Your point is a very good one provided that the person using the mixer has years of experiance as well as a very well trained ear for audio.
A mixer would be very useful for some reasons:
1. You can apply in real time the EQ to your voice, without having to apply the EQ by software later to each file you record. Hearing in real time your EQ'd voice is useful to adjust your voice (how you speak) getting a better final sound.
If these setting are incorrect, you have now embedded them into your recording and if you wanted to change them later, you can't. You can set up most DAWs to achieve the same thing by setting up a track and sending the input signal to the track with reverb or eq so you can hear your voice, but the fx will not be in the recording. This will allow you to change/tweak the settings later on.
2. You hear your voice without delay.
Most AIs have 0 latency monitoring right from the unit and can use the ASIO file. As far as I know, most DAWs can also read the ASIO file which will provide the same results.
"
Audio Stream Input/Output (
ASIO) is a computer
sound card driver protocol for digital
audio specified by Steinberg, providing a low-latency and high fidelity interface between a software application and a computer's
sound card".
3. You can put a gate/compressor between mixer and computer.
Again, unless you have years of experience, this is a very bad idea. I can not tell you how many files I have been sent by someone who had a "Once In A Lifetime Recording" begging me to fix their file due to improper gating/compression. Once that signal runs through the hardware and into the recorder, it is final.
From famous interviews to recording recitals, improper gate/compression settings can destroy the delicate features that make the audio what it is. Once the audio/music/words have been chopped by the gate, it is impossible to repair. It also makes it very difficult to use a noise reduction filter, for the rest of the audio.
There is nothing worse then listening to crystal clear audio when there is silence, only to hear a bunch of background noise come in when the gate opens up. Not an argument, just pointing out some potential pitfalls concerning recording audio with pre set conditions.