MP3 - lossy, universal, finished
MP3 permanently discards audio information the encoder judges you will not miss. At 192 kbps and above most people cannot hear the difference on speech, which is why it dominates podcasting.
The important consequence: every time you re-encode an MP3, you lose more. Editing an MP3 and exporting it again is a second generation of loss. Treat MP3 as a delivery format, never as a working file.
WAV - uncompressed, huge, lossless
WAV stores raw PCM samples with no compression at all. A stereo minute at CD quality is roughly 10 MB, so an hour-long recording lands near 600 MB.
It is the right format while you are recording and editing, because there is no encoding step to degrade anything and no decoding cost when scrubbing a timeline.
FLAC - lossless, compressed, archival
FLAC compresses losslessly, typically to 50–60% of the WAV size. Decode it and you get back the exact original samples, bit for bit.
That makes it the correct archive format: full quality, roughly half the storage. The trade-off is that fewer consumer devices and upload forms accept it than MP3.
What about AAC and M4A?
AAC is lossy like MP3 but more efficient - noticeably better at the same bitrate, particularly below 128 kbps. M4A is usually just AAC in an MP4-style container, which is what Apple Voice Memos and most podcast exports produce.
If you are choosing a lossy delivery format today and the destination accepts it, AAC beats MP3 on quality per kilobyte. MP3 still wins on the sheer number of things that will open it.
The short version
Record and edit in WAV. Archive the master in FLAC. Publish in MP3 or AAC. Never edit and re-export a lossy file if the original still exists.