Seriously, everyone has their own opinion on this subject, for various (and valid!) reasons. At the risk of entering WW III, I'll offer a few thoughts:
(1)
At the extremes, obviously cables do matter. If we make cables from nylon rope or wooden broomsticks, or trust to the random alignment of air molecules between output and input terminals to provide a "carrier" for our signals, well, we're not going to get very satisfying sound. To say that they make absolutely NO difference is blatant stupidity, but...
(2) Assuming we have a reasonable conductor like a metal wire, other things come into play:
Different metals conduct differently, so the metal the wire is made from will have some effect. Try using a thin Nichrome wire as a cable, if you doubt that. Aluminum, copper, and silver (which are the commonly used cable conductors, perhaps tin (plating) could be added) each are different enough for the differences to be audible in most fairly revealing systems (expecially aluminum, which is quite different and not recommended - it's not even safe if not used correctly). Impurities do affect the flow of current, so OFC will sound better than "dirty" copper wire, although whether your system and your ears will detect that difference is another issue.
(3)
The thickness of the wire will make some difference, and that difference will vary as the length of the wire increases. [I'll avoid technical issues of conductance, resistance, inductance (less relevant), impedance here, but this is true: you will want to use a different wire for that 100-foot run to the back of your garden, than you use in the 8-foot run in your living room.]
(4)
The surface of the wire will make a difference. I had some cable that had gotten moisture inside it, underneath the translucent outer insulation. The outside of the wire corroded and probably some mold grew there; it looked nasty. Out of curiosity, I tried it. It sounded veiled, muddy...just very BAD, whereas another length of exactly the same cable that had not been wet sounded fine. The wire underneath that yucky exterior was unaffected; this was a surface effect I was hearing. Likewise, platings and coatings on wires CAN make a difference.
(5)
The law of diminishing returns applies to cables as much as to anything else. Once you reach the stage of 0.999% pure OFC (Oxygen-free copper) versus 0.9999% pure OFC, I strongly suspect that most systems would never reproduce an audible (if even a detectable, at all) difference.
Manufacturers have to compete for business, so they offer all sorts of techno-babble arguments to sell their products. Some of these are valid. Some are scientifically valid yet irrelevant, since the improvements their cables bring will never actually be heard by human ears. Other arguments are almost certainly fraudulent salesmanship (one is born every minute, and every other minute someone is born who will fleece that guy). That issue aside, beyond a certain point you are adding improvements so small and subtle that you have to strain your ears to hear them at all. Some of the improvements beyond a certain point will be beyond human capacity to detect, anyway.
(6)
The more revealing your system is, the more you will be able to hear the differences (if any) that cables can make.
On MOST of the systems listened to by MOST of the people here, beyond a certain point the improvements simply will not be audible, and spending beyond that point makes no sense. For most of us using vintage gear --decent but not all TOTL units fully restored with dedicated power lines, etc... -- most "decent" cables will be entirely satisfactory. However, for someone with an ultra-high-end super-sophisticated system (that probably cost over $100,000), the differences between cables may very well be audible. For a system at that level, spending a few thousand dollars for a cable might make sense -- but for the rest of us, it is a waste of money. We can get 99% of the quality for a much lower price, and our systems can't reveal that last 1%, anyway.
(7)
Psychoacoustic phenomena are widespread and real in audio applications. We listen with our brains, through our ears, and to some extent we hear what we want to hear and are paying attention to. If we just spent a lot of money on cables, we are listening for improvements, so we are more likely to "hear" them. Also, we notice changes and our ears "habituate" to certain types of sound, so if there is a difference, our opinions of whether it is positive or negative may change over time. This makes it difficult to make really objective assesments of cables, and leads to a lot of disagreements.
(8)
There are a LOT of other things --for example, room acoustics and the influence of what else is in in our listening rooms, and where--
that have a bigger influence on what we hear than our choice of cable. To focus on something that --once it has reached a decent level-- has marginal additional impact on our systems, and leave these other factors unconsidered, is foolish: why spend time and money on something that may add perhaps a 1% improvement, when something else even cheaper and easier may improve things 5%? Your choice of source is hugely important. Your choice of speakers is hugely important. Your choice of preamp is probably next, then the power amp. Your listening space is a huge factor, too. Assumiing you are using decent copper wire and not broomsticks, cables come somewhere after that, in the list of worthwhile things to improve. You might get better improvement from paying more attention to damping your gear...and almost certainly will get a MUCH bigger improvement if you look at the electricity that is going into your gear in the first place (run a separate "dedicated" power line and ground, isolate from sources of "noise" like refrigerators, etc...).
Personally, I have some Japanese 12 gauge OFC cables and some DIY Cat 5 twists and a scattering of branded others I've picked up over the years. I've played with them in different systems, and find that while there are some subtle differences, especially when the lengths get longer, a lot of other things make a much bigger difference. In fact, I'm quite satisfied (so far) with my DIY Cat-5 based cables. Yes, I MIGHT get some audible improvement with others, but when I play with vables, it is mostly just for fun and out of curiousity, and often I don't hear any difference at all [admittedly I'm not playing with $10,000 cables, though]. I know I can get better improvements elsewhere than from cables. One day when I get a $100,000-plus system that can reveal that there are three copper molecules out of alignment about 2.6 feet into the left high-end cable [Yes, I'm exaggerating], I'll look at my cables more carefully.
I'm off to put on my flack jacket now...
