I'm the first to admit that I am not fully versed in the subject but my understanding is...

Cutting off tension bars in a region that has high shear (ie..ends of a simply supported beam) creates a stress concentration that can lead to cracking at the cutoff point.

Wight & MacGregor have a discussion of this in their Reinforced Concrete Design textbook. (i think there are older versions on the internet)

They say you start with a calculated 'flexural cut-off point' where your moment drops below some level. This 'flexural cut-off point' is then modified to account for shear, development, and construction requirements to get the 'actual cut-off point'

The development length is the code defined length required in order to develop the bar for its intended tensile load. In the case of a single span simple beam with a uniformly distributed load, your reinforcement would need to extend this development distance (ld) on each side of the span’s centerline before you can begin to curtail the reinforcement (ACI 12.10.4). As an example, this would be the length of (4) bottom bars required to carry the maximum moment.

A cut off point is the location on a beam where a lesser amount of theoretical reinforcing is required. For the same simple span beam, this cut-off point would occur some distance away from the span’s centerline. The amount of steel at this location would be calculated per the reduced moment. As an example, this might be a location where you only need (2) bottom bars to carry the load as opposed to the (4) bars required at center span. Be aware that the code requires you to extend bars past this theoretical cut off point, typically the effective depth or 12 bar diameters (ACI 12.10.3).

You must satisfy the two conditions above, as well as maintaining minimum steel requirements, considerations given to shear at these cut off locations, the potential for additional stirrups... this can get quite involved for different beam configurations.

Note: I’m referencing ACI 318-11 in the sections above

Quote (OP)

If a bar is at a point where it's no longer needed, wouldn't it also need to be developed to this point as well?

That is correct. You'll notice that the development length requirement is for the bars that continue past the cutoff point rather than those that are terminated. ACI commentaries give the reasons for that. The bars that are cut off need to acquire bond strength at least as fast as tensile demand grows in the direction of increasing moment. That's easily satisfied under uniform loading however with code cutoff lengths however.

I like to debate structural engineering theory -- a lot. If I challenge you on something, know that I'm doing so because I respect your opinion enough to either change it or adopt it.