Merged Artificial Intelligence

[Wudang]

The argument that we haven't done X yet and I don't see how X is possible therefore X is impossible does not have a great success rate in science.
Personally I think LLMs are more interesting from a psycholinguistic research perspective than anything else but I'm interested to see what the future brings.

 

[The Great Zaganza]

There is a fundamental flaw in trying to create an artificial human-like intelligence: we have plenty of that type, all prone to the same fallacies limits.

We should try to create a totally alien type of intelligence to think in ways we never would

 

[Wudang]

Yes. I find people saying “this AI is great. It makes mistakes like a human” to be like someone telling me “This car is great. It trips over loose paving and plants your face in the dirt like a human would”.

 

[arthwollipot]

I'm not sure I've seen anyone saying that, though.

 

[Wudang]

Now this looks promising. Via @leibnizopenscience@mastodon.social

90% of science is lost. This new AI just found it

Vast amounts of valuable research data remain unused, trapped in labs or lost to time. Frontiers aims to change that with FAIR² Data Management, a groundbreaking AI-driven system that makes datasets reusable, verifiable, and citable. By uniting curation, compliance, peer review, and interactive...

www.sciencedaily.com

 

[Darat]

I don't know enough to know if they are doing it right but that is the sort of thing we should be using AI for, not trying to replace dirty, smelly, needy humans in customer services.

 

[Opcode]

I don't know enough to know if they are doing it right but that is the sort of thing we should be using AI for, not trying to replace dirty, smelly, needy humans in customer services.

Today is AI Day at University of Texas at Dallas' Jindal School of Management. The opening keynote speaker was Nikunj Nirmal, Head of Engineering for Amazon. He said that he considers AI as "Augmented Intelligence." Instead of replacing humans, it will augment them, making them more efficient workers. This on the same day that Amazon lays off 30k workers due to AI productivity gains.

 

[Dr.Sid]

Of course, they clearly failed to replace ALL the humans. So augmenting it is. Till next month.

 

[arthwollipot]

I don't know enough to know if they are doing it right but that is the sort of thing we should be using AI for, not trying to replace dirty, smelly, needy humans in customer services.

AI is great at sorting through massive amounts of data to find patterns.

 

[Puppycow]

Today is AI Day at University of Texas at Dallas' Jindal School of Management. The opening keynote speaker was Nikunj Nirmal, Head of Engineering for Amazon. He said that he considers AI as "Augmented Intelligence." Instead of replacing humans, it will augment them, making them more efficient workers. This on the same day that Amazon lays off 30k workers due to AI productivity gains.

Wow. Ironic.

It is economically logical though.
IF
(1) AI can actually make workers more efficient, then
(2) Fewer workers should be able to achieve the same results.

It has ever been thus. One man with an excavator can dig a hole more efficiently than 10 men with shovels. The excavator "augments" the operator, making him a more efficient worker.

Questions remain about the quality of AI output, although the quantity is probably there. If your primary concern is quantity rather than quality, then AI can make workers more efficient.

 

[Skeptical Greg]

With an excavator, you still need someone to jump in the hole to dig around pipes and cables. But I'm sure there will be a robot for that someday

 

[Darat]

You still need the 6 people to look down the hole. That's apparently very important around here.

 

[Dr.Sid]

You still need the 6 people to look down the hole. That's apparently very important around here.

Robots have still some issues with bipedal walking, especially over rough terrain. But they can do "just standing there" quite fine. Add some decent camera so they can look down the hole properly, and you can ship it.

 

[EHocking]

Robots have still some issues with bipedal walking, especially over rough terrain. But they can do "just standing there" quite fine. Add some decent camera so they can look down the hole properly, and you can ship it.

Will they need shovels to lean on?
It looks like this is just the thin edge of the wedge.
(or it would if YT worked on this board)

 

[Darat]

Will they need shovels to lean on?
It looks like this is just the thin edge of the wedge.
(
or it would if YT worked on this board)

See:https://internationalskeptics.com/forums/index.php?threads/youtube-embeds-not-working.377918/unread

 

[Darat]

Interesting article about how LLMs think through tasks: https://www.searchenginejournal.com/anthropic-research-shows-how-llms-perceive-text/559636/

 

[theprestige]

Interesting article about how LLMs think through tasks: https://www.searchenginejournal.com/anthropic-research-shows-how-llms-perceive-text/559636/

It's interesting that the LLM model evolves an heuristic, rather than writing a script.

But they lose me at that stupid apostrophe. And alienate me for good when the whole article is written in the service of better subverting of search engine results.

 

[EHocking]

See:https://internationalskeptics.com/forums/index.php?threads/youtube-embeds-not-working.377918/unread

You mean the the thread where my reply to the OP is post #2?
Yeah, I’m aware of it.

 

[NatalieSalat]

More playing with old TV/films and AI:

Brenda the bully in Orphan (2009):








Margaret Ann in "Lime Street" (1985):

 

[NatalieSalat]

A sister and her brother on "Cheers" in 1986:





But she doesn't need him: