One advantage here is that you're definitely getting into some decent peptide concentration without a carrier. If we assume 200pg/cell x 28 cells you're already at 5.8ng of material (at 100% recovery) and just about any instrument can work with that.
My first obvious concern is -- OMG, the data processing must be miserable -- but as someone who is commonly combining hundreds of multiplexed TMT with randomized samples....this is a point where I've got zero room to talk.
They do see advantages from the carrier channels in terms of # of differential proteins between conditions, but the numbers are impressive in both cases. For those of us who aren't reducing and alkylating prior to TMT labeling single cells, the methods here for adding those steps are a welcome surprise.