Here I'm going to try to write down some of my predictions for future technological development, so that I can check them later. I'm assigning things I'm more sure about a higher "certainty level or 'CL'", which I'm thinking will be in a range of 0-9.
Inspired by http://kk.org/ct2/2008/04/digital-things-ive-been-wrong.php.
Second Life (or equivalent): going to be the basis of a lot of socialization in 10 years, if they can make it decentralized (CL 2); primarily among groups of people who know each other in real life, but aren't physically co-located at the moment (CL 2).
Spore: going to be a huge seller (CL 3) and even bigger influence on newly designed games (CL 4) (and maybe other programs (CL 2)) over the next 5 years.
Automated fabrication (one-off production of goods without direct human intervention): starting around 2010 and by 2013, it's going to replace more than 5% of traditional manufacturing, measured by value (CL 4). Price-sensitive goods will continue to be made by traditional manufacturing unless and until automated fabrication gets cheaper than traditional manufacturing plus transportation (CL 5), which I don't think will happen in 5 years and probably not in 10 (CL 3).
Version control: in five years, the mainstream version control systems will include SVN (CL 5), Git (CL 5), and Mercurial (CL 4), but not Darcs (CL 3), Bazaar (CL 3), any other arch variant (CL 4), CVS (CL 2), or Perforce (CL 4). They already don't include RCS, PVCS, SCCS, CSSC, Aegis, Vesta, or BitKeeper. There will be important new alternatives that I haven't heard of yet (CL 4) but none of them will be as popular as Git (CL 4). A lot of things that aren't traditionally considered source code will be stored in these version control systems, like textual documents.
Perl 5 will still be in wider use than Perl 6 five years from now (CL 2), but neither will be in as wide use as Ruby or Python (CL 3).
DiSo will have a lot of users in 5 years (CL 1) but still won't be as popular as Facebook (CL 3) or MSN Messenger (CL 5) and won't be as popular as some social networking system I haven't heard of yet (CL 3).
General-purpose consumer microcomputers that are popular in the EU will cost between one and four barrels of light sweet crude (CL 3), and computers at that price will still execute single-threaded code at under 5 billion basic blocks per second (CL 5). Today's equivalents (which cost a bit more) can typically do about 8 threads at full speed; five years of Moore's law would make that 32, but I think that computers at that price five years from now are likely to be able to run at least 128 threads at this 5G-basic-blocks-per-second speed (CL 4). Multithreading will provide a decisive performance advantage to software written in a variety of ways that aren't currently mainstream, such as Erlang's model, Verilog, and array-processing languages (CL 3), rather than providing an equally large performance boost to software written with shared-everything threads and locks, as Herb Sutter predicted. High-performance software being written in this way, combined with a lot of experience with server virtualization, will create an opportunity for new CPU architectures that don't directly support x86, ARM, or MIPS instruction sets --- whether a relatively traditional design like the Tera MTA or something wild like FPGAs, the SeaForth, tagged dataflow machines, or concurrent linear graph reduction machines --- so at least one such architecture will have total market sales 1% or more of the x86-compatibles' market sales (CL 3).
In five years, devices like cell phones, portable computers with embedded radios in which the user does not have ultimate control over the software, will be the way that most people access the internet most of the time (CL 5; this is already true except that most people don't access the internet yet) and will sell ten times as many units per year as the traditional kind of computer in which the user is ultimately responsible (CL 3).
In five years, electoral campaign success in the US at every level except presidential campaigns will be largely determined by word of "mouth" (using electronic communications whose receivers solicit them, rather than paid advertisements) rather than paid advertisements (CL 3). Presidential campaigns will be too (CL 2). In ten years, both of these will be true (CL 4).
In five years, intellectual property restrictions, not terrorism, other aspects of human rights, or agriculture, will be the most controversial issue in international negotiations. (CL 2)
I don't know what to predict about peer-to-peer systems like Vipul's Razor and BitTorrent. I suppose that they will continue to exist, and the kind of people who thought jazz and tango music were scandalous will continue to oppose them (CL 6). And fragile centralized systems like DNS and Google will continue to exist too (CL 6). Whether the mix will change, and how it will affect society, I have no idea.
In the last few years, there have been several hobbyist UAVs and other robots for remote sensing, of which Art van den Berg's glider is probably the most impressive. In the next five years, there will be more of them (CL 5) and some of them will be even more impressive than van den Berg's project (CL 2). Governments will worry and some will make new laws restricting model aircraft (CL 4). Some people will commercialize the technology (CL 3), use it to advance the state of knowledge in some scientific field (CL 3), or do things with it that benefit s lot of people (CL 3). Possible applications include inexpensive high-quality aerial remote sensing, deep-sea exploration, substituting for radio towers and communication satellites, and rapid, inexpensive delivery of small, light items, especially to inaccessible places.
Five years from now, solar energy will be a cost-effective source of electricity in countries that currently derive much of their electrical energy from natural gas or oil (CL 5), because the prices of those commodities will have risen (CL 6). I mean to say that it will be cheaper than those fossil fuels without any government subsidies. It will probably be cost-effective in countries like the US which get most of their electricity from coal (CL 4), because people will figure out how to bring the cost of solar electric systems down and their efficiency up (CL 4).
Five years from now, other renewable sources of electricity will be cost-effective, too (CL 4). Maybe wind, geothermal, old-oil-well geothermal, tidal, wave power, or even biomass.
Five years from now, digital cameras will surveil all urban public places in the OECD countries (CL 4), probably because random individual people will carry digital cameras around with them and always turned on (CL 3), not just because of cameras placed permanently by the owners of those spaces.
Five years from now, despite the rising prices of fossil fuels (see above, CL 6) long-distance transport will remain affordable for most of the world's food (CL 6), so "locavorism" will not become an economic imperative (CL 6).
Five years from now, China will have experienced at least one recession (two consecutive quarters of negative economic growth) (CL 3), as some aspect or other of its current strategy falters.
Within five years, extrajudicial executions of innocent citizens by the police in both the US and Britain will have provoked considerable furor (CL 3), but the general public in both countries will be in favor of such policies (CL 5).
Within five years, the most-widely-consulted cartographic resource will be a community project like Wikipedia (CL 4), and OpenStreetMap will be the most-widely-consulted community cartographic resource (CL 4). Most important scientific papers in hard-science fields will be published first in open-access media such as arXiv, open-access institutional repositories, or open-access journals (CL 4). The same thing will be true of most important academic papers in less scientific disciplines such as history, psychology, and sociology (CL 2). Significant non-open-access journals will still exist, even in hard-science fields, and will still publish some important new results (CL 5).
Five years from now, most people who enter text into computers will still use QWERTY keyboards or phone keypads (CL 4), not Morse code, Dvorak, speech recognition, or something else exotic.
DHTML will still be a widely-used format for new applications five years from now (CL 4) and will not have been eclipsed in popularity by one of the various new alternatives that are being marketed now (CL 3) such as Silverlight and Flash.
Over the next five years, computer and network security will be a bigger and bigger problem (CL 5) as organized crime makes more and more use of computer security vulnerabilities to get money and power (CL 4).
Five years from now, illegally copied software will still be ubiquitous, even in OECD countries, on computers that are controlled by their users (CL 5).
I'm not sure what's going to happen with agriculture. There are clearly better ways to go about it than the Green Revolution approaches, especially in marginal areas like New Mexico, Iraq, much of Jordan and the West Bank, the llanos of eastern Colombia, and so on. Whether those approaches will catch on, who can say?
Five years from now, there will be at least one change that is as important as anything I've mentioned here, but that didn't occur to me (CL 3).
Five years from now, Communist China still will not have produced any proprietary software that is widely used in the rest of the world, except software bundled with hardware, (CL 4) and probably no large piece of widely-used free software either (CL 3).
Five years from now, quantum computers will not be a practical alternative to conventional computers for general tasks (CL 6), nor will they be in ten years (CL 5), but in five years we'll see them do some impressive things (CL 3). (It's possible that they will instead fail to work; if that happens, it will reveal a fundamental weakness in our understanding of quantum physics (CL 4).) Ten years from now, they will do some practically useful things (CL 2).