When we start learning something, naturally we want to do it well, or at least competently, or at least not so incompetently that people actually laugh. What's interesting though is that, to improve their various endeavours people employ all sorts of different strategies.
Some people will just try to get on with it and work it out. The old joke runs "I don't know if I can play the violin, I've never tried". It sounds like the sort of thing that should never work. But it can, up to a point. If you know roughly what you were aiming to do, then with enough trial and error you can indeed improve at most things. Even on the violin, you could sort of piece together a technique that might produce a functional sound. This is actually quite a popular approach in the various martial arts at the moment, working on the basis that "your body knows what to do". For me, that's nonsense. One look at footage of this sort of thing shows that what actually happens in these "make it up for yourself" moments is that you learn what to do in the course of a series of carefully designed exercises, and because you've learnt it in that way, you don't have the awkward readjustment to make between learning a neat little technique and applying it to a live situation. Generally then, without at least some guidance, this approach is likely to be quite inefficient.
The other end of the spectrum is to work from an accepted basic technique. We might look into what are the accepted techniques for writing a novel, or composing a sonnet, or hitting a cricket ball somewhere into an adjoining river. We might assiduously learn those techniques. We might then, as I suggested above, find that they have only a certain amount to do with the actual execution of the endeavour in any sort of real setting. Generally though, we learn enough to improve, maybe even to become very good. Certainly a lack of gaping technical flaws is unlikely to harm most things.
There are dangers though. One is becoming so obsessed with getting technique right that you lose everything else, whether it's lining up all your poetic feet neatly in an utterly passionless poem or keeping the left elbow nice and high in a forward defensive shot that nevertheless misses the ball. Another problem is that set techniques are simply what the people who designed them thought to be the best approach for getting consistent results from most people. It does not mean that what they are suggesting is automatically the best way of doing things, or that it is right for you. The Sri Lankan cricketer Ajantha Mendis has an incredibly unorthodox technique, but is extremely successful because of it. Had he been subjected to the "You won't succeed if you don't do it properly" approach so common in England, I doubt he would have been half as successful. Even worse, these techniques can date quickly. The sabre technique I was taught as an eight year-old has only a partial connection with what the best sabreurs are doing these days, while free-poetry is so much the norm among top poets that it makes a mockery of most people's basic training in it, which is usually in a more formal brand of the stuff.
Which brings us nicely to one last way of learning things: copy the people who are best at what you want to do. Again here, there are advantages and disadvantages. The proponents of the approach, most notably the NLP people, insist that modelling the techniques of the best practitioners of things is the only way of producing a high level of ability on a consistent basis. You certainly have a certain reassurance that what you're learning has already met with success. It's also quite fun in its way. Spending an afternoon doing impersonations of famous cricketers' bowling actions is far more fun than learning the basics by rote, while stealing good players' best licks has long been a guitarist's main way of improving.
The difficulties are slightly less obvious, but no less real. Firstly, you need to beware of the gap between what the best people say/think they are doing and what they are actually doing. Secondly, have they become the best because of or in spite of their technique? For yet another cricket based example (I assure you that this sort of thing applies to writing too, but this is what comes to mind) there is the case of Brian Lara's backlift.
Much like a golfer, cricket's batters have a backswing before they swing at the ball. The great West Indian batsman Brian Lara had a particularly exagerated backlift, but there are two schools of thought on its significance. One says that he was great partly because this backlift let him hit the ball harder. The other says "fair enough, but he was only able to do it because he was so talented. If you try it, you won't have time to hit the ball. Now go back to doing it properly".
This possibly connects into the last point, which is what you're capable of doing, both physically and mentally. It's obvious that I will never bowl like 6'4" 90mph fast bowler Steve Harmison. It's slightly less obvious that I will never be mentally set up to write in the same vein as Murakami, but it's just as true.
So what, after all this, is the answer? I suspect (fence sitting ahead) that the answer lies in doing a bit of all three. There's nothing wrong with learning a basic technique for things. There's also nothing wrong with learning from experience. If both of those happen to take you in the direction of what the best people are doing, so much the better.