IMHO, the intent is to limit the available development length for positive moment bars (bottom bars) beyond the point of inflection for continuous beams and beyond the center of support for simply supported beams. In other words, a #8 bar with an ld of 48" would not develope its full strength. It would be limited to the d of the beam or 12db, limiting the strength to 12"/48" or 1/4 of full strength.
My understanding is that there is a lack of data to support that long end anchorage length will be fully effective in developing a bar that only has a short length between the point of inflection and the point of maximum stress (ie - the point where tension in the bar is initiated and the point where it reaches its max stress).
But you can still use hooks or deformed bar anchors without satisfying equation 12-5.
The inflection point is where +M changes to -M. But this point might vary in location. If you use very large bars, and don't extend the bar far enough past the inflection point, there could be a case where load fluctuations move the inflection point over such that your tension bar is no longer developed into the concrete far enough and the bar lets go from the concrete, becoming incapable of resisting the tension.
Say you have several spans with constant positive moment reinforcing. I'm still not sure why a large bar size at the point of inflection would be a problem in this situation.
Well, if the reinforcement is constant across all the spans, then this section is satisfied since you have developed the bar waaaay past where it needs to be developed.
