|
|
Post by MartinT on Aug 8, 2023 10:42:17 GMT
Politicians don't believe in transparency.
|
|
|
|
Post by MikeMusic on Feb 25, 2024 16:47:29 GMT
What AI tool can you recommend to summarise YouTube videos ?
To save time, maybe lots of time
I'm tired of going through videos that are not what they say they are
Videos of 30 minutes, 60 minutes can usually be easily summarised
I haven't been using ChatGPT or any other AI tool and I reckon I'm missing a trick or three
Perhaps an AI can be trained to summarise
"This is not what the title says"
"This takes over an hours to say one thing, which is......"
even
"This is rubbish"
|
|
|
|
Post by MartinT on Feb 25, 2024 17:05:29 GMT
|
|
|
|
Post by MikeMusic on Apr 10, 2024 9:23:30 GMT
New Scientist article
AI has its data checked by people who often don't know the field they are working in
The answers which could be right, wrong or somewhere in between are then in the answers given by AI forevermore
So AI even more faulted than I thought
|
|
|
|
Post by Slinger on Apr 10, 2024 13:16:35 GMT
AI is fuelled by enormous amounts of raw data. The internet is big. You just won't believe how vastly, hugely,mindbogglingly big it is. I mean, you may think a New Scientist article is big, but that's just peanuts to The Internet. (With apologies to Douglas Adams) The point is, an AI model begins with scraping up an unimaginably huge amount of raw data. Now, what else do we all know about the Internet apart from how immense it is? That everything ever committed to it, from E=mc 2 to Auntie Lakshmi's recipe for Murgir Jhol, is the absolute, complete and utter truth...  |
|
|
|
Post by John on Apr 10, 2024 13:51:35 GMT
42
|
|
|
|
Post by MartinT on Apr 10, 2024 14:29:29 GMT
Think of it as we do now with Wikipedia.
Highly useful, but fact-check everything.
|
|
|
|
Post by MikeMusic on Apr 10, 2024 14:42:18 GMT
Quote from a review of the book
Code Dependant by Madhumita Murgia
"data annotator ... labelling bones for medical AI without knowing human anatomy"
"other labellers sometimes have to guess at the least wrong answer"
|
|
|
|
Post by MikeMusic on Apr 10, 2024 14:43:37 GMT
Think of it as we do now with Wikipedia. Highly useful, but fact-check everything. Ok if you have plenty of time and give a damn When you are surrounded by fire and bogey men the fastest answer could well be the one to go for |
|
|
|
Post by Slinger on Apr 10, 2024 15:03:04 GMT
Think of it as we do now with Wikipedia. Highly useful, but fact-check everything. But what does one fact-check with, and why not go there first?  |
|
|
|
Post by MartinT on Apr 10, 2024 15:07:57 GMT
Because AI, like Wikipedia, can give you a base from which to start  |
|
|
|
Post by rfan8312 on May 21, 2024 13:03:23 GMT
With some of the conversations happening on TAS recently I thought I'd ask a general question about AI.
Is it possible that we unconsciously created AI to solve the world's problems for us?
With its processing power seems like it could put pieces together (for example in all of our writings in libraries and articles and books written each year ) that are right in front of us but that our minds dont have the processing power to assemble into one picture.
|
|
|
|
Post by MartinT on May 21, 2024 14:26:02 GMT
Being cynical, I think the long-term goal for AI was to replace people in jobs where people make mistakes and are more expensive.
|
|
|
|
Post by Slinger on May 21, 2024 16:03:58 GMT
Being cynical, I think the long-term goal for AI was to replace people in jobs where people make mistakes and are more expensive. AI was developed to replace people in jobs where they have to pay people, unlike the invention of "robots," the original and far more philanthropic (original) aim of which was to take over menial tasks, thereby giving the working classes more leisure time. ...and not having to pay them. |
|
|
|
Post by MikeMusic on Sept 27, 2024 10:25:23 GMT
Interesting article in New Scientist
Broadly saying that AI won't go on to become any more than it currently is
The applications for use will expand of course, but that's our lot they say
|
|
|
|
Post by MartinT on Sept 27, 2024 11:00:41 GMT
No-one has put an AI learning algorithm into a really powerful Quantum Computer yet.
Give it time...
|
|
|
|
Post by nicholas on Sept 27, 2024 14:46:59 GMT
No-one has put an AI learning algorithm into a really powerful Quantum Computer yet. Give it time... Once that happens maybe we'll get an answer for Fermi's Paradox? |
|
|
|
Post by MartinT on Sept 27, 2024 15:16:59 GMT
Once that happens maybe we'll get an answer for Fermi's Paradox? 42? |
|
|
|
Post by Slinger on Sept 27, 2024 15:18:47 GMT
I'm sorry, Dave...
|
|
|
|
Post by daytona600 on Sept 28, 2024 13:10:27 GMT
|
|
