Sounds like you already know at least one person who didn’t see this coming.

edit: oh blimey I didn’t realise this was an 8-month necro, oops

I approve of their very helpful and comprehensive documentation for LLMs: https://jepsen.io/llm/index

There’s a ridiculously simple test you can use to figure out whether AI mandates at a given company are well-considered, or are being shoved out the door by out-of-touch buffoons: What happens when an engineer documents that AI made a task worse? Are they taken seriously, or are they marginalized and eventually managed out?

Even easier test: is it an AI mandate?

“imagine themselves” was the important bit

I think the poster ad thing is a bit big for that. The idea of this fantasy is that the sloperator should get zero value from their defacement of public space, so it would have to meaningfully obscure any pertinent details like address etc. while also very helpfully telling the sloperator what their oversight was.

So what’s with the phenomenon of people that are dismissive of the capabilities of chatbots and freely list their constant fuckups, maybe even habitually calling them ‘clankers’, and yet still constantly fucking using them, comparing harnesses, blah blah blah? It’s not our usual cocaine analogies because these people are hardly evangelists… maybe still covered by ‘gambling addiction’? It’s just baffling to me. You humbly suggest that maybe they shouldn’t be making so much use of a chatbot that constantly bullshits them, and you don’t get any actual objection to that… and yet they continue on.

I poked around the repo some more a while ago… it’s somewhat enlightening into why the story is Like That. I should probably write up a sneer on it, though it was already out of the stubsack by that point…

Just so long as the future involves gas turbines.

A mere four paragraphs after noting the claims of AI labs “most certainly cannot be” taken at face value, half of those being just a blockquote, the author nonchalantly asserts that OpenAI’s attack on HuggingFace was the result of software “going rogue”. Bravo!

Another year, another “ink”tober done and dusted! Now I can focus on the important things in life, like video games. And draw nothing of any description for the next eleven months.

The last one (today’s 🧐) came out pretty well:

I saw a slop ad hanging on a fence today, and I’m in half a mind to buy some spray paint just so I can write “SLOP” over it

Never interrupt your enemy when they are making a mistake – you’ll lose the opportunity to take notes and copy them.

But what if the sufferhammers gave you velocity?

https://theluddite.org/post/spectacle-of-computation.html

The natural world does not come with a quantified representation of itself, but we are constantly awash in so much data that we forget that it must come from someone with their own interests, flaws, and ideologies, no matter what they claim, or how much computation they put it through.

What if you did productive and creative labour, that benefits your employer and not you, for free? Now that’s impact!

Whatever maintenance work @[email protected] did on Pivot seems to have stopped the RSS feed from lagging behind by about a day. Now when an article is posted to TechTakes, it’s already in my reader! Yay!

Mitchell Baker: I need to be paid $7 million because that’s what I could be paid elsewhere

Then go elsewhere?

Turns out slop prose isn’t any less galling after you put it in quotation marks.

Maybe the very first sentence of your story shouldn’t have me stopping to work out whether you meant January or August.

You’re fixated on charitably reading between the lines when far more pertinent – the entire point of this discussion here – is that the possibility is being raised at all. Is water conscious? Should we consider the ramifications of drinking fluids that react to potential gradients? Why would anyone be conjecturing the consciousness of computer software if they did not already think it were plausible, and why would they spend any time discussing it in preference to anything else?