Email or username:

Password:

Forgot your password?
Odoben's posts Post Back to profile
Top-level
Odoben

I keep reading about how LLM generated pages are already flooding a considerable chunk of the internet but at the same time I rarely do encounter those pages so my internet usage habits must be fairly good. Or maybe I'm just adapting to using general search engines less.

3 Feb 2024 at 11:43 | Open on meow.social
7 comments
Odoben

Machine learning has immense potential to improve a lot of tools used by creators; things like content-aware fill, removing backgrounds from images, colour-correcting video, applying filters, reducing image noise, cleaning up audio. Just don't use it to generate all the content from scratch...

17 Mar 2024 at 19:57 | Open on meow.social
Odoben

"The lawyer got in trouble for citing cases that didn't actually exist because they asked ChatGPT for help writing their legal brief."

"It's almost like you shouldn't off-load your own human-level work onto something with roughly the synapse count of a flatworm and think it'll do a good job."

2.5admins.com/2-5-admins-191/

21 Apr 2024 at 14:34 | Open on meow.social
Odoben

I stopped using Reddit when they took away the alternative client I was using. I stopped using Twitter when they took away the alternative client I was using. It's not like I decided to stop what I was doing, they just took it away so they probably don't want me to be a user anyway. :blobcatgooglyshrug:

3 May 2024 at 22:04 | Open on meow.social
Odoben

Here's an example of an actually genuinely awesome use of AI machine learning in science: AlphaFold
deepmind.google/technologies/a

28 May 2024 at 21:47 | Open on meow.social
Odoben

youtu.be/3ihjz7g1OQM
Listening to this interview with Google's director of generative AI kind of got me to actually start being okay with calling the current ML tools "AI" and start being interested in it generally. Because there's cool use cases besides plagiarism as a service.

We just really, really need to solve the copyright issues.

8 Jun 2024 at 14:04 | Open on meow.social
Odoben

what a title

Applications of these systems have been plagued by persistent inaccuracies in their output; these are often called “AI hallucinations”. We argue that these falsehoods, and the overall activity of large language models, is better understood as bullshit in the sense explored by Frankfurt (On Bullshit, Princeton, 2005)

A paper titled "ChatGPT is bullshit" published in Ethics and Information Technology Abstract: "Recently, there has been considerable interest in large language models: machine learning systems which produce human-like text and dialogue. Applications of these systems have been plagued by persistent inaccuracies in their output; these are often called “AI hallucinations”. We argue that these falsehoods, and the overall activity of large language models, is better understood as bullshit in the sense explored by Frankfurt (On Bullshit, Princeton, 2005): the models are in an important way indifferent to the truth of their outputs. We distinguish two ways in which the models can be said to be bullshitters, and argue that they clearly meet at least one of these definitions. We further argue that describing AI misrepresentations as bullshit is both a more useful and more accurate way of predicting and discussing the behaviour of these systems."
28 Aug 2024 at 7:46 | Open on meow.social
Odoben

New ASUS NUC has a Copilot button on the PC itself 🤡

ASUS mini PC with front IO visible: power button, audio jack, 2x USB-A, 1x USB-C, Copilot button.
7 Sep 2024 at 15:16 | Open on meow.social
Go Up