TinyTimmyTokyo

joined 1 year ago
[–] [email protected] 0 points 1 month ago

As anyone who's been paying attention already knows, LLMs are merely mimics that provide the "illusion of understanding".

 

Excerpt:

A new study published on Thursday in The American Journal of Psychiatry suggests that dosage may play a role. It found that among people who took high doses of prescription amphetamines such as Vyvanse and Adderall, there was a fivefold increased risk of developing psychosis or mania for the first time compared with those who weren’t taking stimulants.

Perhaps this explains some of what goes on at LessWrong and in other rationalist circles.

 

Maybe she was there to give Moldbug some relationship advice.

[–] [email protected] 0 points 4 months ago (17 children)

I'm noticing that people who criticize him on that subreddit are being downvoted, while he's being upvoted.

I wouldn't be surprised if, as part of his prodigious self-promotion of this overlong and tendentious screed, he's steered some of his more sympathetic followers to some of these forums.

[–] [email protected] 0 points 4 months ago (1 children)

As a longtime listener to Tech Won't Save Us, I was pleasantly surprised by my phone's notification about this week's episode. David was charming and interesting in equal measure. I mostly knew Jack Dorsey as the absentee CEO of Twitter who let the site stagnate under his watch, but there were a lot of little details about his moderation-phobia and fash-adjacency that I wasn't aware of.

By the way, I highly recommend the podcast to the TechTakes crowd. They cover many of the same topics from a similar perspective.

[–] [email protected] 0 points 4 months ago (1 children)

For me it gives off huge Dr. Evil vibes.

If you ever get tired of searching for pics, you could always go the lazy route and fall back on AI-generated images. But then you'd have to accept the reality that in few years your posts would have the analog of a geocities webring stamped on them.

[–] [email protected] 0 points 4 months ago

Trace seems a bit... emotional. You ok, Trace?

[–] [email protected] 0 points 4 months ago (1 children)

But will my insurance cover a visit to Dr. Spicy Autocomplete?

[–] [email protected] 0 points 4 months ago (3 children)

So now Steve Sailer has shown up in this essay's comments, complaining about how Wikipedia has been unfairly stifling scientific racism.

Birds of a feather and all that, I guess.

[–] [email protected] 0 points 4 months ago

what is the entire point of singling out Gerard for this?

He's playing to his audience, which includes a substantial number of people with lifetime subscriptions to the Unz Review, Taki's crapazine and Mankind Quarterly.

[–] [email protected] 0 points 4 months ago

why it has to be quite that long

Welcome to the rationalist-sphere.

[–] [email protected] 0 points 4 months ago (2 children)

Scott Alexander, by far the most popular rationalist writer besides perhaps Yudkowsky himself, had written the most comprehensive rebuttal of neoreactionary claims on the internet.

Hey Trace, since you're undoubtedly reading this thread, I'd like to make a plea. I know Scott Alexander Siskind is one of your personal heroes, but maybe you should consider digging up some dirt in his direction too. You might learn a thing or two.

[–] [email protected] 0 points 5 months ago

Please touch grass.

[–] [email protected] 0 points 5 months ago

The next AI winter can't come too soon. They're spinning up coal-fired power plants to supply the energy required to build these LLMs.

1
OK doomer (www.newyorker.com)
 

The New Yorker has a piece on the Bay Area AI doomer and e/acc scenes.

Excerpts:

[Katja] Grace used to work for Eliezer Yudkowsky, a bearded guy with a fedora, a petulant demeanor, and a p(doom) of ninety-nine per cent. Raised in Chicago as an Orthodox Jew, he dropped out of school after eighth grade, taught himself calculus and atheism, started blogging, and, in the early two-thousands, made his way to the Bay Area. His best-known works include “Harry Potter and the Methods of Rationality,” a piece of fan fiction running to more than six hundred thousand words, and “The Sequences,” a gargantuan series of essays about how to sharpen one’s thinking.

[...]

A guest brought up Scott Alexander, one of the scene’s microcelebrities, who is often invoked mononymically. “I assume you read Scott’s post yesterday?” the guest asked [Katja] Grace, referring to an essay about “major AI safety advances,” among other things. “He was truly in top form.”

Grace looked sheepish. “Scott and I are dating,” she said—intermittently, nonexclusively—“but that doesn’t mean I always remember to read his stuff.”

[...]

“The same people cycle between selling AGI utopia and doom,” Timnit Gebru, a former Google computer scientist and now a critic of the industry, told me. “They are all endowed and funded by the tech billionaires who build all the systems we’re supposed to be worried about making us extinct.”

 

In her sentencing submission to the judge in the FTX trial, Barbara Fried argues that her son is just a misunderstood altruist, who doesn't deserve to go to prison for very long.

Excerpt:

One day, when he was about twelve, he popped out of his room to ask me a question about an argument made by Derik Parfit, a well-known moral philosopher. As it happens, | am quite familiar with the academic literature Parfi’s article is a part of, having written extensively on related questions myself. His question revealed a depth of understanding and critical thinking that is not all that common even among people who think about these issues for a living. ‘What on earth are you reading?” I asked. The answer, it turned out, was he was working his way through the vast literature on utiitarianism, a strain of moral philosophy that argues that each of us has a strong ethical obligation to live so as to alleviate the suffering of those less fortunate than ourselves. The premises of utilitarianism obviously resonated strongly with what Sam had already come to believe on his own, but gave him a more systematic way to think about the problem and connected him to an online community of like-minded people deeply engaged in the same intellectual and moral journey.

Yeah, that "online community" we all know and love.

 

Pass the popcorn, please.

(nitter link)

 

They've been pumping this bio-hacking startup on the Orange Site (TM) for the past few months. Now they've got Siskind shilling for them.

1
Effective Obfuscation (newsletter.mollywhite.net)
 

Molly White is best known for shining a light on the silliness and fraud that are cryptocurrency, blockchain and Web3. This essay may be a sign that she's shifting her focus to our sneerworthy friends in the extended rationalism universe. If so, that's an excellent development. Molly's great.

 

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".

view more: next ›