The receiving IR remotes with an SDR by connecting an IR photodiode across the antenna input sounded like nonsense to me. I googled and could it be you meant people using the IR receiver feature of the RTLSDR's? It may still require connecting a photodiode, but to a different set of inputs.

https://medium.com/@rxseger/receiving-ir-signals-with-rtl-sd...

As to connecting a photodiode to the antenna input, I don't see how that would work, but that may well be due to my limited understanding and imagination.

Do you mean using the photodiode in a photovoltaic mode? Also, presumably you'd have to bypass the tuner and hook to the direct sampling pins on an RTLSDR? Even with direct sampling, wouldn't the 38kHz of IR remote modulation get filtered out by the DC blocking?

A photodiode (BPV23NF iirc) connected straight to the dongle's SMA connector. Yes, I believe it would be operating in photovoltaic mode, where the incident IR light from the remote control will induce a small voltage. Yes, I had direct sampling mode turned on (but the rtl-sdr.com V3 can do this through the normal antenna port). I pointed the remote at the sensor (admittedly quite close) and saw a signal centered on 38 kHz in the waterfall, and was able to export the binary pulses.

This video shows a Passive RF/Fiber connector that seems to work pretty reliably for broadcast tv.

https://www.youtube.com/watch?v=wIN7DVGBbKM

I imagine, provided the IR's frequency can be sampled by the SDR, it would look like fairly wide band bursts that could be decoded? Especially if you just treat the SDR as a ADC Oscilloscope