"By the year XXXX" means "before the year XXXX", right?

Despite the dictionary entries, there is no clear definition of whether "by 1959" means "before 1st January 1959" or "before 31st December 1959". The meaning is simply not intended to be that precise. (You will note that all the dictionary examples are talking about times, like 5pm, rather than years.)

I think that most people would take your first sentence as meaning "during 1959 discrete transistors were considered sufficiently reliable..." without implying whether it was precisely true at the start of the year. The phrase is intended to be a generalization anyway, so precise definitions are not going to work.

Your suggested change would make it clear that before 1959 discrete transistors were not sufficiently reliable, and in 1959 they were. The original version would be true if transistors had been considered sufficiently reliable in 1957, where your revised version would not.

The original wording ('by 1959, discrete transistors were considered sufficiently reliable') will normally be understood to imply (even though it does not say so explicitly) that there was a process of transistors becoming more and more accepted as more and more reliable; that process may have lasted several years. The OP's reading of the sentence misses that idea of a relatively lengthy process of growing acceptance.

At some point, the process crossed the threshold of their being 'considered sufficiently reliable'. Because the threshold is vague, the author of the article cannot vouch when precisely it was crossed, and has therefore chosen the wording that is compatible with its having been crossed in say, November 1958, or February 1958, or September 1957, etc., but definitely rules out its having been crossed in, say, 1960.

Is the meaning of that wording compatible with the threshold having been crossed in, say, April 1959? The wording is ambiguous in that respect, as has already been explained in DJClayworth's answer. By 1959 could be interpreted as by 1st January 1959 or as by 31st December 1959.

Now it is true that the sentence does not, strictly speaking say anything about what happened after 1959. It is logically possible that something was discovered in say, 1960, that led to them not being any more accepted as reliable. The idea of a lengthy process of their growing acceptance, however, strongly suggests (even though it, admittedly, does not logically entail) that the process was not suddenly reversed.

To see that more clearly, consider the fact that we often say to teenagers something like 'by the age of 18, you will be mature enough to . . .'. That sentence does not strictly speaking say anything about what happens after the age of 18. It does not logically entail that, at the age of 25 one will be mature enough for it. It is logically possible for the maturity to decline after the age of 20, so that at the age of 25, one is less mature than at the age of 18. But that logical possibility never crosses our mind when we utter or hear 'by the age of 18, you will be mature enough'; we all understand that sentence to conversationally implicate (even though it does not logically imply) that one will continue to be mature into the indefinite future. Analogously, almost everybody will understand the original sentence about transistors to conversationally implicate that they continued to be regarded as reliable.