Fast reactors are breeders.
The problem with a true breeder is proliferation, as it requires removing enriched fuel from the reactor. The concepts use a fertile blanket of material that needs to be replaced periodically.
Fast reactors are intrinsically ISO breeders, which create more fuel from the production of higher actinides e.g. plutonium, which can be burned in the reactor.
This has the downside that the enrichment of the reactor is constantly changing and changes the reactivity, this is part of what makes them harder to control.
Like any engineering decision it's a compromise, harder to control, but doesn't need to be refuelled for 20yrs.
But taking from something like LFTR, where they reprocess and remove fission products from the reactor, automatically during operation, will help with regulating the chemistry of the core and help with control.
The real advantage of fast reactors is they can be run on "spent fuel" from thermal BWR and PWR reactors with minimal reprocessing.
The problem is they're making the same mistake AEA made in the 60's, liquid metal cooling, they're either volatile (sodium, lithium etc) or expensive e.g. Lead + Bismuth. Molten Salts offer a better compromise IMO, but are also not without their risks.
All this is un-proven is the real issue and the major blocker being research investment.