It hasn't really been an issue for me so far. Things usually seem back to normal within 24 hours, at least, I never noticed a big drop that I could blame on code change.

Switching from the standard code to asynchronous and later to responsive didn't slow me down one bit. Never noticed any real downside to switching either. Upside is, I started to make money on devices that weren't really producing before.

Things usually seem back to normal within 24 hours

In the past, as long as I did not change channels and placement (which I did not this time), this was true for me as well. This was the first time it took time to settle without changing channels or placement.

netmeg & webcentric, but how about other kinds of changes? How long do you wait to see the effects of say a placement change or a channel change?

A couple weeks, but it really depends on the time of year as my sites are heavily seasonal. But honestly, the channel changes seem to pop right back too, and I haven't changed any placements for a couple years - the last time I did was because AdSense predicted a significant revenue increase if I moved an ad slightly and went to a bigger size - and not only were they right, but it improved within two hours.

I regularly move ads around (well, not every day but often enough to see what location, color or ad size has to do with anything). I can usually tell in a day or so if my changes were effective or not. On the other hand, because of fluctuations in the quality of ads being served at any given time, I find that it doesn't pay to make these kind of changes more than once every couple of weeks or so. In other words, I'm giving things a couple of weeks in case Adsense is generally in a rut or is doing something out of the norm that might skew my perception of the results. There's no such thing as a normal Adsense day to me and so it pays to take a larger sample when calculating the effectiveness of any move.

Added: I'm also finding the "month" to be a good period of time when making comparisons with historical data e.g. this January vs last January. This month vs last month is not necessarily the best way to judge sometimes because each month of the year really represents a different frame of mind for advertisers and for consumers.

General rule of thumb, give it 2-4 weeks. Compare to LAST year, not last week. Take into account traffic differences. Some changes take time. Saturday's traffic may be different from Tuesday's traffic.

P.S. I use that example because Saturday is my slowest day, and Tuesday is my busiest.