(no April's fools joke included)
Back in the late 1930's the Royal Air Force's Fighter Command deemed it too risky to train its pilots in combat larger than 1 vs. 1, due to the risk of collision accidents. Its training accident rate was low, but the German Luftwaffe with its much less experienced senior pilots trained in battles of 4 vs. 4, or even squadron against squadron. It developed there and over Spain tactics for small units of fighters, as opposed to mere aerobatics for individual fighters. This made the Luftwaffe's fighter force a formidable threat to the Royal Air Force in 1940 even though its senior personnel had much less flying experience (there was no Luftwaffe before 1935, and no covert Luftwaffe before 1933).
Meanwhile, the Italians apparently focused more on individual aerobatics than any other air force (save for maybe the Japanese), and as a result badly neglected to push for higher practical and top speeds of its fighters.
The M1 Abrams main battle tank introduced in the early 1980's was technologically well ahead of its predecessor M60 Patton, but when new the crews still had to learn how to make use of its strengths. During the early exercises M1 Abrams-equipped units were actually less effective than M60 Patton-equipped ones. This is inconceivable to laymen, but actually a very common outcome and one of the reasons why hoping for "wonder weapons" is often foolish.
Back in WW2 it was considered best to train soldiers into infantrymen for six months, and pilots into fighter pilots with 400 flying hours. Due to urgent demand of forces in contact with the enemy many soldiers were sent to serve as infantrymen after six weeks of infantry training and many pilots were sent to serve as fighter pilots after about a hundred flying hours on a fighter. This contributed greatly to excessive casualty rates among the green infantrymen and rookie pilots. Less than ten per cent of fighter pilots scored more than half of the kills in the air, and it was even more extreme in the German and Japanese air forces where experienced pilots were never withdrawn (for good) to serve as trainers till the end of the war. The lopsided kill results were so extreme because most rookie pilots lost their plane on their first four contacts with the enemy, and all those inexperienced hostile pilots who weren't able to avoid getting shot down provided a large choice of easy targets to experienced pilots.
Infantry tended to accumulate the most losses in armies historically, with very few exceptions (steppe warfare with almost nothing but cavalry, occupation armies). Good training - not only long training - is pivotal to the infantryman's survivability. The German army of WW2 trained its soon-to-be infantrymen harshly, and even more so from 1942 onwards because of the harshness of the Eastern Front. A short-term result was a horrible accident rate. A newly set-up or refreshing infantry company would typically suffer one or two soldiers killed or permanently disabled during training until sent to the front (again). This looks horrible, but the life-saving effect on the front was much greater than a 1-2% mortality during training. The German army considered its infantry as qualitatively superior until about 1944, in great part due to superior (relative to opponents, not relative to what would have been possible!) training.
There are some lessons from this*, especially when you look at it from the background that there were short (approx. two years) but intense arms races prior to both World Wars:
Do not expect a relevant improvement of combat capability from a very new weapon system at a critical time. Consider recently re-equipped units as in training and not combat-ready, and keep enough combat-ready forces available at any time.
Be prepared to switch to more expensive (and especially more risky) training when you expect a major peer war in the near term with a high probability. Sweat and blood in training may save much, much more blood in battle.
Devise your training and force build-up plan such that enough time (and budget!) is given for training of combat troops. The intensity of training during a pre-war arms race is likely not sustainable during peacetime in general, whereas the advisable long duration of training may not be sustainable during an arms race or war.
Lesson #4 (drawing from lessons #2 and #3)
Armed services should be fully competent at three modes of operation: Peacetime routine, accelerated expansion in expectation of a war in the near term and wartime mode.
The legal framework (red tape, what's permissible, acceptable risks et cetera) and training regime need to know these three modes - with three different sets of regulations/manuals**. The planners need to have plans for all three modes, with transitions between the modes happening at unpredictable times.
*: I used anecdotes and didn't bother to look up the sources again, but you can find similar historical stories in many places and many times.
**: + a manual to point out the differences..