Raczej uważam, iż jestem odpornym na zalewającą nas http://www.racjonalista.pl/forum.php/z,0/d,1/s,533403#w533705 głupotę http://www.racjonalista.pl/forum.php/z,0/d,43/s,600722#w603757 ale czasem trudno dać sobie na tzw. „wstrzymanie”. Może ktoś z Państwa wyjaśniłby mi…
Ciekawa prezentacja, ale.
AI nie jest rozny od wielu innych technologii.
Noz mozna uzyc do krojenia tortu ale tez do
AI pomaga nam zrozumiec jak my ludzie myslimy
Sa ludzie ktorzy ostrzegaja przed rozwojem AI.
Jakis czas temu byli przeciwnicy kolei twierdzac ze lokomotywy beda zabijalu krowy .Pewna doza sceptycyzmu jest potrzebna ale obecne roboty sa tak prymitywne ze obawa o utrate rzadu swiatem na rzecz robotow jest literacka fantazja .W USA pojawilo sie kilka filmow z superinteligientnymi robotami. Sa one popularne bo robia dobra kase a nie dlatego aby ostrzegac ludzi
Japonia przoduje w zastosowaniu robotyki ze wzgledu na demografie , coraz wiecej starych ludzi ,coraz mniej roboczej sily.Najdluzej ludzie zyja na wyspie Okinawa .
Amerykanski dziennikarz zapytal 100 letniego Japonczyka na Okinawie Czemu zawdziecza swoje dlugie zycie? Odpowiedz byla: ja nie wiem ale w nastepnym domu mieszka moja matka. Ona moze wiedziec.
Additional comments .
The speaker of the attached video pointed out correctly that the core of the AI capability is the current development of supercomputers .They have very large memories(terabytes) and operate in parallel performing trillions of arithmetic and logic operations per second. The fastest machine on the planet performs around 50 PetaFlops .This is 50 *10 to the power 15 Floating Point operations per second.
A hand held supercomputer can do one trillion Flops ,
10 to the power 12.
However we need to keep in mind that computers are much faster than human brain but so far are inferior in terms of reasoning, logic and common sense.We can say that current supercomputer is less intelligent than a cocroach (pluskwa).Cocroach can find food, hide from preditor and mate.Very sophisticated functions developed over milions of years of evolution.
Supercomputer can beat a grandmaster in chess but
it plays very differently than human . It can remember a vast number of games and tries a huge number of options.
We think well but we aer slow .Signals in brain travel
about a foot per second. Electronic circuit is many times faster(somewhat less than the speed of light).
Tanks Krzysztof for this esej. It is more cheerful than
What we can see now in AI may be just a part of a whole complex intelligent data processing. That is why its speed is so fascinating. I would definitely support further research and development. Fears are fears and dealing with them is just a part of any popularization campaign. They may look scary, but in a long term run they may even help as long as they are manageable.
The video speaker mentioned an interesting historic fact.
Several timers computer scientists predicted that AI would soon achieve human capabilities. Several times they were wrong.
I worked several years at The Boeing AI Laboratory
but decided that there were better more useful things to do. One of them is Hihg Performance Computing that advances science and technology .So far AI has been a wishful thinking and poor investment. Be careful.
Marvin Minsky who coined the name AI has died .Others
make good money selling a snake oil to naive believers.
However Japanese have made some reasonable progress. Their robots can dance or climb stairs.
All what we can say is that human capabilities for AI driven robots are not around the corner. Invest in AI only if you plan to live very long time!
Good point, some matching areas between computing technology and cognitive processes does not necessarily conclude that AI can be reached. I have never believed that AI is waiting around the corner. In my article published in the „Futures” magazine I pointed out that the Big Blue win over Kasparov was a misleading signal for AI followers. This is the same mechanism as with any other human imperfection promised to be fixed by marketing campaign, just like medical treatment giving a hope for less attractive people to look like top models. The fact is that most people have average cognitive skills, so there is a huge market of dreamers hoping to be smarter thanks to AI. But there is no such progress yet, and even more – so far there is no evidence this can be done anytime. Even driverless cars failed, nothing to say about more sophisticated mental tasks..
Słabe to, mnostwo błedow.
Za to krytyka polskiego wszystkowiedza Lucka bardzo błyskotliwa. Czy ja naprawdę muszę czytać te Pana "okruchy geniuszu" ?