Can two people see the same photon?

Seeing = detecting photons that happen to interact with your retina.

You can't see photons when they are just travelling nearby. Take lasers for example. When someone is using laser pointer, the only reason you see the beam is that photons collide with dust and air particles and therefore their direction is changed. For example into your eye. Otherwise you wouldn't see anything.

It isn't possible for two people to see the same photon.

In theory, in the most perversely contrieved case, and if you are willing to cheat a bit, it would be possible. In any half-reasonable, realistic setting, the answer is a clear, definite "No". Indeed, people cannot even see single photons at all (contrary to urban myths).

How does seeing a photon work? The photon has to hit your eye, specifically one of the billion rhodopsin molecules in one of the several-million retinal cells, then something-something, and then a nerve impulse maybe, if some conditions hold goes through the roughly-one-million ganglion network in the retina, and maybe makes it to the brain. Maybe. And maybe the visual cortex makes something of it.
The "maybe" part and the fact that a single cell has billions of G-proteins going active and inactive every second, and that there's a continuous flow of cGMP up and down is the reason why you cannot really see a single photon. That just isn't reasonably possible, if anything it's placebo effect or mere suggestion.

So what's that something-something mentioned previously? The photon flips the cis-bond at position 11 in retinal to trans. Which, well, takes energy, and absorbs the photon. This triggers a typical G-protein cascade, with alpha subunit going off and blah blah, resulting in production of cGMP at the end. If the cGMP concentration goes above some threshold, and if the cell isn't currently refractive, then the cell fires an AP. That's a big "maybe". Then comes something-something ganglion cells, which is the other big "maybe" part above.

The photon is "gone" after that. No second person could possibly see it.

Now of course, no absorption is perfect, there's an absorption maximum for each type of rhodopsin, and even at that it isn't 100%. Outside the maximum, the absorption is far from 100%. Which means that the photon is emitted again, and it could, in theory, in the most improbable case, hit another person's eye, why not. But of course we have to cheat a bit here because it strictly isn't the same photon.
Unless we are willing to cheat, the answer must therefore be "not possible".

Candles do not give off single photons. Preparing light sources that can emit single photons is tricky.

The photon contains "one photon" (some small quantity of electronvolts) of energy. The energy in a photon is directly propotional to its frequency, so two photons of the same "color" have the same energy. The process of absorbing a photon transduces "one photon" of energy from the electromagnetic field to the detector. Consequently, if either human detects the photon, there is no energy left to be detected by the other human.

In "Direct detection of a single photon by humans", J.N. Tinsley et al. directly measure the event of conscious detection of single photons. Subjects in that experiment

• did (barely) better than chance (51.6% (p=0.0545)) correctly identifying photon present and photon absent events) when observer confidence in event was excluded and
• did better than chance (60.0% (p=0.001)) when confidence was included.

Interestingly, "the probability of correctly reporting a single photon is highly enhanced by the presence of an earlier photon within ∼5 s time interval. Averaging across all trials that had a preceding detection within a 10-s time window, the probability of correct response was found to be 0.56±0.03 (P=0.02)."

Of course, not every photon that strikes the retina is transduced. "Based on the efficiency of the signal arm and the visual system, we estimate that in ∼6% of all post-selected events an actual light-induced signal was generated ..." So we expect to see improvements over random chance in the neighborhood of 6%, and all numbers reported above are in that neighborhood.