TechTakes

Many will point out that AI systems are not yet writing award-winning books, […]

Holy shit, these chucklefucks are so full of themselves. To them, art and expression and invention are really just menial tasks which ought to be automated away, aren’t they? They claim to be so smart but constantly demonstrate they’re too stupid to understand that literature is more than big words on a page, and that all their LLMs need to do to replace artists is to make their autocomplete soup pretentious enough that they can say: This is deep, bro.

I can’t wait for the first AI-brained litbro trying to sell some LLM’s hallucinations as the Finnegans Wake of our age.

I maintain that people not having to work is a worst-case scenario for silicon valley VCs who rely on us being too fucking distracted by all their shit products to have time to think about whether we need their shit products

one of the follies of the early non-research internet was that the folks in charge of the more influential communities (Something Awful comes to mind) all tended to be weird fucking libertarian assholes insisting that debating fascists in a “free and fair environment” (whatever that is) must be a good thing, we can’t just ban them on sight for some reason. generally these weird libertarian assholes were motivated by typical weird libertarian asshole things — greed, or being fascists themselves.

and all of these horseshit policies around not banning fascists ended in complete disaster for those communities and those libertarian shitheads (SA again), but somehow they’re practically the only element of those early communities that carried over to the modern internet, likely because caring about community quality doesn’t make money, but pandering to nazi fuckwits does.

it probably goes without saying, but I take the apparently radical stance that nobody needs to interact with fascists and assholes and I won’t give their bullshit transit on any system I control. it’s always been surprising to me how many sysops consider “kick out the fucking fascists” an unworkable policy when it’s very easy to implement — it only gets hard when you allow the worst fucking people you know to gain a foothold.

You laugh, but people in incel circles are heralding the nascent arrival of better than real AI girlfriends.

Now I am laughing harder.

I seen someone on LinkedIn yesterday talking about how you'll be able to "chat with your database"

Like, what if i don't want to be friends with a database? What next ? a few beers with my fridge? Skinny dipping with my toaster?

Both options would be horrible, of course, but they'd both be better than spending another five fucking minutes on LinkedIn.

this was dunked on like 7 years ago already, maybe earlier

You know, I have no direct information as to what kind of background those people might’ve had for their trips, but 1337% it was wrong (or this person is just full of it). Set and setting. Always. And dear god if I came out of a trip with that? What a fucking abysmal prospect

And yet holy fuck that’s quite possibly a strong driver in all this insanity isn’t it?

You could guess my internal state right now, and you’d need zero psychedelics to do it

Ah yes the famous trial where HRC was prosecuted and convicted, I remember that, very analogous to Trump’s case, yes.

Big Yud has been huffing AI farts for so long he’s starting to hallucinate like one!

edit the replies once again prove that a blue check is an instant -10 point to credibility.

My moderate distaste for people who capitalise words To Imply Some Other Meaning without Directly Stating It knows no Bounds.

Why are techbros such shit at Lojban? It's a recurring and silly pattern. Two minutes with a dictionary tells me that there is {seldikca} for being charged like a capacitor and {nenzengau} for charging like a rechargeable battery.

Also Yud: I can't be a cult leader, because I dress funny.

i really, really don't get how so many people are making the leaps from "neural nets are effective at text prediction" to "the machine learns like a human does" to "we're going to be intellectually outclassed by Microsoft Clippy in ten years".

like it's multiple modes of failing to even understand the question happening at once. i'm no philosopher; i have no coherent definition of "intelligence", but it's also pretty obvious that all LLM's are doing is statistical extrapolation on language. i'm just baffled at how many so-called enthusiasts and skeptics alike just... completely fail at the first step of asking "so what exactly is the program doing?"

Straight line on a lin-log chart, getting crypto flashbacks.

Similar vibes in this crazy document

EDIT it's the same dude who was retweeted

https://situational-awareness.ai/

AGI by 2027 is strikingly plausible. GPT-2 to GPT-4 took us from ~preschooler to ~smart high-schooler abilities in 4 years. Tracing trendlines in compute (~0.5 orders of magnitude or OOMs/year), algorithmic efficiencies (~0.5 OOMs/year), and “unhobbling” gains (from chatbot to agent), we should expect another preschooler-to-high-schooler-sized qualitative jump by 2027.

Last I checked ChatGPT can't even do math, which I believe is a prerequisite for being considered a smart high-schooler. But what do I know, I don't have AI brain.

Who is more evolutionarily fit: a deaf person who appreciates technological progress and has kids, or an unloved eugenics enjoyer posting on hacker news?

I expect SneerClub to provide my alibis when reading one of these finally makes me snap.

Apparently deaf people never have kids. Hmm. Alright. Okay.

This guy isn't even the first techbro to suggest recently that if disabled people are allowed to have kids, eventually we will all be disabled. I know very very little about genetics, but I'm still pretty sure it doesn't work that way.

"You know, we just had a little baby, and I keep asking myself... how old is he even gonna get?"

Tegmark, you absolute fucking wanker. If you actually believe your eschatological x-risk nonsense and still produced a child despite being convinced that he's going to be paperclipped in a few years, you're a sadistic egomaniacal piece of shit. And if you don't believe it and just lie for the PR, knowingly leading people into depression and anxiety, you're also a sadistic egomaniacal piece of shit.

Always a good sign when your big plan is virtually identical to L. Ron Hubbard's big plan in the 70s.

Trying to figure out if that comment is a bit or not, and at least one of the poster's other comments on the video is this literary masterwork:

some people have observable x chromosones and others have observable y chromosones, and those categories are good to make useful distinctions, just like I make a distinction between a chair and a sofa... I still don't believe gender is real, and I can still observe sex. Sure, sex could be an illusion, but it doesn't matter. The chair is most certainly an illusion.... At what point does wood become a chair? Is a three legged chair a chair? A two legged? A one legged? A none legged? Is a seat and a chair the same thing? What if I break the seat in half? Is it still a chair?

The accessibility community is pretty divided on AI hype in general and this feature is no exception. Making it easier to add alt is good. But even if the image recognition tech were good enough—and it’s not, yet—good alt is context dependent and must be human created.

Even if it’s just OCR folks are ambivalent. Many assistive techs have native OCR they’ll do automatically, and it’s better, usually. But not all, and many AT users don’t know how to access the text recognition them when they have it.

Personally I’d rather improve the ML functionality and UX on the assistive tech side, while improving the “create accessible content” user experiences on the authoring tool side. (Ie. Improve the braille display & screen reader ability to describe the image by putting the ML tech there, but also make it much easier for humans to craft good alt, or video captions, etc.)