I tried to avoid that comment, but now that it's been quoted, let's dig in, shall we? FCEVs are apparently only working on cost reduction, while BEVs still lack some capabilities that FCEVs and ICE cars have. So, what can we put under capabilities?
Luxury and performance? There are many more BEVs in the luxury category with many more bells and whistles available than FCEVs, and the highest-end BEVs rival or exceed ICE cars in most respects. The world's most advanced and versatile driver aids are in BEVs, the world's best gadgety interfaces are arguably in either a Tesla or high-end Cadillac, BEVs are with some distance the quickest cars in any segment and although they are not quite the fastest, for all intents and purposes there is no lack of performance in any metric relevant to car sales and general car usage, even in the cheapest offerings.
Versatility? Yes, BEVs are very much trailing ICE offerings, and have been for as long as they existed. There are a grand total of 2 BEVs right now that you can buy and that can tow anything, and nothing beyond ~3klbs. There are no people carriers worth talking about (I'm discounting the Evalia), there are no trucks, no semis, no haulers, in general most corners of the market have at most one or two serious offerings, with many lacking any. But in the FCEV market, this is even worse. There are currently 1 cars you can buy. Sorry, 1 CAR. Three you can lease, with a fourth one coming soon (Nexo), but phasing out another one (ix35). These cars stretch across a grand total of two market segments, neither particularly popular (and sedans getting much less popular as of late).
So what possible capability could we be talking about? FCEVs weigh the same or a bit more than BEVs, so it certainly doesn't have anything to do with that. Ah, of course, the only thing we could be talking about is the only USP of FCEVs: recharge time. FCEVs can do 0.7kg/min right now, which is an equivalent of about 45mi/min or 2700mi/hr. The best BEV on the market right now can only achieve 120kW or 500mi/hr. But wait, what is that, only 0.7kg/min? I thought a hydrogen refill only took '3 minutes', but 0.7kg/min means about 8min for a full refueling. In fact, average time spent at fueling stations excluding waiting time is well over 10 minutes according to the EPA's 2017 report. If this were all tech we figured out and only needed to make cheaper, why is there still this 'capability gap' in refueling time? Well, I'm of course being facetious: there are still some serious issues that have to be worked out in FCEV refueling. Nothing super hard, it's got to do with thermal shock, frost seizure, pressurization speed and that kind of stuff. But these issues put some really hard limits on what you can achieve with a hydrogen filling nozzle for the time being - a long time being. Meanwhile, electric energy transfer is not really bound by any physics like that. Sure, on a direct line without buffering you probably won't see any fast chargers exceeding the current 350/500kW CCS/Chademo standards (which, by the way, equal current hydrogen fueling speeds). But we will almost certainly be seeing >1MW water-cooled and >5MW superconducting chargers for heavy duty vehicles in the next 5-10 years. At that point you're talking about recharge speeds that can simply never be achieved with pressurized hydrogen at point of sale without some really big breakthroughs.
The next generation of BEV will literally equal or exceed the next FCEV in every respect, including maximum recharge speeds and certainly in range (which the high end already does). There will be literally no USP for hydrogen anymore. And let's be clear: this is all not a limitation of current battery or BEV-related tech, this is just infrastructure and compatibility.
But, to put some more nails in this coffin, there is almost no system in FCEVs that is not bound to a significant degree by lacking technical capabilities. The idea that this is just a cost issue is beyond ridiculous. For a reasonably-sized car, the total volume and in particular the shape of the fuel tanks is prohibitive. The Mirai already does some really funky suspension magic to fit their tanks, and it has zero breathing room in the car to fit anything else, certainly not something like a third tank. This means for all intents and purposes, 700MPa tanks are limited to around 5kg or 300mi worth of hydrogen in perpetuity. The next generation of tanks, going up to 1050MPa, is already about 5 years overdue and according to the 2017 report there has no progress - barely a mention even. Even so, this will at best get us up to about 400mi of real range on about 140kg of tanks and piping, plus about 200kg of fuel cell and cooling. This gets us to a net energy density compared to batteries of about 1mi/kg, worse than bare lithium ion cells and only slightly better than complete battery modules. Considering the tank shape and strength requirements prohibit much in the way of progress here, you completely lose every advantage of hydrogen as far as energy density goes, but on top of that you have no hope of improving over time. Whereas batteries still have a lot of room to improve, with e.g. 600mi BEVs on the roadmap for 2021-2022 (yeah, the Roadster is supposed to launch in 2020 but nobody believes Elon time).
FCEVs are currently reliant on low-temperature fuel cell stacks, which have some inherent efficienty disadvantages. This means they need a lot of cooling, which together with the fuel tank shape limits to a large degree how much form freedom designers have, and what kinds of vehicles you can expect to build. Batteries on the other hand can be placed anywhere, in any shape pretty much.
PEM fuel cells, especially if we ever expect to go off PGM catalysts and need to go towards lower activity nickel-based catalysts, have inherent ramp-up/down times proportional to the time it takes for solvated hydrogen to move across the electrodes. Either that, or you need to permanently waste a lot of hydrogen to keep the electrodes saturated. This limits responsiveness of the car and makes the cars much less appealing to drive compared to BEVs to anyone appreciative of a quick car. BEVs certainly sell on responsiveness and quickness compared to ICE cars. Of course, this is why I 100% expected a plug-in hydrogen hybrid, because then you suddenly do have a big traction battery to do regen and hard pulls on.
Fuel cells have significantly limited lifespans. Much less than most batteries, unless you drive a Leaf
Typical lifespan right now is in the low thousands of hours, and it is the single most expensive part of your car to replace. Let's be clear here; replacing a 64kWh battery will probably cost you $10k, replacing a 100kW fuel cell stack is closer to $50k. Battery prices are going down much faster than fuel cells, so this is not a good prospect for second hand buyers. Same goes for the fuel tanks by the way, they have to be recertified and possibly replaced after 5, 7 or 10 years depending on your jurisdiction.
None of these issues are just 'cost reduction' problems; they are straight up capability issues. We haven't even touched the infrastructure issues either. All-in, FCEVs are really not ready for primetime yet, and by the time they are they will certainly have been surpassed by every BEV on the market.