A very different kind of seaplane.
I put "boats" in scare quotes because the idea would be to make the fuselage a lot less boat-like. Just some minimal creasing of the lower hull so that the floating hull could reach a speed of say 20-30 knots without too much drag--but no "step." Instead, you have retractable hydrofoils extending from the wings or lower hull (just like standard wheeled landing gear, but with longer, faired struts to submerge the foils well below wave troughs when the hull is just clearing wave crests.
The foils, I'm imagining, are quite dense (presumably made of stainless steel) rather low-aspect ratio elliptical planform wings which can be tilted to get various angles of attack.
Start by extending the foils while afloat (if starting from shore or in very shallow water, taxi out to deeper water with them retracted first!) and using the main engines (or perhaps an auxiliary engine running a submerged screw) "taxi" up toward foil "takeoff" speed. You might as well crank the foil angle of attack up to nearly stall angle, because every bit of lift you can get from them, even way below the speed where there is enough lift to get the fuselage out of the water completely, raises it at least somewhat, thus lowering the drag (and dynamic force) on the fuselage. It comes at the cost of hydrodynamic drag on the foils, but this is after all similar to what happens when the plane is taking off so presumably the engines can give the necessary thrust. At foil takeoff, the hull lifts out of the waves completely; now it is time to start lowering the AoA of the foils, which thus lowers the hydrodynamic drag, while the fuselage is now completely clear--and the air drag is still negligible compared to that at full (airborne) takeoff speeds. Now the acceleration increases as less and less thrust is consumed by the foil drag. Eventually you will pass the optimum lift speed for the foils and drag will start to increase again--bigger foils give earlier lift out of the water but more drag at high speeds, smaller ones mean longer delays to clear the fuselage but lower drag at high speeds. Meanwhile the airspeed is increasing; when well above stalling speed, you "toss" the nose of the plane up by smartly raising the AoA on the forward foils (you will have main foils under each wing and a small one on the tail for a "taildragger" type layout; I think this works better than a "tricycle" type with a nose foil and two main foils farther back, better mainly because of this maneuver and the reverse for landing). This raises the AoA of the wings, lifts the main foils quickly clear of the water, then you check the rising angle of the plane by in turn lifting the tail foil. Now you are airborne, still pretty low but well above the waves, and able to go for a higher lift coefficient (combination of high airplane angle of attack and flaps) than in higher flight because of the ground effect (same thing that makes ekranoplans possible). Accelerating just above the wave tips you reach speeds where high lift is easily achieved, off you go. Meanwhile you have been retracting the hydrofoils into wells on the wings or fuselage (wings much favored, you don't want draggy wells on the fuselage unless they can be totally clear of the water at all times.) They are foils so they won't be as draggy as wheels would be.
Landing (assuming you know you are on a safe approach--you know the bottom depth is greater than foil depth and there are no snags--or assuming you are taking some risk because landing without this preparation is that important...)--approach into the wind, slowing nearly to stall speed. Ease into ground effect, angle the plane up more--drag will slow you more but not as much due to ground effect. Lower the foils, then let the plane down enough to submerge the rear foil. This is tricky because this foil must slice cleanly into the water with minimum drag until it is deep enough to operate reliably, then suddenly raise the lift to bring the tail up and hence level the plane--which losing lift, drops its forward foils into the water in turn--these too have to enter just so. I think some sort of automatic control has to enable these insertions. Then again when reaching foil operation depth the front foils engage and stop the nose-downward pitch and descent of the plane--it is now gliding on the submerged foils, straight and level. As the plane slows you come down to the optimal minimum-drag foiling speed--maybe you throttle the engine to hold it there and taxi, maybe you want to slow down immediately so you let drag slow you more, raising the foil angles which further increases water drag, until these are nearly stalling. At that point you either stall them deliberately (which means a rather sharp brake and also the hull drops into the water, which means even more sudden high drag and possible damage to the hull) or hold them short of that angle and let the hull slowly sink as dynamic lift fades as speed bleeds off gradually. Soon you are a low-speed boat again.
Since you have the necessary forces available from the foils to both lift the fuselage out of the water before hydrodynamic forces will really start battering it, and to achieve a good range of wing AoA (some sources tell me the real purpose of the "step" on a standard flying boat design is mainly to allow the plane to rotate the nose upward for takeoff without it being shoved back down by forces on the rear of the hull) the fuselage can be much more like a standard landplane one. It needs some extra strength to be sure, and it needs to be corrosion-resistant and water-tight. The former will cost in weight and hence either fuel or payload, the latter might cost extra money.
The foil gear itself will also probably weigh somewhat more than wheel gear. But once airborne with the foils retracted, the plane is no more, or at least very little more, draggy than a landplane with its landing gear retracted.
Making this thing an amphibian is a matter of installing wheel gear alongside the water gear.
Drawbacks--well, an unseen underwater snag would be deadly at high foiling speeds (and probably a very bad thing at low speeds too, but that just cripples the airplane--it doesn't wreck it). But flying boats have similar hazards--they are at risk just putting down on any spot of water that looks OK from above. For safe operations one would have people at the destination scout out the water and clear any snags. For fancy military operations and the like, a 1960s or later version could carry a small drone plane--a UAV--to scout it out, landing in the prospective path and scanning it with sonar while the plane circles above pending the verdict.
I've estimated the foil sizes and lift/drag curves before, the main showstopper I think would be cavitation. Water is 800 times the density of air, therefore tremendous lift can be generated on very small areas at very low water speeds--the trouble is such high density lift corresponds to really low pressures, and when dynamic flow lowers water pressure enough, the water boils at the ambient temperature--this is cavitation. Cavitation involves bubbles whose noise of formation and combination and re-collapsing apparently packs quite a wallop, as it can pit propeller blades. Not to mention that water flashing to gas will play havoc with the flow streamlines on the foils and probably increase drag to murderous levels.
I suspect that there can be design workarounds--either using foils big enough to avoid cavitation (it also would help to submerge them deeper--but both these suggestions raise drag especially at high speeds) or perhaps designing the foil so that it is supposed to cavitate and the resulting flow pattern gives efficient lift, damn the pitting!
But I have never heard of anyone trying to use the kinds of foils I am talking about for water take-off and landing. I think I may have heard of someone trying the other kind of hydrofoil--which are basically V-or-U shaped long foils which gradually lift more or less out of the water as speed varies, and can't be controlled (or elegantly retracted) which would obviously only be any good for rather slow airplanes.
Anyway in an ATL without WWII and the massive investment in landplane infrastructure, I would think sooner or later someone would try this approach to watercraft, to try and get the best of both worlds.
When I raised this suggestion some years ago on an alternate aviation site (one mainly dedicated to LTA) someone there sniffed that we shouldn't even think of hydrofoils, we should just develop air cushions for landing/takeoff--that way we automatically have something that can work on either a paved runway or water. Well, the problems of an air-cushion landing/taxiing system for an airplane are not trivial to solve, apparently, from the total lack of any such systems actually being used on any type of airplane, no matter how obscure or experimental. I have downloaded some old documents (1970s) from USAF proposals to develop such systems, but never a verdict on why apparently no one has gone ahead and done so.
Frankly this might be a better idea than the foils--but then again, it was hardly practical before say 1960 (and apparently still not as of 2010!) whereas I'd think the hydrofoil thing could have been tried in the 30s if not in the 20s or even earlier.
Certainly it seems like a better idea than the hydroplaning skis tried on the Convair Sea Dart. These, essentially waterskis, created as one might expect severe pounding when the plane approached takeoff speeds or had just landed. I do wonder why they didn't consider submerging the planes for a smoother ride. Would cavitation be the reason they could not work and so no one has tried it?
Teh Google is my friend. Apparently someone has tried it, it worked, and then the US Navy decided it didn't want any more seaplanes around then, so forget about it...