TinyTimmyTokyo

joined 2 years ago
[–] TinyTimmyTokyo@awful.systems 1 points 2 months ago

We should be trying to stop this from coming to pass with the urgency we would try to stop a killer asteroid from striking Earth. Why aren’t we?

Wait, what are we trying to stop from coming to pass? Superintelligent AIs? Either I'm missing his point, or he really agrees with the doomers that LLMs are on their way to becoming "superintelligent".

[–] TinyTimmyTokyo@awful.systems 10 points 2 months ago (4 children)

Why do AI company logos look like buttholes?

(Blog post written by a crypto-turned-AI bro, but the observation is amusing.)

[–] TinyTimmyTokyo@awful.systems 15 points 2 months ago

Maybe Elon can install Grok as the copilot of his private jets.

[–] TinyTimmyTokyo@awful.systems 6 points 3 months ago

Check out the by-line. Big surprise!

[–] TinyTimmyTokyo@awful.systems 15 points 3 months ago

"Thought process"

"Intuitively"

"Figured out"

"Thought path"

I miss the days when the consensus reaction to Blake Lemoine was to point and laugh. Now the people anthropomorphizing linear algebra are being taken far too seriously.

[–] TinyTimmyTokyo@awful.systems 2 points 3 months ago (1 children)

As a fellow Usenet junkie from way back, now I'm curious which newsgroups Yarvin hung out in.

[–] TinyTimmyTokyo@awful.systems 5 points 3 months ago* (last edited 3 months ago)

Yeah, it was a brain fart.

[–] TinyTimmyTokyo@awful.systems 5 points 3 months ago (3 children)

I always tune into Casey Newton and Kevin Roose's podcast to get my latest fix of AI hype, now that they've moved on from crypto hype and multiverse hype. Can't wait to see what the next hype cycle will bring!

[–] TinyTimmyTokyo@awful.systems 6 points 3 months ago

Scott talks a bit about it in the video, but he was recently in the news as the guy who refused to sign a non-disparagement agreement when he left OpenAI, which caused them to claw back his stock options.

[–] TinyTimmyTokyo@awful.systems 6 points 3 months ago (5 children)

I'm fascinated by the way they're hyping up Daniel Kokotajlo to be some sort of AI prophet. Scott does it here, but so does Caroline Jeanmaire in the OP's twitter link. It's like they all got the talking point (probably from Scott) that Daniel is the new guru. Perhaps they're trying to anoint someone less off-putting and awkward than Yud. (This is also the first time I've ever seen Scott on video, and he definitely gives off a weird vibe.)

[–] TinyTimmyTokyo@awful.systems 27 points 3 months ago (3 children)

After minutes of meticulous research and quantitative analysis, I've come up with my own predictions about the future of AI.

 

Not 7.5% or 8%. 8.5%. Numbers are important.

 

Non-paywalled link: https://archive.ph/9Hihf

In his latest NYT column, Ezra Klein identifies the neoreactionary philosophy at the core of Marc Andreessen's recent excrescence on so-called "techno-optimism". It wasn't exactly a difficult analysis, given the way Andreessen outright lists a gaggle of neoreactionaries as the inspiration for his screed.

But when Andreessen included "existential risk" and transhumanism on his list of enemy ideas, I'm sure the rationalists and EAs were feeling at least a little bit offended. Klein, as the founder of Vox media and Vox's EA-promoting "Future Perfect" vertical, was probably among those who felt targeted. He has certainly bought into the rationalist AI doomer bullshit, so you know where he stands.

So have at at, Marc and Ezra. Fight. And maybe take each other out.

 

Rationalist check-list:

  1. Incorrect use of analogy? Check.
  2. Pseudoscientific nonsense used to make your point seem more profound? Check.
  3. Tortured use of probability estimates? Check.
  4. Over-long description of a point that could just have easily been made in 1 sentence? Check.

This email by SBF is basically one big malapropism.

 

Representative take:

If you ask Stable Diffusion for a picture of a cat it always seems to produce images of healthy looking domestic cats. For the prompt "cat" to be unbiased Stable Diffusion would need to occasionally generate images of dead white tigers since this would also fit under the label of "cat".

 

[All non-sneerclub links below are archive.today links]

Diego Caleiro, who popped up on my radar after he commiserated with Roko's latest in a never-ending stream of denials that he's a sex pest, is worthy of a few sneers.

For example, he thinks Yud is the bestest, most awesomest, coolest person to ever breathe:

Yudkwosky is a genius and one of the best people in history. Not only he tried to save us by writing things unimaginably ahead of their time like LOGI. But he kind of invented Lesswrong. Wrote the sequences to train all of us mere mortals with 140-160IQs to think better. Then, not satisfied, he wrote Harry Potter and the Methods of Rationality to get the new generation to come play. And he founded the Singularity Institute, which became Miri. It is no overstatement that if we had pulled this off Eliezer could have been THE most important person in the history of the universe.

As you can see, he's really into superlatives. And Jordan Peterson:

Jordan is an intellectual titan who explores personality development and mythology using an evolutionary and neuroscientific lenses. He sifted through all the mythical and religious narratives, as well as the continental psychoanalysis and developmental psychology so you and I don’t have to.

At Burning Man, he dons a 7-year old alter ego named "Evergreen". Perhaps he has an infantilization fetish like Elon Musk:

Evergreen exists ephemerally during Burning Man. He is 7 days old and still in a very exploratory stage of life.

As he hinted in his tweet to Roko, he has an enlightened view about women and gender:

Men were once useful to protect women and children from strangers, and to bring home the bacon. Now the supermarket brings the bacon, and women can make enough money to raise kids, which again, they like more in the early years. So men have become useless.

And:

That leaves us with, you guessed, a metric ton of men who are no longer in families.

Yep, I guessed about 12 men.

 

Excerpt:

Richard Hanania, a visiting scholar at the University of Texas, used the pen name “Richard Hoste” in the early 2010s to write articles where he identified himself as a “race realist.” He expressed support for eugenics and the forced sterilization of “low IQ” people, who he argued were most often Black. He opposed “miscegenation” and “race-mixing.” And once, while arguing that Black people cannot govern themselves, he cited the neo-Nazi author of “The Turner Diaries,” the infamous novel that celebrates a future race war.

He's also a big eugenics supporter:

“There doesn’t seem to be a way to deal with low IQ breeding that doesn’t include coercion,” he wrote in a 2010 article for AlternativeRight .com. “Perhaps charities could be formed which paid those in the 70-85 range to be sterilized, but what to do with those below 70 who legally can’t even give consent and have a higher birthrate than the general population? In the same way we lock up criminals and the mentally ill in the interests of society at large, one could argue that we could on the exact same principle sterilize those who are bound to harm future generations through giving birth.”

(Reminds me a lot of the things Scott Siskind has written in the past.)

Some people who have been friendly with Hanania:

  • Mark Andreessen, Silion Valley VC and co-founder of Andreessen-Horowitz
  • Hamish McKenzie, CEO of Substack
  • Elon Musk, Chief Enshittification Officer of Tesla and Twitter
  • Tyler Cowen, libertarian econ blogger and George Mason University prof
  • J.D. Vance, US Senator from Ohio
  • Steve Sailer, race (pseudo)science promoter and all-around bigot
  • Amy Wax, racist law professor at UPenn.
  • Christopher Rufo, right-wing agitator and architect of many of Florida governor Ron DeSantis's culture war efforts
 

Ugh.

But even if some of Yudkowsky’s allies don’t entirely buy his regular predictions of AI doom, they argue his motives are altruistic and that for all his hyperbole, he’s worth hearing out.

view more: ‹ prev next ›