The Deep Blue computer that beat Kasparov was far from a simple desktop PC that you could order from Dell. It was based on 30 to 40 years of research and development and was designed specifically for the one task of beating a chess master. A crack team of IBM's best people was put together to build and programme this computer to beat a chess champion.

Deep Blue was the very first IBM computer to beat a champion after decades of previous attempts. Newer versions of Deep Blue have now been created and each incarnation is more efficient than the previous one.

Chess virtuosos play out moves in their head, and from years of experience they can recognise detailed patterns and moves, while the IBM machine worked like some neurotic Woody Allen character, analysing every single potential move and outcome in the blink of an eye and reexamining every possibility after every other subsequent move of a piece on the board.

It is this pattern recognition, which humans are very good at, that has set us apart from the machines, but the fuzziness of pattern recognition is now coming to a computer screen near you.

Now when you go shopping on Amazon. com, the site will learn your tastes more fully the more often you buy from it. When you start buying a few items, Amazon suggests other books based on what other people have purchased, but the more you use the site the cleverer it becomes, and it will eventually start emailing you about new products that it thinks youmight like.

Music sites such as Last. fm and Pandora do the same with music. The more music you listen to with them, including music you say you dislike, the more they'll tailor their service to you, to the point where they will design a personal radio station for you, playing music they think you'll like. The kickback for these sites of course is that they'll offer to sell you this music that you like. Good for them and good for you.

Google too is going into this game with its search service.

Last month, Google switched on its web history service so that if you are logged into your Gmail account as you do searches on Google, it will store what you searched for and what you clicked on. Over time it will rejig the search results in its listings to results it thinks you'd prefer. This helps you out and it also helps Google to identify more valuable information which it can share with the rest of its users.

But we could end up depending too heavily on recommendations. In Britain recently, a woman who took the word of her car's satellite navigation system as gospel ended up driving onto a railway track and being hit by a train. We need to realise that complex systems can create complex errors.

All of these suggestion and guidance services are a simple way of easing even the most ardent luddite into trusting automated technology a little bit more. However, these services do not consider the randomness of taste that humans have. So while these services will give us more of what we like, there isn't the option to try something we might not have considered before.

The technology is still not complex enough to beat the type of character played by Jack Black in the film High Fidelity, who gets people to try something new and different from what they typically like.

This is why, despite researchers loading all the best pop tunes into a massive number-crunching machine and trying to get the computer to produce a cheesy pop song, they still cannot produce anything near the dodgiest of You're a Starwinners. This is also why . . . so far . . . computers cannot create anything as good as the Mona Lisa. Computers do not have a creative flair or the ability to recognise beauty in something unseen. They will not 'know it when they see it'.

Technology has come a long way since Deep Blue and is helping us to make more informed decisions, but without gut feeling and instinct, technology will still be more assistive than truly creative. To put it in terms of an out-of-fashion election slogan, a lot done, a lot more to do.

THE PUBLIC SERVICE DEBATE ON RTE

EOIN O'DELL of Trinity College recently petitioned RTE to give away the raw content of the Prime Time election debates so that others could distribute it on the web and put the debates in podcasts or on YouTube, disseminating the content to as wide an audience as possible.

RTE respectfully declined the request, saying the archive of the debates can be seen in a low-quality and restrictive video format on its website. But with TV cards in computers and DVRs under TV sets, high-quality formats are already available on YouTube and on filesharing networks.

As a public service broadcaster, maybe RTE should give the people what they want.