Do the antennae protrude out of tiny little fairings/blisters? It is very hard to find an image without the booster stage.
No, there are no little visible fairings in this case, it is there, but it is really built in and not easily noticed. If you can find a clear image of the back of the missile sans booster you will see a small notch in the back ring, near #3 tail fin. That notch is the location of the antenna that looks back towards the beam source.
But remember the Talos is a pretty large missile and has a thick walled body to build things into. On smaller RF beam riding missiles there are typically blisters / radomes on the rear at some point, outboard of the exhaust cone and sometimes slightly forward. Some RF beam riders have put the antenna on the rear edge of the rear control / stabilization surfaces.
Was this from the outset or early on in the missile's development/operational life?
I have no idea if it was always conceived of as semi-active in the final phase of flight, but I think from very early in the design cycle this was a goal.
What function do the antennae serve? I was told it was some kind of interferometer...
Yes, they are an RF interferometer. They work in pairs to determine the direction of the received illumination energy from the target.
The illuminating RF reflects off the target, the RF interferometer determines the angle of arrival of that RF, and the autopilot steers the missile towards the reflection.
Roughly, the best accuracy you can achieve with a beam rider is the beam width at the target distance. Once the beam width at the target distance exceeds the kill radius of the warhead your kill probability goes down substantially.
Like making sure you kill the target you mean to kill and not something else.
Active seeker missiles tend to fall into two categories, lock on before launch, or lock on after launch.
Lock on before launch makes everything good. You can direct the missile to start tracking the desired target, queue the missile to the current target location, confirm that the missile has locked onto a target, and also confirm the missile is on the right target, all while the missile is still on the rail. Launch it, and let it go do its thing, fire-and-forget.
But missiles tend to have limited power sources in flight. That means the on-board radar is not very powerful. And they tend to be small in diameter, limiting the size of the antenna, meaning low antenna gain. Both of these factors (and others) mean the radar on an active seeker tends to have limited range and there might be other, less costly and less complex, techniques you can use at such ranges.
So the more common approach is to have the missile on-board active radar take over for the last phase of flight only, and provide some other form of guidance for the first parts of the flight to get the missile within its limited radar range. There are several popular techniques, such as command guidance, beam rider, inertial nav, etc, that have been used for this mid-course guidance.
But now the missile is out there on its own, way down range, and it has to search for and acquire a target to intercept. And if you do your design job right, it will acquire something. But is it the right something? Your original target will have moved by then, and you can point the missile at a predicted location, but is that prediction still good? One technique is to let the missile go inertial nav and provide periodic telemetry to the missile to an updated target position. But then, if you are going to do that and keep tracking the target all the way to intercept (or just short of it) why not use one of the more simple (design) techniques like command guidance or semi-active all the way home. One real benefit is at relatively long ranges, when techniques like command guidance fall down in accuracy, or where the target might be actually outside the range at which you can keep energy on him, like below the horizon (this does not have to be a low altitude target, at long ranges very high flying aircraft might be below your radar horizon), so you cannot illuminate for a semi-active seeker.
Don't get me wrong, there have been successful active seeker AA and SA missiles, but they tend to be less common than command guided or semi-active. AMRAAM and Phoenix are two real successes. Active seekers are more common in SS and AS applications.
But we seem to have gotten pretty far afield of anything related to the P-61
T!