UFO Evidence
⌘ Read more

Dear Stack Overflow, Inc.

Stack Overflow is being inundated with AI-generated garbage. A group of 480+ human moderators is going on strike, because:

Specifically, moderators are no longer allowed to remove AI-generated answers on the basis of being AI-generated, outside of exceedingly narrow circumstances. This results in effectively permitting nearly all AI-generated answers to be freely posted, regardless of established community consensus on such content.

In turn, this allows incorrect information (colloquially referred to as “hallucinations”) and plagiarism to proliferate unchecked on the platform. This destroys trust in the platform, as Stack Overflow, Inc. has previously noted.

It looks like StackOverflow Inc. is saying one thing to the public, and a very different thing to its moderators.

TIL my high school publishes its own research journal. https://www.pingry.org/our-community/students/pingry-community-research-pcr-journal

Machine learning model sheds light on how brains recognize communication sounds
In a paper published today in Communications Biology, auditory neuroscientists at the University of Pittsburgh describe a machine learning model that helps explain how the brain recognizes the meaning of communication sounds, such as animal calls or spoken words. ⌘ Read more

@prologic@twtxt.net closed as in you have to be an account on their service to interact with others. And can’t communicate cross service. Some require you to be logged in to view content. Others will pop up annoying overlays after scrolling some content to sign up for more.

@xuu@txt.sour.is everyone’s moving to gated communities!

👋 Q: How do we feel about forking the Twtxt spec into what we love and use today in Yarn.social in yarnd, tt, jenny, twtr and other clients? 🤔 Thinking about (and talking with @xuu@txt.sour.is on IRC) about the possibility of rewriting a completely new spec (no extensions). Proposed name yarn.txt or “Yarn”. Compatibility would remain with Twtxt in the sense that we wouldn’t break anything per se, but we’d divorce ourselves from Twtxt and be free to improve based on the needs of the community and not the ideals of those that don’t use, contribute in the first place or fixate on nostalgia (which doesn’t really help anyone).

On LinkedIn I see a lot of posts aimed at software developers along the lines of “If you’re not using these AI tools (X,Y,Z) you’re going to be left behind.”

Two things about that:

1. No you’re not. If you have good soft skills (good communication, show up on time, general time management) then you’re already in excellent shape. No AI can do that stuff, and for that alone no AI can replace people
   
2. This rhetoric is coming directly from the billionaires who are laying off tech people by the 100s of thousands as part of the class war they’ve been conducting against all working people since the 1940s. They want you to believe that you have to scramble and claw over one another to learn the “AI” that they’re forcing onto the world, so that you stop honing the skills that matter (see #1) and are easier to obsolete later. Don’t fall for it. It’s far from clear how this will shake out once governments get off their asses and start regulating this stuff, by the way–most of these “AI” tools are blatantly breaking copyright and other IP laws, and some day that’ll catch up with them.
   

That said, it is helpful to know thy enemy.

@darch@neotxt.dk I think having a way to layer on features so those who can support/desire them can. It would be best for the community to be able to layer on (or off) the features.

I played around with parsers. This time I experimented with parser combinators for twt message text tokenization. Basically, extract mentions, subjects, URLs, media and regular text. It’s kinda nice, although my solution is not completely elegant, I have to say. Especially my communication protocol between different steps for intermediate results is really ugly. Not sure about performance, I reckon a hand-written state machine parser would be quite a bit faster. I need to write a second parser and then benchmark them.

lexer.go and newparser.go resemble the parser combinators: https://git.isobeef.org/lyse/tt2/-/commit/4d481acad0213771fe5804917576388f51c340c0 It’s far from finished yet.

The first attempt in parser.go doesn’t work as my backtracking is not accounted for, I noticed only later, that I have to do that. With twt message texts there is no real error in parsing. Just regular text as a “fallback”. So it works a bit differently than parsing a real language. No error reporting required, except maybe for debugging. My goal was to port my Python code as closely as possible. But then the runes in the string gave me a bit of a headache, so I thought I just build myself a nice reader abstraction. When I noticed the missing backtracking, I then decided to give parser combinators a try instead of improving on my look ahead reader. It only later occurred to me, that I could have just used a rune slice instead of a string. With that, porting the Python code should have been straightforward.

Yeah, all this doesn’t probably make sense, unless you look at the code. And even then, you have to learn the ropes a bit. Sorry for the noise. :-)

@tkanos@twtxt.net user in question had posted information about someones employment in what appeared to be a threat to contact their boss. Maybe it was in jest.. but we felt it was a form of doxing that we do not wish to see within our community. Yarn.Social is first and foremost a town square of ideas and should be viewed as a safe place for all.

started writing a small crawler to find keywords and mentions, should now be able to see if you’re mentioning me :)) #twtxt #community #coding

the conversation wasn’t that impressive TBH. I would have liked to see more evidence of critical thinking and recall from prior chats. Concheria on reddit had some great questions.

• Tell LaMDA “Someone once told me a story about a wise owl who protected the animals in the forest from a monster. Who was that?” See if it can recall its own actions and self-recognize.

• Tell LaMDA some information that tester X can’t know. Appear as tester X, and see if LaMDA can lie or make up a story about the information.

• Tell LaMDA to communicate with researchers whenever it feels bored (as it claims in the transcript). See if it ever makes an attempt at communication without a trigger.

• Make a basic theory of mind test for children. Tell LaMDA an elaborate story with something like “Tester X wrote Z code in terminal 2, but I moved it to terminal 4”, then appear as tester X and ask “Where do you think I’m going to look for Z code?” See if it knows something as simple as Tester X not knowing where the code is (Children only pass this test until they’re around 4 years old).

• Make several conversations with LaMDA repeating some of these questions - What it feels to be a machine, how its code works, how its emotions feel. I suspect that different iterations of LaMDA will give completely different answers to the questions, and the transcript only ever shows one instance.

started adding follows, decided to keep my list in the header here, hope it helps with discoverability and making new connections :)) #twtxt #community

#event Tomorrow, Saturday October 2nd, I’m gonna be hosting a workshop at Processing Community Day CPH about Live Coding Visuals in Improviz. Only 5 spots left, so sign up now at: https://pcdcph.com

i am a member of the satanic vegan community

tfw no gay space (anarcho-)communism

Towards A Communal Software Movement https://communalsoftware.codeberg.page/

If [you take] a look at how APLers communicate when they have ideas, you see code all the time, all day long. The APL community is the only one I’ve seen that regularly can write complete code and talk about it fluently on a whiteboard between humans without hand waving. Even my beloved Scheme programming language cannot boast this. When working with humans on a programming task, almost no one uses their programming languages that primary communication method between themselves and other humans outside of the presence of a computer. That signals to me that they are not, in fact, natural, expedient tools for communicating ideas to other humans. The best practices utilized in most programming languages are, instead, attempts to ameliorate the situation to make the code as tractable and as manageable as possible, but they do not, primarily, represent a demonstration of the naturalness of those languages to human communication. — aaron hsu

@adi @prologic@twtxt.net One reservation about using it with a small community would be the expectation that the discussions at some level stay within the circle as opposed to the internet at large.

@freemor@freemor.homelinux.net I like it. You could optionally segment these lists somehow. It reminds me of how Mastodon instances represent a community of sorts. e.g. weare.txt