Depends on the kind of game, but in general I think lossless audio with high bitrate is good enough for most things.

I wfor one would love lossless audio in a audio games.  The sound would be worth the space loss for sure.  Experience the way the sound was ment to be.

Same as @5.
I have never been able to detect any difference, so I'd prefer a smaller size. Sounds are often the biggest part of a game.

Huh. I was not expecting to be one of the few "yeah, flac needs more love" responses. Even though I do like smaller file sizes, personally do not need hyper "realistic" simulation-quality content, etc, I also don't like destroying information. I kinda wish flac was much more common for that alone.
Also, license shenanigans.

For sounds/audio environments, I'd say lossless all the way. If people don't like it they can get a larger hd and more ram, sata-based ssd's are becoming more budget-friendly with the advent of pciE-based drives. AS for ram/cpu it's all based on the machine you get nowadays for in the case of laptops. The reason I bring this up now before someone can say, but my flee-market machine... is because just like the whole windows xp fude, there are going to be the folks that may object that either have their 4gb ram windows 7/10 machines, or still have their machine that got the free upgrade to windows 10, and I for one would not be able to do much comparison between lossy vs lossless audio handling unless I brought out my old Pentium t4300-based machine with an hdd. At least if audio games are to be potentially brought into 64bit land for more abstract audio environments (look at City of Flesh as an example of high-resource yet great audio handling, then we are going to have to start raising the bar for system requirements along with the video gaming industry (minus the gpu requirements of course).
Regarding audio handling, Three-D Velocity handled it in a bit of a hybrid though not a one-size-fits-all. All of the sounds of the game were in wave, and the long cutscenes were ogg. A game of the present could use a combination of wave and flac (wave for smaller audio/game sounds and flac for longer bits) or go flac-exclusively.
@CAEJones re: licenses, flac does not need royalties. Hell, Fraunhofer IIS stopped collecting royalties for mp3 in 2017, so the era of privately owned codecs, at least for your standard-issue audio, is definitely behind us.

@Zoren: With the NVidia gtx1650 with 4gb of dedicated memory shipping with some higher end laptops, that isn't a very tall order, as long as we can trust people to make the proper judgement call when buying machines. This dell xps15 7590 with 16gb of ram and the core i7 6core hyperthreading cpu was around $1700, and the mac population probably knows better than anyone that so-called entry-level macs are sold at close to the same price and don't have nearly as much going for them.
Regarding hardware, fortunately the software-hardware audio landscape is changing ofr the better. This machine comes with Waves Max audio pro and some good speakers, capable of running asio drivers and complex reaper setups, and virtual surround is becoming powerful enough that Waves can pump an Abby Road Studio studio monitor setup through a pair of headphones. So gpu-powered hrtf doesn't seem too far fetched.

Hello,
I don't have the good ear that is required to separate a lossless compressed audio from a lossy compressed audio file. So for me it does not matter at all, if there wouldn't be the quality loss truth I'd say lossy compression would be more benifitial in the way that it results in smaller audio files. Of course, the matter changes when it comes to unpacking those lossy compressed files to edit them.

@Iceguard: Hard drive space is not as much an issue, only for a small minority of people still on hdd's or t100 netbooks. Ram/cpu is more the issue, with a lot of people still using 4gb ram and duel core cpu's.

@16 true, but personally I would be willing to use my hardware resources for that. As for other people, that's up to them to decide

HRTF on the GPU is definitely possible already. The difficulty is writing the shaders necessary to do that. I never thought of using the GPU for audio processing... what a novel idea.
The main issue is the shaders. To send code to the GPU you have to give it a shader, which means you need to pull in a hole graphics library. That's not too bad of a dependency, since every OS nowadays has one. The shader, on the other hand, will be a tad more difficult. The language is quite limited i what it can do...
While it would be awesome you'd be putting the extra job of learning graphics programming onto audio game developers shoulders too, and graphics programming tutorials usually don't teach you how to do compute pipelines.

I'm all with jaybird here, we could do so much, with so little, with out sacrificing quality or storage. I'm also a fan of the retro things. Nice to hear that someone knows about that old calqulator!