do you have to reveal the algorithms you come up with or is a working prototype sufficient ? how detailed do those feasability reports you have to submit back to them have to be ?
Built one for some clients a while ago. clients were dot commies so went bust...and i dont have anything useful to do with it anymore.
its totally modular and does everything. built in Java+HTML (java is totally separate from the HTML) and connects to any database/OS known to man. I can prolly customize it for you. not open source but you own the hardware+software and its plug n play. no admin required (has watchdog timer, automated backups + bandwidth throttle/selective service Q) and secure (AES 256 bit/RSA 2048 bit). Its been deployed on lotsa hardware including clusters and tested upto 200K transactions/min. If interested let me know. it'll cost you around $5K/site cos i'll have to customize it some depending on what features you want on each site. i can handle most of the credit card setup and help you in the process..ive done it dozens of times anyway. I can also ship you preinstalled versions on custom hardware. zurk42 at hushmail dot com
its very simple. LCD or lowest common denominator. intelligence is indicated when LCD is NOT the path chosen to solve a particular problem AND the problem is solved anyway. natural selection always chooses the lowest common denominator or shortest path to solve a particular problem set. Intelligence does not. intelligence definitely requires either a brain or some centre of processing analogous to a brain. A CPU or turing machine will do if it can emulate a brain like process. Consider a parrot which wants a cracker. Everytime it wants a cracker is sez "Polly wants a cracker" at which time a human feeds it a cracker as a reward. Natural selection gave the parrot wings to fly and eat berries from trees. That was a lowest common denominator solution i.e. berry on tree -> parrot must eat -> give parrot wings. However the parrot is also intelligent and understands that crackers dont grow on trees and there is NO way it can get a cracker without pleasing its human keeper. So in order to get a cracker it must please its human keeper and to do that it must reproduce a set of sounds accurately in order to obtain the cracker. so intelligence is : cracker with human -> learn sound to please human -> reproduce sound when human in room with cracker -> parrot gets cracker. while natural selection is : cracker with human -> fly to human and attack human to get cracker -> human too large -> get some other food item. crude analogy but i think i made my point.
just out of curiosity...has anyone done a realistic model of a single neuron ? or at least a complete emulation of a single neuron in software ? that includes all the growth and electrical/chemical effects ? i check minduploading.org (MURG list at http://minduploading.org/mailman/listinfo/murg) but they dont seem to have accurate models.
and now consider this -- youve sold yourself to the highest bidder and gone and done something that compromized your ethics for what ? a job that pays 40-50K/yr ? its a dog eat dog world but compromising your ethics really isnt worth it. you can always get another job -- hell construction/trucking/sales positions are always available. and in virtually all other jobs you dont have to compromize your ethics. In case anyones wondering, yes ive had this situation before where i had to fudge some benchmarks. i said no, and i expected to get fired (which i did of course). what did i learn from this ? [a] theres other jobs out there. [b] being someones whore isnt worth it. [c] i still have my self respect (and a lot more respect from my ex-co-workers).
should be fairly easy to do -- heck i could do it. the only problem is finding someone with $4 mil to invest in a project like this as a test case. it wont take $20 mil unless you include airconditioning, the HALON fire suppression system and building leases.
dual processor linux workstation ? no problem -- heres the specs. 1 x Asus A7M266-D socket A motherboard. 2 x Socket A AMD MP processors (1900+ should be ok) 4 x 184-pin DDR Crucial PC2100 1024MB RAM DIMMs (4 gig RAM -- max it out) 1 x Matrox Marvel G400-TV video card (ok not the best for gaming but it does lotsa video stuff and is well supported in linux). 2 x Fujitsu MPG3409AH 40.9GB HDDs (mirror/RAID em via software IDE RAID) 1 x Pioneer IDE DVD-R drive for backups and reading DVDs and CDs. 2 x 3COM 3c905B Ethernet controller. Any decent LCD monitor or CRT monitor-- choose your poison. Its not the fastest system but it'll do what you want. For frills you can also add: 1 x matrix orbital LCD with floppy drive plate 1 x Yeong Yang 0221XB Black Cube Server Case with redundant 400W PSUs 1 x Axis network camera for video conferencing 1 x 802.11b wireless PCI adapter (make sure its got a prism 2 chipset) Try this config. i've done it before and it works. its not a screaming bleeding edge game machine..ive sacrificed speed for reliability. ive also made it more pricier for more reliable components. shouldnt burn a hole in yer pocket tho. i'd buy something like this for my next system but i have my eye on a dell inspiron 8200 as soon as i get reasonable employment. bleh. recessions suck.
Palm did the exact same thing initially -- they gave away the OS (ROM dump of several versions of debug OS), devtools, emulator etc etc. They still do if you sign the NDA and fax it to em. They sold the hardware with the non debug OS.
do the same. give away a DEBUG version of your tools which cant be used in production and SELL the complete non debug production version. just dont make the debug version an eval or time limited. let it print gobs of debug messages and make it bleeding edge.
these guys typically need a version control system or equivalent to do the job. with more than a dozen ppl chucking everything into a single directory makes a huge mess. i've actually been writing my own CMS with version control/web based GUI editor etc etc and it looks pretty good. if anyone is interested email me for a license. its still beta but it should do the job.
are you stupid ? have you looked at monster.ca's job forums ? Canada has a nearly 65% unemployment rate for IT...the highest in north america. moving here ? youve got to be kidding.
uuh..maybe because in order to decrypt you have to solve the problem ? a mathematical knot cant be solved so you can never decrypt the information in the first place...rendering the entire scheme futile.
and security by obscurity never works. we already do padding.
hmm...maybe. we dont know if intelligence and entropy *are* linked only when considering intelligent organisms tho. how do you prove its not linked when considering an intelligent organism ?
i dont think autopilot.sf.net can be used with one of these to fly. it seems to be limited to normal RC heli models not this weird 4 propeller gyro stabilised carbon fibre flying machine.
The motion an handling characteristics of a 4 propeller machine v/s a single (large) bladed chopper are way different.
how is http://nathandial.com/mirror/globalenv/ better from a research PoV ? real research pushes the boundaries -- if we can somehow find a useful solution to eliminate bugs its prolly going to be non obvious so jumping on weird tehcniques like this may be a valid solution to eliminate bugs in the near future. it doesnt seem practical now of course but what if they could get bugs to soun different from the main portions of the program in some way ?
the main challenge is being able to autonomously control the heli indoors..i.e. avoid hitting walls with rototr blades etc. follow the horizon autopilots have been done to death in most of the UAV competitions. object tracking/scene recognition is the real challenge.
dude. next time post under your own logged in name when posting such a godamn incoherent post. autocoding aint going anywhere. not anytime soon anyway. and it has nothing to do with.NYET
computers cant program themselves either. if they could solve the deep theorems as you prove it they would be able to program themselves pretty easily -- all computer programs are essentially deep theorems.
Slashdot Mirror
do you have to reveal the algorithms you come up with or is a working prototype sufficient ?
how detailed do those feasability reports you have to submit back to them have to be ?
now THATS cool. anyone done DOOM in javascript ?
Built one for some clients a while ago.
clients were dot commies so went bust...and i dont have anything useful to do with it anymore.
its totally modular and does everything. built in Java+HTML (java is totally separate from the HTML) and connects to any database/OS known to man. I can prolly customize it for you. not open source but you own the hardware+software and its plug n play. no admin required (has watchdog timer, automated backups + bandwidth throttle/selective service Q) and secure (AES 256 bit/RSA 2048 bit).
Its been deployed on lotsa hardware including clusters and tested upto 200K transactions/min.
If interested let me know. it'll cost you around $5K/site cos i'll have to customize it some depending on what features you want on each site.
i can handle most of the credit card setup and help you in the process..ive done it dozens of times anyway. I can also ship you preinstalled versions on custom hardware. zurk42 at hushmail dot com
so true. *sigh* the current economic models are a pile of utter crap.
- former economist and now electrical engineer.
its very simple. LCD or lowest common denominator. intelligence is indicated when LCD is NOT the path chosen to solve a particular problem AND the problem is solved anyway. natural selection always chooses the lowest common denominator or shortest path to solve a particular problem set. Intelligence does not. intelligence definitely requires either a brain or some centre of processing analogous to a brain. A CPU or turing machine will do if it can emulate a brain like process.
Consider a parrot which wants a cracker. Everytime it wants a cracker is sez "Polly wants a cracker" at which time a human feeds it a cracker as a reward. Natural selection gave the parrot wings to fly and eat berries from trees. That was a lowest common denominator solution i.e. berry on tree -> parrot must eat -> give parrot wings. However the parrot is also intelligent and understands that crackers dont grow on trees and there is NO way it can get a cracker without pleasing its human keeper. So in order to get a cracker it must please its human keeper and to do that it must reproduce a set of sounds accurately in order to obtain the cracker.
so intelligence is : cracker with human -> learn sound to please human -> reproduce sound when human in room with cracker -> parrot gets cracker.
while natural selection is : cracker with human -> fly to human and attack human to get cracker -> human too large -> get some other food item.
crude analogy but i think i made my point.
just out of curiosity...has anyone done a realistic model of a single neuron ? or at least a complete emulation of a single neuron in software ? that includes all the growth and electrical/chemical effects ?
i check minduploading.org (MURG list at http://minduploading.org/mailman/listinfo/murg) but they dont seem to have accurate models.
and now consider this -- youve sold yourself to the highest bidder and gone and done something that compromized your ethics for what ? a job that pays 40-50K/yr ?
its a dog eat dog world but compromising your ethics really isnt worth it. you can always get another job -- hell construction/trucking/sales positions are always available. and in virtually all other jobs you dont have to compromize your ethics.
In case anyones wondering, yes ive had this situation before where i had to fudge some benchmarks. i said no, and i expected to get fired (which i did of course). what did i learn from this ? [a] theres other jobs out there. [b] being someones whore isnt worth it. [c] i still have my self respect (and a lot more respect from my ex-co-workers).
how does the job situation look in lawyer land right now BTW ?
bad or good ? getting any offers ?
should be fairly easy to do -- heck i could do it. the only problem is finding someone with $4 mil to invest in a project like this as a test case.
it wont take $20 mil unless you include airconditioning, the HALON fire suppression system and building leases.
dual processor linux workstation ? :
no problem -- heres the specs.
1 x Asus A7M266-D socket A motherboard.
2 x Socket A AMD MP processors (1900+ should be ok)
4 x 184-pin DDR Crucial PC2100 1024MB RAM DIMMs
(4 gig RAM -- max it out)
1 x Matrox Marvel G400-TV video card (ok not the best for gaming but it does lotsa video stuff and is well supported in linux).
2 x Fujitsu MPG3409AH 40.9GB HDDs (mirror/RAID em via software IDE RAID)
1 x Pioneer IDE DVD-R drive for backups and reading DVDs and CDs.
2 x 3COM 3c905B Ethernet controller.
Any decent LCD monitor or CRT monitor-- choose your poison.
Its not the fastest system but it'll do what you want.
For frills you can also add
1 x matrix orbital LCD with floppy drive plate
1 x Yeong Yang 0221XB Black Cube Server Case with redundant 400W PSUs
1 x Axis network camera for video conferencing
1 x 802.11b wireless PCI adapter (make sure its got a prism 2 chipset)
Try this config. i've done it before and it works.
its not a screaming bleeding edge game machine..ive sacrificed speed for reliability. ive also made it more pricier for more reliable components. shouldnt burn a hole in yer pocket tho.
i'd buy something like this for my next system but i have my eye on a dell inspiron 8200 as soon as i get reasonable employment. bleh. recessions suck.
uuh... did you really testify before congress ?
what was it like ?
Palm did the exact same thing initially -- they gave away the OS (ROM dump of several versions of debug OS), devtools, emulator etc etc.
;)
They still do if you sign the NDA and fax it to em.
They sold the hardware with the non debug OS.
do the same. give away a DEBUG version of your tools which cant be used in production and SELL the complete non debug production version.
just dont make the debug version an eval or time limited. let it print gobs of debug messages and make it bleeding edge.
That'll be $400 for my consulting fee. Thanks.
some more references :
http://www.ifi.unizh.ch/ailab/projects/evol/
these guys typically need a version control system or equivalent to do the job. with more than a dozen ppl chucking everything into a single directory makes a huge mess.
i've actually been writing my own CMS with version control/web based GUI editor etc etc and it looks pretty good. if anyone is interested email me for a license. its still beta but it should do the job.
are you stupid ?
have you looked at monster.ca's job forums ? Canada has a nearly 65% unemployment rate for IT...the highest in north america.
moving here ? youve got to be kidding.
uuh..maybe because in order to decrypt you have to solve the problem ?
a mathematical knot cant be solved so you can never decrypt the information in the first place...rendering the entire scheme futile.
and security by obscurity never works. we already do padding.
any other ideas ?
hmm...maybe. we dont know if intelligence and entropy *are* linked only when considering intelligent organisms tho. how do you prove its not linked when considering an intelligent organism ?
you can probably double the power with a nicol prism and two of those thingies.
coherent optics r00l.
i dont think autopilot.sf.net can be used with one of these to fly. it seems to be limited to normal RC heli models not this weird 4 propeller gyro stabilised carbon fibre flying machine.
The motion an handling characteristics of a 4 propeller machine v/s a single (large) bladed chopper are way different.
how is http://nathandial.com/mirror/globalenv/ better from a research PoV ? real research pushes the boundaries -- if we can somehow find a useful solution to eliminate bugs its prolly going to be non obvious so jumping on weird tehcniques like this may be a valid solution to eliminate bugs in the near future.
it doesnt seem practical now of course but what if they could get bugs to soun different from the main portions of the program in some way ?
thats one hell of a maintainence issue. means they gotta keep spending cash which they may not have in the future.
the main challenge is being able to autonomously control the heli indoors..i.e. avoid hitting walls with rototr blades etc.
follow the horizon autopilots have been done to death in most of the UAV competitions.
object tracking/scene recognition is the real challenge.
or when you use deadlocks and semaphores.
He did a lot more than just eliminate GOTOs.
dude. next time post under your own logged in name when posting such a godamn incoherent post. .NYET
autocoding aint going anywhere. not anytime soon anyway. and it has nothing to do with
computers cant program themselves either. if they could solve the deep theorems as you prove it they would be able to program themselves pretty easily -- all computer programs are essentially deep theorems.