Sök igenom hela webbplatsen
Denna webbplats använder kakor för att fungera optimalt, analysera användarbeteende och för att visa reklam (om du inte är inloggad). Genom att använda LibraryThing intygar du att du har läst och förstått våra Regler och integritetspolicy. All användning av denna webbplats lyder under dessa regler.
Hide this

Resultat från Google Book Search

Klicka på en bild för att gå till Google Book Search.


Understanding Moore's Law: Four Decades of Innovation

av David Brock

MedlemmarRecensionerPopularitetGenomsnittligt betygDiskussioner
412,900,936 (4)Ingen/inga
Senast inlagd avraany, antao, DKL_ref, jiffernut

Gå med i LibraryThing för att få reda på om du skulle tycka om den här boken.

Det finns inga diskussioner på LibraryThing om den här boken.

"I think there is a world market for maybe five computers."

Thomas Watson, chairman of IBM, 1943

Ah, what about Moore’s Law?

In a very important sense Watson was right. He was right because he was looking at the subset of those problems that scientists were actively trying to solve circa 1943 that the first digital computers could address. And circa 1943 no scientist or engineer was actively trying to solve any problem that was hopelessly beyond the reach of the chalkboard, slide rule, and tabulating machine. To a huge extent the proliferation of computing power has been a supply-side phenomenon: now that this power exists, what can we use it for? And at last we know: we can use it to steal and disseminate the naked pictures that other people can now create and store in vulnerable places. Those dang computer scientists! They should have had the ethical training that would have stopped me from doing that...

Software is applied mathematics.

The hardware side of computer science is about how to implement the mathematical principles given the current limitations of engineering. As Leslie Lamport, a computer researcher with a mathematical background, put it: "Some readers may need reminding that numbers are not a string of bits and 2 to the power of 33 multiplied by 2 to the power 33 is 2 to the power 66 and not overflow error".

I agree that mathematicians and physicists are much more likely to make the important breakthroughs in computing. This has certainly been the case in the past - Turing (mathematician), Dijkstra (physicist – (*)), Codd (mathematician).

My feeling is that engineers and computer scientists often equate an increase in raw computing power with an increase in intelligence. Moore's "law" states that the number of transistors in a dense integrated circuit doubles approximately every two years. The fact that I can now fit the computing power of a 1980s mainframe in my pocket doesn't make it any cleverer.

The part of AI that I find most interesting is the robotics research attempting to emulate how birds fly or fish swim, with the goal of building machines that are as energy efficient as their natural equivalents. This involves building software that emulates the behaviour of the bird or fishes brain, to control the machines movements. Is this intelligence? I would argue that it is.

I've always found the idea of intelligence tests and IQ a little strange. How can you measure something as complex and ill-defined as intelligence with a single number?

Rodney Brooks' subsumption architecture, IIRC, did some interesting work on this way back in the 80s. It's an interesting topic, particularly as it highlighted some of the less justifiable assumptions made in AI (such that, if you solve the symbolic reasoning concept, you can worry about someone else handling the sensors-to-symbol conversion). My own personal interest has always been more in emergent behaviour from relatively simple agent AIs, which does somewhat use that same assumption about the environment; I've always gone by the (non-exclusive) maxim that intelligence can be defined as something which leads to rational, goal-orientated behaviour (somewhat in line with Wooldridge's definition of an intelligent agent). But there are so many other aspects to consider, too.... learning, autonomy, opportunism, probabilistic reasoning, etc.

Arguably we need a different word from intelligence to cover these axes, otherwise our definition of intelligence becomes - ironically - artificial.

NB (*):Dijkstra in the essay "The Humble Programmer" relates how he wasn't allowed to put computer programmer as his profession on his marriage license because that profession didn't currently exist in the Netherlands. However, I do think that their hard science or mathematical backgrounds were important. Dijkstra chose his profession, but he remained highly critical of the computer industry. ( )
  antao | Sep 6, 2020 |
inga recensioner | lägg till en recension
Du måste logga in för att ändra Allmänna fakta.
Mer hjälp finns på hjälpsidan för Allmänna fakta.
Vedertagen titel
Alternativa titlar
Första utgivningsdatum
Viktiga platser
Viktiga händelser
Relaterade filmer
Priser och utmärkelser
Inledande ord
Avslutande ord
Förlagets redaktörer
På omslaget citeras
Kanonisk DDC/MDS
Kanonisk LCC

Hänvisningar till detta verk hos externa resurser.

Wikipedia på engelska


Inga biblioteksbeskrivningar kunde hittas.


Populära omslag



Medelbetyg: (4)
4 1

Är det här du?

Bli LibraryThing-författare.


Om | Kontakt | LibraryThing.com | Sekretess/Villkor | Hjälp/Vanliga frågor | Blogg | Butik | APIs | TinyCat | Efterlämnade bibliotek | Förhandsrecensenter | Allmänna fakta | 164,616,746 böcker! | Topplisten: Alltid synlig