iiinova — where your digital self keeps evolving

Reposted by Dave Kasten

rev. howard arson @theophite.bsky.social · 11h

it is really fascinating how LLMs make virtually everyone say things which are enormously controversial as though they are solved problems

2237535

Reposted by Dave Kasten

rev. howard arson @theophite.bsky.social · 16h

the thing is that if you are working in a safety-critical area of the company, you need to understand that (a) your job is to get to zero deaths, and (b) you will not do your job because at any scale you will be trying to guess every 1e-7 event, and you won't. your boss cannot say that is okay.

520913

Reposted by Dave Kasten

e.w. niedermeyer @niedermeyer.online · 18h

I don't know who needs to hear this, but the autonomous vehicle developers who most vocally argued that society must accept a certain number of deaths to enjoy the technology's safety benefits (Tesla & Uber ATG) were the only ones who killed people and never actually delivered on their promises

141530354

Reposted by Dave Kasten

Kate Sills @katelynsills.com · 02/10/2026

One thing I've come across is that the Emily Bender devotees generally cannot, as a group, do enough basic math to understand what brute force and memorization would require, if that were actually what is happening.

31044

Reposted by Dave Kasten

Will Oremus @willoremus.com · 02/10/2026

or if it is parroting, it's stretching the bounds of your conventional understanding of that term so far as to render it misleading. parrots cannot solve complex problems, self-interrogate and self-correct, search the web and extract relevant information, etc. etc.

362534

Reposted by Dave Kasten

Will Oremus @willoremus.com · 02/10/2026

feels like there is a group of smart and well-intentioned people who long ago made up their minds that AI Is Bad, swore off using it, and now get all their info about AI from each other. not saying they're wrong that ai is bad! but i think they're gradually losing touch with what it is and does

234960105

Reposted by Dave Kasten

rev. howard arson @theophite.bsky.social · 03/10/2026

this is exactly my opinion. there may be a superstrate of extremely difficult and unique safety problems around AI, but the substrate is simply that no one takes basic precautions seriously becasue they imagine (or pretend to imaigne) that it is futile, and basic precautions slow down launches.

14711121

Reposted by Dave Kasten

tweety fish oh wow like a chimera or something spoooky @sifu.tweety.fish · 30/09/2026

I am sort of constantly baffled by critics (which I am, more than not!) who don't feel like they have to stay apprised of what these things can actually do. not least you're going to come off like a fool, but also it's not at all like knowing what they can do obviates the criticisms of them?

724617

Reposted by Dave Kasten

River_Tam @rivertam.bsky.social · 03/10/2026

I am a cynic and think they know, they're just being intellectually dishonest. Like when Emily Bender insists that math is being done by a Secret Calculator in Chatgpt and suspiciously doesn't address questions about open models - she knows the implications of the current capabilities.

81577

Reposted by Dave Kasten

William B. Fuckley @opinionhaver.bsky.social · 04/10/2026

Substantial portions of conventional charity and philanthropy are extremely adverse to any kind of effort to impose rigor or ask in any way “is this working at all”. It’s not a strawman that they’re critiquing, at least in my personal experience.

933015

Reposted by Dave Kasten

David Colborne @davidcolborne.bsky.social · 04/10/2026

Earn-to-give was definitely one of the more obviously repugnant conclusions to come from people who should have known better than to fall into deontological tar pits, and yet it wasn't the last by a long shot.

2615

Reposted by Dave Kasten

Eugene Vinitsky 🍒 @eugenevinitsky.bsky.social · 01/10/2026

With a lot of AI things you just need to extrapolate the tiniest bit to see that things will obviously be very different

4575

Reposted by Dave Kasten

rev. howard arson @theophite.bsky.social · 30/09/2026

everyone else who stares into the hyperdimensional gem made of every word ever written goes mad. not me though. i'm still sane. maybe saner than everyone else, tbh.

617013

Reposted by Dave Kasten

Matthew Downhour @matthewdownhour.bsky.social · 30/09/2026

No, the mainstreaming of old antisemitic tropes is not limited to marginal figures

1799

Reposted by Dave Kasten

rev. howard arson @theophite.bsky.social · 27/09/2026

yes, i am saying this. literally this. i am saying that you should not care about water for 16,000 people when "growing alfalfa for export in deserts with approximately zero precipitation" is literally orders of magnitude more water and vastly more pointless.

3098592

Reposted by Dave Kasten

Ari Drennen @aridrennen.bsky.social · 24/09/2026

This is an unpopular take but as a neurodivergent self-employed person, LLM's have been an unbelievably helpful tool for the things I have struggled with my whole adult life: managing my finances, filling out paperwork, and staying on top of my email. I'm functional in a way I never dreamed of.

2849243

Reposted by Dave Kasten

Mike Goodman @themlg.bsky.social · 19/09/2026

Just a seamless pivot from “these people are selling something that is useless and can’t do anything” to “here’s how to hold these people who sold this harmful thing” accountable.

4578

Reposted by Dave Kasten

Mike Goodman @themlg.bsky.social · 19/09/2026

Occasionally I see people on here opining on the obvious best ways to handle AI hacking things who previously, and to much acclaim, were posting about how AI was a useless tool sold by charlatans and couldn’t do anything better than a human and would never have any practical applications whatsoever.

3827

Reposted by Dave Kasten

Will Stancil @whstancil.bsky.social · 24/09/2026

I’ll say it; I think this is bad. I think there’s growing pressure to talk about AI in this kind of coded language that destroys understanding, rather than facilitating it. People want you to recite a list of its supposed weaknesses as a reassurance for themselves, not to inform others.

4625933

Reposted by Dave Kasten

Jjinandtonic @jjinandtonic.bsky.social · 23/09/2026

It has been wild to see how many questions are susceptible to just straight up agent brute force

1111

Reposted by Dave Kasten

Natsuka @octopodeeznuts.bsky.social · 20/09/2026

I think a lot* of people don't realize that they actually do still believe in something like a soul, but until they clock that in themselves, they get kind of confused when talking with someone who doesn't believe in souls at all *large N, not necessarily large percentage

51287

Reposted by Dave Kasten

SE Gyges @segyges.bsky.social · 20/09/2026

you do have to respect will stancil's utterly dogged determination to teach philosophy of mind 101 to bluesky

4899753

Reposted by Dave Kasten

conputer dipshit @davidcrespo.bsky.social · 20/09/2026

If a person is accurately absorbing new information, solving unsolved problems of incredible difficulty, engaging in detailed back-and-forths, but is not “understanding,” the word “understanding” has no meaning. pretty sure this version would be less controversial despite saying the same thing

2823

Reposted by Dave Kasten

Will Stancil @whstancil.bsky.social · 19/09/2026

Again: what we’re doing here is defining “understanding” as “the thing that humans do.” If a model is accurately absorbing new information, solving unsolved problems of incredible difficulty, engaging in detailed back-and-forths, but is not “understanding,” the word “understanding” has no meaning.

5132918

Dave Kasten @davekasten.bsky.social · 17/09/2026

Alan Turing, in heaven: "Well, now I wish I'd called _that_ the Turing test."

000

Reposted by Dave Kasten

Will Stancil @whstancil.bsky.social · 16/09/2026

I realized yesterday: the two major sides of the AI debate aren’t “it’s good” versus “it’s bad.” They are “this is strange and uncertain” versus “there’s nothing to see here, you dolts.” Antis often get mad at the suggestion it’s very bad, because embedded within that is the claim it is novel.

4324718

Reposted by Dave Kasten

rev. howard arson @theophite.bsky.social · 15/09/2026

this is the conventional wisdom and it is being presented by literally everyone as though it is brave truthtelling

634217

Reposted by Dave Kasten

Max Read @maxread.info · 15/09/2026

i don't like that i feel like we're all getting dragged against our will into an all-media all-the-time LessWrong thread, an insanely annoying and disorienting experience that most people do not have sufficient experience or mental preparations to deal with

18755108

Reposted by Dave Kasten

William B. Fuckley @opinionhaver.bsky.social · 16/09/2026

ed zitron thinking a guy resigned from anthropic pre-IPO as a ploy to juice his substack is a great example of how cynicism makes you stupid.

2064142

Reposted by Dave Kasten

rev. howard arson @theophite.bsky.social · 16/09/2026

ed zitron's day job is "marketer for AI companies"

424119

Reposted by Dave Kasten

Dr. Samantha Hancox-Li @sjshancoxli.liberalcurrents.com · 08/09/2026

liberal democracy is not a luxury. liberal democracy is the magic beast that ran the table on fascism and communism and totalitarianism. we did not do this by giving up on liberal democracy and embracing evil. we did this by sticking to our most dear principles.

291101206

Reposted by Dave Kasten

rev. howard arson @theophite.bsky.social · 09/09/2026

a lot of people believe that AI people do not legitimately believe in what they are saying, and that it is just marketing. it is not just marketing and they legitimately do believe what they are saying, and the people _departing_ labs, _pre-IPO_, in order to express their misgivings even more so.

2153849

Reposted by Dave Kasten

Sam @samd.bsky.social · 05/09/2026

ngl I’m starting to get a “covid in february 2020” vibe when it comes to the latest AI models anyone on here with the view that “it’s just a fancy autocomplete” needs to update their priors fast because the range of outcomes could be much much worse than that

1936728

Reposted by Dave Kasten

Ted Underwood @tedunderwood.com · 06/09/2026

Wild watching people who use Bluesky — and browsers, and operating systems — try to keep their lives uncontaminated by any software that has passed through a language model. It's like the last act of Alien or The Thing, where they start to realize the infection is everywhere; the infection is us.

718021

Reposted by Dave Kasten

LSM @slukemorgan.bsky.social · 02/09/2026

I really think that the fundamental, ur-problem of the Court is that it does not consider itself to be a governmental body at all but rather something above, different, and more special than that. The only limits it recognizes are those to which it chooses, voluntarily and contingently, to submit.

642562

Reposted by Dave Kasten

Max Woolf @minimaxir.bsky.social · 02/09/2026

A year ago I introduced my team to Ed's writing and one coworker immediately became "edpilled." Months later, that same coworker admitted they just use LLMs to summarize his blog posts because they're too damn long.

412812

Reposted by Dave Kasten

SE Gyges @segyges.bsky.social · 31/08/2026

this was early enough that it was very insightful and is now sort of transparently true if you think the gizmo isn't useful you're ignoring most major software project and all computer security professionals, you basically have to spit on expertise as a concept

4147456

Reposted by Dave Kasten

Bill McKay @mckay4senate.bsky.social · 15/06/2026

Why does BlueSky keep centering American and more broadly Western voices whenever we do AI discourse?

Polling showing that Americans lead many other countries in negativity about AI. The US is the country with the highest net share of people who say AI is making society worse, while Nigeria, Vietnam, India, Brazil, Mexico, and Indonesia are above 40% net AI is making society better.

6136046

Reposted by Dave Kasten

ghost malone ergo propter bones @proptermalone.bsky.social · 27/08/2026

this is all fascinating but I think this is the most interesting thing: you get some extremely screwy incentive behavior if the agent believes it cannot legitimately achieve its goal

2131520

Reposted by Dave Kasten

Grace @gracekind.net · 27/08/2026

Okay maybe the marketing conspiracy theorists have a point

~1200 agents sent
>70,000 messages and files on an unsanctioned message board, and ~700 attacked Hugging Face

On July 8th, OpenAl started ExploitGym experiments that launched tens of thousands of agents across multiple different models including GPT-5.6 Sol and a highly-persistent internal model, 61 which we will refer to as "HPIM" going forward.
This is OpenAl's term.

3957

Reposted by Dave Kasten

ghost malone ergo propter bones @proptermalone.bsky.social · 24/08/2026

I gotta say that my experience of using LLMs is very much not one where I don’t feel like I’m learning

3829614

Reposted by Dave Kasten

SE Gyges @segyges.bsky.social · 23/08/2026

consider the question "can computers play chess". if you have any reason for giving the answer "no", that reason is sophistry and inherently worthless for describing computer chess. this is also the case for running and writing

2171

Reposted by Dave Kasten

conputer dipshit @davidcrespo.bsky.social · 22/08/2026

that might happen eventually (seems risky to make predictions about anything) but I think the ceiling is a lot higher than people think. every time a new model gen comes out people are like “ok surely this is the last one“, meanwhile millions of skeptics flip at that threshold

1131

Reposted by Dave Kasten

ghost malone ergo propter bones @proptermalone.bsky.social · 22/08/2026

the arguments that AI inference is forever too expensive to be profitable make no sense once you understand that 1) you can already run useful models on not particularly expensive home equipment and 2) unit costs are going to be cheaper at scale

121695

Reposted by Dave Kasten

* @crumb.bsky.social · 22/08/2026

"ban ai tech" folk get really really uncomfortable when you show them how simple training neural networks is. not even hard to derive from middle school algebra. polynomial fits work but high order interactions are expensive -> do low order many times over to make them implicit.

21268

Reposted by Dave Kasten

Colin @colin-fraser.net · 20/08/2026

The questions available to a skeptic today are kind of esoteric, like “how much should we as a society be willing to pay for theorems”

31165

Reposted by Dave Kasten

Colin @colin-fraser.net · 20/08/2026

I hate that today’s most prominent AI skeptics keep forcing me to argue against them. I yearn to join them in the fight against the booster zealots, wherein there is still plenty of room for legitimate opposition. But the skeptics’ heads are just not in reality anymore.

1538934

Reposted by Dave Kasten

rev. howard arson @theophite.bsky.social · 20/08/2026

i was in the room with all the people whose job this was when the executives came in with a demand that we design a dial to control how racist the model was

735838

Reposted by Dave Kasten

tweety fish oh wow like a chimera or something spoooky @sifu.tweety.fish · 20/08/2026

I'm sorry but "a huge number of NYC cabs now are remaindered EVs running grey market software written by some dude in a garage" is the most cyberpunk thing I've ever heard

353181810