent software? Well, now some of our common programming languages (perl, prolog, etc.) include such tables as basic data types, and the programmers use them routinely. But nobody considers the resulting software "intelligent"; it's merely more complex computer software, but basically still just as mechanical and unintelligent as the first adding machines.
So my prediction is that we'll never have Artificial Intelligence. Every new advance in that direction will always be reclassified from "intelligent" to "merely mechanical". When we have computer software composing best-selling music and writing best-selling novels or creating entire computer-generated movies from scratch, it will be obvious that such things are merely mechanical activities, requiring no actual intelligence.
Whether there will still be things that humans are intelligent enough to do, I can't predict.
Those who do study history are doomed to stand helplessly by while everyone else repeats it.
AI already has successes. But, as an AI researcher friend of mine points out, once they succeed it's no longer 'AI'. Things like packet routing, used to be AI. Path-finding, as in games, or route-finding, as with GPS: solved. So yes, AI will never arrive, because AI is always 'other than the AI we already have.'
A.
That's not CASE. CASE was about translating requirements into code without human involvement. Your examples are all about abstraction and APIs.