Discussion Topic |
|
This thread has been locked |
healyje
Trad climber
Portland, Oregon
|
|
|
Feb 26, 2015 - 02:03am PT
|
Lordy, lots of optimism over scifi 'AI' and computationalism in general. Not going to happen.
And AI isn't dead, it just figured out it had to make a living like everyone else. It specialized even more than it was on various problem spaces and broke out into optimization/operations research, expert systems, search/pattern recognition, inference engines, various forms of machine learning, etc., etc.
Plenty of it going on all over today. But as far as computation-based consciousness or sentience? Again, ain't going to happen. 'Smart machines' are and will happen; sentient ones are not and will not.
|
|
Tvash
climber
Seattle
|
|
|
Feb 26, 2015 - 06:40am PT
|
uhoh. Rev. Healy's prophesizing again.
Move over Largo. All hands stand by for Im So Expert I Can Predict The Future For All Time.
Such predictions are aided by updating one's awareness of what's happening in AI from, say, the mid 90s, however.
|
|
WBraun
climber
|
|
|
Feb 26, 2015 - 07:30am PT
|
'Smart machines' are and will happen; sentient ones are not and will not.
YES
|
|
Tvash
climber
Seattle
|
|
|
Feb 26, 2015 - 08:06am PT
|
Such predictions reveal a poor understanding of development arch for AI, what's driving it, or new technology in general for that matter.
Development of larger integrated systems suddenly accelerates when enough of its subcomponents are commoditized and their interfaces standardized. Designers and researchers from all over the world can pull such subsystems off the shelf, focus on the hard problems, collaborate, and innovate.
One of the holy grails of AI has been to reduce life's 'friction' - going to the bank, parking, searching for information. AI attempts to satisfy evolved human needs in an efficient manner.
If we include the need to be loved, understood and cared for, something our fellow humans are apparently really, really bad at, then a sentient artificial being that 'knows you better than you know yourself' (as Google already does) would have enormous value, particularly in a rapidly aging population like ours.
This will likely morph into artificial life companions. We already know that people will readily 'love' a machine. Or any inanimate object, for that matter.
Make love to a machine, you say?
A huge percentage of American women already do on a regular basis - and not one that will also cook dinner and clean up afterwards while engaging in the perfect brand of witty banter.
Yet.
The market for autonomous machines is gi-f*#king-gantic, as anyone in the industry will tell you. Robot cars (and the end of roadway expansion). Warbots. Hazardous/remote environment exploration/surveying. Amazon's competitive need to break the 48 hour delivery barrier. The fully burdened cost of labor.
Healy's view smacks of the early 90s, when the hangover of what robots COULDN'T do came home to roost in American industry as a result of technological limitations and resultant high system maintenance/capital costs that never quite made up for the labor savings - just as offshoring was becoming easier and more common. I designed automated assembly lines at the time and saw this first hand.
To state none of this will happen in 20 years is probably safe. But 100 years? 500 years? How far into the future does one's ego need to stretch before one is forced to admit to to the same level of universal ignorance about what the future holds we're all doomed to share?
|
|
Ed Hartouni
Trad climber
Livermore, CA
|
|
|
Feb 26, 2015 - 08:12am PT
|
ever watch the system message console of your machine?
looks like the "monkey mind" to me... a discursive rap on what the state of the machine is, in realtime, as multiple processes execute and vie for the attention of the "central control" which decides what to do...
but then, I am a materialist after all, and susceptible to the fantasy of a material explanation of phenomena... as opposed to the fantasy that "something else" (or should that be "no-thing else") runs the show.
If "no-thing else" does run the show, it's hard to imagine why we would discuss any of this for so long. Perhaps those making the appeal to that explanation should be happy and just not try to describe it... the only thing they have is to criticize those who believe otherwise. They literally have "no-thing" to put up as a counter argument.
Anything is possible if "no-thing" is the answer.
|
|
Tvash
climber
Seattle
|
|
|
Feb 26, 2015 - 08:22am PT
|
Finally, the prediction that 'it can't be done' is already wrong.
Nature has already done it.
In two very different ways, even.
But humans aren't very good at studying, understanding, mimicking, and innovating on nature after all. No - we just don't go there.
|
|
Tvash
climber
Seattle
|
|
|
Feb 26, 2015 - 08:26am PT
|
Oh, and eat this, Healy:
[Click to View YouTube Video]
Who wouldn't want to share a romantic dinner with that thing?
|
|
High Fructose Corn Spirit
Gym climber
|
|
|
Feb 26, 2015 - 08:28am PT
|
Thanks, MikeL, for the reply.
I was thinking Cosmos might give us more common ground for discussion. No worries, will just have to find other means, eh?
PS As an avid squatter back in the day, I also varied my sets in my workouts between low rep times and high rep times. I enjoyed mixing it up and valued the blend feeling I got benefits from it across several sports. Knees still in great shape, kow.
Maybe I'll do a few sets today. Thanks. :)
.....
"But humans aren't very good at studying, understanding, mimicking, and innovating on nature after all. No - we just don't go there."
You forgot the irony/sarcasm font, bro.
|
|
healyje
Trad climber
Portland, Oregon
|
|
|
Feb 26, 2015 - 08:37am PT
|
Such predictions are aided by updating one's awareness of what's happening in AI from, say, the mid 90s, however.
You should reel back from, say, 2090 and too much scifi. Also consider ripping the Kurweil singularity drip from your arm. If you were current on what's going on in the space you wouldn't say such ridiculous stuff.
|
|
Tvash
climber
Seattle
|
|
|
Feb 26, 2015 - 08:39am PT
|
Is there an echo in here?
Anyway, Joe, I know yeez well enough to know there's no prying the Joecentric universe from yer stubborn clutches.
You know the future. All I ask is that you wield that awesome power for good.
If you only knew the present better - but we've got Ed for that with his pesky data fetish.
Ketchup, anyone?
|
|
healyje
Trad climber
Portland, Oregon
|
|
|
Feb 26, 2015 - 08:56am PT
|
Ah, but I don't know the future. But I've been in the business for thirty years so I know the tech, know the biology, the trending in the field - hence know the past. Knowing the past gives me some pretty good insight into the rate of change, what's possible relative to both the tech and the biology, and leaves me of the strong opinion computational sentience is not in the cards even though all kinds of smart/intelligent software and machines will be. At our current rate of change - which is fairly accelerated - we will be overcome by pandemics or an asteroid several times over before getting anywhere near sentient machines.
|
|
Tvash
climber
Seattle
|
|
|
Feb 26, 2015 - 09:15am PT
|
In exactly what capacity have you been 'in the business'? And what business, exactly?
I spent 18 years in new product development - from design of robot assembly and test cells for high volume manufacturing to new product design (first as a mechanical engineer, and finally managing entire product development effort - all for human-worn or handheld product applications), to managing large software development programs for startups in a variety of companies from big (ATT) to 20 person startups.
That doesn't mean I can predict the future, however.
How about you?
You'll forgive my pals in new product startup environments and genetic research if they are somewhat more sanguine about the future of technology, even as they recognize the challenges. It's probably the typical forest/tree thing so common among the overspecialized/no-longer-current.
|
|
MH2
Boulder climber
Andy Cairns
|
|
|
Feb 26, 2015 - 09:40am PT
|
But as far as computation-based consciousness or sentience? Again, ain't going to happen. 'Smart machines' are and will happen; sentient ones are not and will not.
Despite the emphatic statement, healyje does not say WHY it cannot happen.
Back in a day when I followed computer science and neuroscience more closely, no responsible investigator would make general-purpose intelligence or consciousness a goal to be implemented in a machine or understood in a theoretical sense. That destination is too ill-defined. How would you be able to show that you were even headed in the right direction?
However, curiosity about intelligence and consciousness has led some otherwise sensible people to devote time to them, usually in the twilight of their careers, like Sir Francis Crick, Roger Penrose, and my hero Rodolfo Llinás. There is a fascination which younger stronger minds are better able to ignore.
Who would want a sentient machine badly enough to spend money on it? In the present day it looks to me like the entertainment industry might be a candidate. A lot of work and money has gone into CGI and animation. In addition to generating realistic facial expressions, it might be useful to generate realistic emotions and dialogue, too.
Yes, it is an sf speculation already treated in that realm. I like Terry Bison's Voyage to the Red Planet.
|
|
healyje
Trad climber
Portland, Oregon
|
|
|
Feb 26, 2015 - 09:52am PT
|
You'll forgive my pals in new product startup environments and genetic research if they are somewhat more sanguine about the future of technology...
They're sanquine because they are currently funded. That funding in the space is currently driven in part by thinking similar to yours and by VC investors who similarly drink Kurwweil's daily singularity pablum. Unfortunately for them, however sanquine, the vast majority of these startups fail (and biotech startup funding is currently down year-over-year with fewer VC involved). Tech startups only slightly better.
Show me one company currently on the AI AngelList that's even remotely connected to your proposition...
|
|
Tvash
climber
Seattle
|
|
|
Feb 26, 2015 - 09:52am PT
|
Who would want a personal digital expert companion?
Everyone with smartphone, apparently - about 2 billion peeps so far.
People clearly prefer their smartphone, even as dumb as they currently are, to real people. Just look around.
Hierarchical systems are developed in the much same way organisms evolve - from the bottom up, sub system by sub system.
How long did it take nature to figure out the cell level chemistry before multiple cell organisms could evolve? 3 billion years? Multicellular has only been around a sixth of that.
We're in that stage with artificial sentience now. Once there is a robust enough library of subsystems readily available - development of complete systems will likely accelerate.
Do we need a complete theory of consciousness to create it artificially?
Nope.
You want some key AI funders today?
How about Apple, Google, Amazon, and the DOD?
Yup, a lot of small startups fail - but the cream of the crop get bought. NEWSFLASH: VCs actually plan for this through robust risk assessment, diversification, and more conservative rounds of funding tied to performance. Tech is also cheaper to develop than it was during the bubble. Way cheaper.
The VC industry has adapted to the failure of the bubble and is currently growing rapidly worldwide.
But that's not the only way new tech gets funded of course, as i just mentioned.
That, and 'The Singularity' is not machine sentience. Just an FYI.
|
|
WBraun
climber
|
|
|
Feb 26, 2015 - 10:15am PT
|
artificial sentience
That means it's not real.
The real ingredient is missing, and since you do not know the real ingredient you make stupid claims.
Thus you're ultimately playing with matches and have no credit except to manipulate the gross material elements and energy.
But can't create them as they are already there to begin with.
Thus you can't create sentience either since its already there to begin with.
Thus you're ultimately a poseur claiming a post dated check in the future and only giving gross technology .....
|
|
Tvash
climber
Seattle
|
|
|
Feb 26, 2015 - 10:18am PT
|
Or just dating a(n artificial) chick in the future, more accurately.
Could be an option for ya, Werner. In my kinder, gentler imagined future, even the most outlier personalities get some robot love.
Until they turn on us, of course. You still gotta be nice.
|
|
MikeL
Social climber
Seattle, WA
|
|
|
Feb 26, 2015 - 11:41am PT
|
Ward:
That was a hilarious post. I cried.
HFCS:
Yeah, we’ll find another topic to get together on.
|
|
rbord
Boulder climber
atlanta
|
|
|
Feb 26, 2015 - 11:48am PT
|
Scientists discover black hole 12 billion times bigger than the sun! How does that affect our belief systems? Are we all still masters of consciousness and truth? I won't hold my breath waiting, in the face of my unwavering certainty, for the discovery of the next 12 billion fold increase in Information, or the next 12 billion fold increase after that. I've learned to believe that what I have is enough to answer life's mysteries.
|
|
|
SuperTopo on the Web
|
Let us know!