Free Republic
Browse · Search
General/Chat
Topics · Post Article

To: enumerated

BTW, it tool 4o less than a second to produce that answer.


32 posted on 01/31/2025 8:46:59 AM PST by RoosterRedux ("There's nothing so inert as a closed mind" )
[ Post Reply | Private Reply | To 29 | View Replies ]


To: RoosterRedux

Well, the speed is no surprise at all. I don’t think any reasonable person would doubt AI is superior in that regard. :-)


37 posted on 01/31/2025 9:00:33 AM PST by enumerated (81 million votes my ass)
[ Post Reply | Private Reply | To 32 | View Replies ]

To: RoosterRedux
"BTW, it tool 4o less than a second to produce that answer." (I presume that by "tool" you mean "took" and by "4o" you mean "GPT-4o".)

The speed is impressive, but I would be interested to know what sort of internal caching GPT-4o does for questions and answers.

Since your relatively sophisticated question, "Explain why increasing the minimum wage hurts entry level workers", is on a topic of current interest, it is possible that the answer is already in GPT's cache and quickly accessible via the three important linguistic components "Subject: increasing [the] minimum wage", "Verb: hurts" and "Object: entry level workers". (Computer parsing of such sentences has been around for many years.)

In other words, GPT-4o might be able to give the answer without consulting its full database and performing complicated analyses on what it finds therein.

It would also be interesting to know if GPT-4o makes use of a sophisticated CDN (Content Distribution Network) to offload -- and thus speed up -- some of the "simpler" queries to regional caches. (This has been done for years for video streaming and other data-hungry services.)

61 posted on 01/31/2025 5:32:33 PM PST by powerset
[ Post Reply | Private Reply | To 32 | View Replies ]

Free Republic
Browse · Search
General/Chat
Topics · Post Article


FreeRepublic, LLC, PO BOX 9771, FRESNO, CA 93794
FreeRepublic.com is powered by software copyright 2000-2008 John Robinson