Elon says it, so it must be true:
Humans must become cyborgs and develop a direct high-bandwidth connection with machines or risk irrelevance and obsolescence, says Tesla and SpaceX founder Elon Musk.
Musk's latest cheery thoughts were imparted at the World Government Summit in the UAE. "Over time I think we will probably see a closer merger of biological intelligence and digital intelligence," Musk said, according to CNBC.
The main thrust of Musk's argument seems to hinge on the limited bandwidth and processing power of a single human being. Computers can ingest, transfer, and process gigabytes of data per second, every second, forever. Meatbags, however, are severely limited by an input/output rate—talking, typing, listening—that's best measured in bits per second. Thus, to risk being replaced by a robot or artificial intelligence, we need to become machines.
(Score: 3, Insightful) by VLM on Wednesday February 15 2017, @07:37PM
Meatbags, however, are severely limited by an input/output rate—talking, typing, listening—that's best measured in bits per second
Probably more like kilobits/sec for the higher functioning.
I grew up in/around the BBS era and 300 bps modem is annoyingly slow, 1200 is slow but usable, 2400 is where I start to break down depending on text content. The advanced modems 9600/14.4/28.8/56 are well above what I can read.
If its something like a menu I know by heart or I'm merely looking at a portion of a command response I can keep up with a 9600 baud serial terminal hooked up as a console or when I was young hooked up to a VAX etc.
I am well aware I can outread 99% of the population and outtype maybe 95% of the population and I top out around 4.8K down and maybe 100 baud up. Doing ham radio stuff I can backstuff a PSK31 session quite easily so I know I can easily type a modest integer multiple of 31 bits/sec.
I would question if any of that bandwidth would be useful. Certainly a couple tens of megs of video bits per second is mostly just crap not worth anything. There's the concept of "goodput" and I suspect there is not a hell of a lot of data useful for people in a goodput sense, as demonstrated by social media.
Its interesting how little productivity gain there is over long terms of time. Sure my phone displays menus instantly but its all useless crap that doesn't matter, and it takes 10 minutes to slowly boot (must be the NSA monitoring software). In the old days at 300 baud logged into compuserve or early BBSes people didn't mind text appearing at a little faster than "TV scroll bar" speeds.
(Score: 2) by mhajicek on Thursday February 16 2017, @12:31AM
But can you type coherently at 2400 bps? Musk was primarily talking about output rates from the brain into a computer. As for whether that would be useful, in CADCAM work about 5% is figuring out what to do, the other 95% is telling the computer to do it.
The spacelike surfaces of time foliations can have a cusp at the surface of discontinuity. - P. Hajicek
(Score: 2) by TheLink on Thursday February 16 2017, @08:27AM
It would be interesting if it turns out you could get some people to do video output, stereoscopic 3d output or even 3D voxel output.
Then the very skilled and talented ones could do stuff like project their "daydreams" for people.
Probably have to start young while many brain cells haven't specialized yet. e.g. attach the I/O stuff to a suitable region and with some stimulation, training and practice some of the subjects end up with the I/O channels.
For the rest of us we can probably manage "thought macros". e.g. thinking of "a purple elephant" followed by "a yellow flower" (or whatever you've trained it on) unlocks "one-shot memory store" mode, then you and your wearable computer looks at the image you want to store (face, business card, etc) then think of stuff or a particular thought pattern you want to associate with that image and then think of the macro you've assigned for "Go!/OK" and voila your wearable computer will retrieve that image when you re-think of that thought pattern/macro.
Depending on the tech or situation you may actually be able to use the "natural" thought pattern you have when you see the image, or you use an artificial thought macro/mnemonic. For important stuff you might actually want to use a particular thought macro, whereas for stuff you just want to quickly store/recall for convenience you might use the "natural" thought pattern (assuming the tech can detect enough data points for such stuff).