Worst and Best Predictions on Technology

prostoalex writes "Dow Jones News asked several mahor scientists and technologists about their worst and best predictions of the future. The story, republished at Yahoo! Finance Singapore quotes Lester Thurow, Professor of management and economics, Massachusetts Institute of Technology's Sloan School of Management; Nicholas Negroponte, Founder and director, Massachusetts Institute of Technology Media Lab; Glover Ferguson, Chief scientist, Accenture; Alan Nugent, Chief technology officer, Novell; Peter Cochrane, Director, ConceptLabs; Michael Earl, Dean, Templeton College, University of Oxford. There seems to be a common agreement on having overrated the ability of machines to talk back to users and vice versa."

Comments

The worst thing is that it's all boring nowadays

[Anonymous Coward]

Nothing really imaginative on the horizon.

CPUs, for example - who cares anymore? 95% of us couldn't care less about upgrading, and the chances are that the last upgrade was just for the sake of it, not because we really needed a faster processor. Not like 10 years ago, when it was, 'Wow! The Pentium is going to be a big leap forward'.

Memory - so cheap, who cares anymore? Even 5 years ago, I was thinking, 'Wow! I've finally managed to afford 128 megs of RAM!!!' Most other people had 32 or less. Now, who cares? I could afford a gig of RAM, but what's the point?

Hard disks - mine is about 20% full, and has been for months. No need to upgrade.

Monitors - the few people who actually need a screen bigger than 17 inches can now afford them. LCD monitors are no longer a novelty.

Mice - optical mice are no longer a novelty

Bandwidth - OK, so ADSL is still 'exciting', but for how long? In two years, anybody who wants it will have it.

Optical storage - recordable DVD is here. CD-R is rediculously cheap. Who needs more storage than that?

OK, that's hardware, what about software?

Linux kernel - it's excellent. However, the excitement of a few years ago is dwinding. Don't get me wrong, Linux is excellent, but now that we've got a really good free *nix, the fun of developing a really good free *nix isn't there.

GNU/Hurd - maybe oneday this will become interesting :-)

Windows - I hate Windows, but at least the launch of 95 was interesting. The lack of initial enthusiasm for 98 was interesting. After that, it got boring. Now, it's just more and more waffle about DRM. It's *boring*.

The only things I can see on the horizon that might be interesting are:

* IPV6
* Linux on non-i386 platforms.

Slashdot should do this!

[Snaller]

Make a subject where the users can enter their predictions about the future - then we return in ten years and check it out :)

Re:Slashdot should do this!

[InfoVore]

You should check out the Foresight Exchange.

Basically it is an idea stock market. When you become a member, you receive a small amount of fake investment money. You can then buy and sell against ideas posted by other members. The premise is the the closer an idea is to being true/possible, the higher its value will be in the market. Ideas do have adjudicators who are responsible for judging when and if a stock has met its criteria and can be pulled off the exchange.

Here is an example of the top 10 traded ideas on Foresight Exchange now:

Rank Volume % Symbol Short Description

1 26234 83.4% T2007 True on Jan 1 2007

2 1034 3.3% BBRP Bal Bdgt 2002 w/2000 GOP Pres

3 803 2.6% USIraq US attacks Iraq in a year.

4 437 1.4% HURR02 Atlantic Tropical Storms 2002

5 371 1.2% ObL1yr Osama bin Laden 1 year after

6 275 0.9% $bill U.S. Prints New Dollar Bill

7 222 0.7% SCHRDR Schröder Remains Chancelor

8 193 0.6% Clone Human Clone before 2005

9 160 0.5% King Prince Charles remains heir

10 154 0.5% SLvl 1 m rise in Sea Level

Problem with all predictors - no metrics.

[wowbagger]

I wish the article had presented a bit more background on these guys predictions than "Here's the worst, here's the best, here's the current". That really doesn't let me gauge whether these guys are making good predictions or not.

Consider Slashdot posts: You might say that my highest rated post is 5, my lowest -1, and my most recent is 3. But, does that give you any real feel for whether you want to read my posts? Now, if you said that my mean post value was 3.5, my mode was 4, and that only 10% of my posts are rated less than 2 (NOTE: all figures are made up - I don't keep that close track on my moderations) then you might be able to judge better.

Simillarly, when judging someone's ability to predict where things are going, I'd like to know what their ratio of hits to misses are. If somebody is right no more often than they are wrong, then I can weight their prediction accordingly.

That's one of the problems I had with Tomorrowland at Disney - it's nothing but a bunch of predictions from the past. I'd rather they have done a "Yesterday's Tomorrow" - for every decade show what people thought the future was going to look like, along with a reality check. Show the things they got wrong (flying cars), the things they got right (television), and the things they completely missed (computers).

OT: is anybody else having problems getting to /.? For the past week I've had a timeout on about 1 in three connections to /., both from work and from home.

If you build it, they will come.

[sketerpot]

Since lots of CPU, RAM, hard drive space, and other stuff are now available, who cares? I do. Once more computer capabilities become available, software will be made that will take advantage of them. I know that games are pushing technology forward because they are always finding some way to fill all of the capacity of modern computers. Doom 3, for example, now has souped-up lighting because computers can handle it. Realtime raytracing (and raytracing in general) could definately benefit from faster processors--and as processors become faster, neat new computationally expensive things will become widespread, because they can be.

I agree with you on some other things that could become interesting. IPV6 could allow IP addresses everywhere, which will probably be taken advantage of. It also supports packet prioritization, which would be very good for VoIP and related technologies.

Linux already runs on several non-i386 processors, and it is commonly used on these in, for example, embedded systems. Embedded systems, I think, are quite exciting. And Linux (or one of the *BSD's) will probably be the kernel of choice for those, since the idea of putting an OS on one of those is to allow the device to be programmed easily and not be noticed by the user. From that perspective, Linux is obviously superior to any harder-to-develop-for OS that you have to pay for, like Windows.

I'm still excited about the future. Are you?

Hard problems

[Gerry Gleason]

Yes, I think it is very interesting that so many AI problems continue to be much more difficult than many predict. Even successes like chess playing programs beating all human players, although it has happened, the way it is done is not particularly satisfying.

Thurow is an economist, not a scientist or engineer, which is why his predictions about biotech are particularly bad. The science is on the edge of a lot of new understanding and breakthroughs, but that will only put us up against the really interesting and hard problems. As if we would be able to find genes that more or less directly influence something as subtle as IQ.

I find the predictions about the future importance of web services and the junk about "insight" to be particularly inane. On the first, nobody should forget that GM and Ford are still about the only companies that represent a percentage of the U.S. economy. Manufacture of physical goods (and commodities production, etc.) will continue to be the drivers of economies.

In my opinion, the most important trend is a favorite of this forum. The growth factors that have been working for Free software are fundamentally exponential, even if the constant factor is small. If it isn't killed off by legal/social influence of current big players, and I don't think this is likely if it is even possible, then the exponential term will eventually dominate.

When this plays out, the companies that make their reputations by being the best at efficiently building and servicing products that are mostly designed in the "Creative Commons". People will pay for quality in goods and services, and there will always be value in good execution. Customers do not value "insight" as described in one prediction. They find this sort of thing invasive and manipulative, and you won't be able to keep it secret.

It was when I was chasing down some secondary links from the GNUradio interview that I came across the stuff about the value of a network increasing at greater than linear rates. You get O(N) for broadcast networks, O(N^2) in peer to peer networks, but the exponential (O(2^N)) comes in when you have group forming networks (GFN).

When you think about it, this is what drives the GPL software phenominon. Every project fork or new initiative forms a new group or groups in the network, and every project is a nucleus for new group formation. The only way this could be stopped is to destroy to possibility of the group forming that leads to the exponential growth. While this might be possible, our robust institutions that support free speech make this very difficult if not impossible.

So my prediction is that Linux on the desktop will overtake Windows in the next ten years, and the RIAa and MPAA will finally lose out to the best interests of the actual artists they claim to support. Also, derivitives of GNUradio will be core technology in establishing cooperative wireless mesh networks. This is the only prediction of any of the pundits in the article that will come true.