Check this out...

Want a great source of US political news on Mastodon?

US Politics in Real Time @uspolitics is a super useful curated account produced by Pell @pell, who describes themselves as "just another programmer."

@uspolitics provides summaries with links to a wide array of US political news stories from around the web.
 "Priority topics: Climate, Labor, Economy, Fascism, LGBT+, Internet regulation."

#USPolitics #politics #news #CuratedAccount

@uspolitics @pell

Edited post to remove reference to AI, as it was creating confusion. In addition to human curation, @uspolitics uses, "Llama2 (a type of AI model) [which] helps make decisions on story quality and creates hashtags, formatting, and summaries."

This use of an LLM to select, summarize and link to published news stories differs from more problematic uses of AI technology that distill information from a wide array of unattributed sources, which may or may not be reliable.

@mastodonmigration @uspolitics @pell I'd still pass on it as a news source. I ignore all AI summaries because they generally seem wrong, weirdly vacuous, or both.
@jredlund @mastodonmigration @uspolitics give it a shot! I've spent a lot of effort to ensure it double checks everything against positive sources. It might change your mind!

@pell @jredlund @uspolitics

Large Language Models LLMs (a term we should be using instead of AI, because there is nothing intelligent about them) are just a class of computer programming. Since people have been coding up solutions, some applications have been very helpful and others dangerous scams. It is not LLMs that are inherently bad, it is the abuse of this class of programming by some snake oil purveyors that is casting a bad light on this entire area of technology.

@mastodonmigration @pell @uspolitics I know that this technology will not go away. However, it is not just the snake oil salespeople that cause harm. LLMs are already putting journalists, artists, and writers out of work. The work produced by LLMs and programs like Midjourney and DALL-E2 (are these Large Image Modelers?) is not superior to that done by humans in most cases, it is just faster and cheaper. And people looking for easy side hustles are flooding various creative markets with junk.
@mastodonmigration @uspolitics @pell
Do they label which used Ai for the summary?
@KnightDamptey @mastodonmigration @uspolitics Hey, it's Llama2. More info is in the account header, which I update frequently.

@pell @mastodonmigration @uspolitics

Thanks.
But I meant will the post say if it was generated by a human or by an Ai?

@KnightDamptey @mastodonmigration @uspolitics yes, it's labelled in the bot's description. The summaries are created by an AI but they aren't exactly written exclusively by it, they are checked/merged from the news org's own human written Open Graph metadata.