Apparently the dry EEG machine used in that Force Trainer game I linked earlier work pretty well, though. Well enough at least that it can detect the difference between a focused, calm mind and a busy mind.
Not accurate enough to build an entire game around, unless that game is simply a mediation app, which is why I was suggesting it would only work if it were incorporated as low-hanging value-added additions to game.
Permadeth
I bought that for a friend of mine - It was novel, but didn’t work very well.
I think she ended up throwing it away because at a glance it looked like a bong.
These things are just noise machines, though. They’re a gimmick. You can take any sufficiently noisy data set and make it do something, but it will never achieve real control. A good analog is motion controllers like the Wii. Yes, swinging that controller is sorta reacted to by the game, and it sorta does what you wanted, but there’s very little timing, precision, or repeatability. The console is swimming in a river of useless noise from the accelerometer and trying to draw a very broad trend from it. The rest is the game erring hugely your favor so that it’s fun. Compared to a brain interface, accelerometer control systems are trivial. People way underestimate how hard this problem is because we’ve all seen tradeshow gimmicks like that that sorta seem to work. We assume it’s a short distance from there to an actual tight control interface.
This is like looking at a pop bottle rocket and feeling like the Saturn V should be a pretty straightforward from there. In fact, the bottle rocket is 50 years of development behind Saturn V. Again- rocketry is easy (seriously) compared to reading brain signals.
It seems to be part of a inexplicably pattern with Valve - get excited about something, do projects related to that, then drop them all before they go anywhere. It applies to all their hardware ventures - all their controllers, Steam machines, etc. (with the exception of VR, which probably has more to do with their partners) and probably most of their games in development, too.
It’s interesting that I’ve been seeing waves of interest, about every decade, around BCI and gaming, each time looking more impressive than the last, but still not going anywhere. (With a similar, 20 year cycle for VR.)
Yeah, it seems like he’s talking about using BCI to create high-resolution visual hallucinations to replace monitors? That’s, um, not exactly happening any time soon. I’m not even sure why that came up in the conversation.
The interesting thing I’ve been seeing lately is claims about some technologies that have been “20 years away” for the last 60+ years getting a new wrinkle, in that neural networks seem to have made some advancements such that it might finally be true. (Neural networks are pretty helpful here, with some aspects of BCI, but not most of what he’s hinting at.) I guess we’ll see in the next couple of decades, but in this case I’m not holding my breath.
The problem is that the actual input data being put out by existing sensor systems is so one-dimensional, it only works in addition to conventional control set-ups. Given how much work is required to get that one input, and how easily it’s replaced (e.g. by a pressure sensitive button or a game pad/stick, all of which are more reliable and useful), it can’t get beyond being a gimmick. Apparently recent game controllers had pressure sensitive buttons but they’ve now dropped them because so few games actually made use of them, so a more complicated version of that is a non-starter, really.
To be fair, most of those things already existed when they promised it would be available at some future date. (E.g. video calls, voice recognition existed in the '50s/'60s.) They either worked too poorly and/or were too expensive and/or lacked the social context they needed to take off. But it was plausible we would some day “have them” because well, we already did as things that worked (more or less) in the lab, just not as commercial products.
On the other hand, there are perpetually “20 years away” technologies that aren’t even working in the lab, now, and we don’t have a road map for even that. E.g. self-sustaining fusion, the kind of BCIs Gabe is talking about.
All of which they abandoned, too. (I assume the link isn’t being sold anymore, I know the controller isn’t, and the link isn’t exactly useful without it.) They made half-hearted efforts at these and then dropped them all.
The Steam Link hardware is discontinued, but weirdly, its software is still being updated. They’re focussing on the Android/SmartTV versions of it now, I believe.
(I’m a big fan of the Steam Link myself - especially for how cheap I got the device - and the Steam Controller is… well, good enough 
)
This topic was automatically closed after 5 days. New replies are no longer allowed.
