dgerard

joined 2 years ago
MODERATOR OF
techtakes
buttcoin
sneerclub
sorted by: new top controversial old
Vibe nuclear — let’s use AI shortcuts on reactor safety! in c/techtakes@awful.systems
[–] dgerard@awful.systems 6 points 19 hours ago

That's a great question! If it's for an AI power station, it turns out that 3.6 roentgen is a nutritional requirement.

Is there anything else I can help you with?

  • Health benefits of radioactive glue on pizza
  • Why South Memphis needs a nuclear reactor built in 6 months
  • Why you need an anime waifu who glows in the dark
Vibe nuclear — let’s use AI shortcuts on reactor safety! in c/techtakes@awful.systems
[–] dgerard@awful.systems 3 points 23 hours ago (2 children)

plutonium, looks like:

US offers nuclear energy companies access to weapons-grade plutonium - Oct 21st https://www.ft.com/content/2fbbc621-405e-4a29-850c-f0079b116216 https://archive.is/Pc949

The Department of Energy on Tuesday published an application that nuclear energy groups can use to seek up to 19 metric tonnes of the government’s weapons-grade plutonium from cold war-era warheads.

...

At least two companies, Oklo, which is backed by OpenAI’s Sam Altman, and France’s Newcleo, are expected to apply to access the government’s plutonium stockpile.

may I just say:

JESUS FUCKING CHRIST

However, experts have raised concerns about the commercial use of plutonium and the risk of the material falling into the wrong hands.

NO SHIT

29
Whoops! Microsoft’s new Windows AI agent platform lets in malware (pivot-to-ai.com)
 

and you thought Windows was supposed to run software

https://www.youtube.com/watch?v=tAeNFkZJVxc&list=UU9rJrMVgcXTfa8xuMnbhAEA - video
https://pivottoai.libsyn.com/20251119-whoops-microsofts-windows-ai-agent-platform-lets-in-malware - podcast

time: 4 min 27 sec

IQ maximalism is going great - ooops, turns out Raven's Progressive Matrices was just a personality test in c/sneerclub@awful.systems
[–] dgerard@awful.systems 6 points 1 day ago

he's already done "don't worry about your <100 score, IQ tests are meant to show black people are dumb"

Vibe nuclear — let’s use AI shortcuts on reactor safety! in c/techtakes@awful.systems
[–] dgerard@awful.systems 2 points 1 day ago

I'm sure there will be no issues setting up nuclear fuel handling at ports worldwide. Well, maybe one or two.

Vibe nuclear — let’s use AI shortcuts on reactor safety! in c/techtakes@awful.systems
[–] dgerard@awful.systems 2 points 1 day ago (4 children)

Trump was ready to give some Sam Altman project highly enriched uranium, though I'm not clear on whether that was 20% (already considered a serious proliferation risk) or full bomb-grade 95%.

Vibe nuclear — let’s use AI shortcuts on reactor safety! in c/techtakes@awful.systems
[–] dgerard@awful.systems 3 points 1 day ago (8 children)

helps to be the US Navy and not be worried about costs

ship-sized SMR power is quite expensive!

Vibe nuclear — let’s use AI shortcuts on reactor safety! in c/techtakes@awful.systems
[–] dgerard@awful.systems 6 points 1 day ago (2 children)

yeah, even the green case for nuclear - which has been around for a long time - falters on wind and solar with battery just being hilariously cheaper. At this point the funding problem is interconnects.

On Incomputable Language: An Essay on AI by Elizabeth Sandifer in c/sneerclub@awful.systems
[–] dgerard@awful.systems 4 points 1 day ago (1 children)

ah, but it was a very defective essay if you don't bother reading it

Stubsack: weekly thread for sneers not worth an entire post, week ending 23rd November 2025 - awful.systems in c/techtakes@awful.systems
[–] dgerard@awful.systems 6 points 1 day ago

I posit the Culture having the Californian accents, because they're liberal colonialism writ large but they're nice about it, until they aren't

Stubsack: weekly thread for sneers not worth an entire post, week ending 23rd November 2025 - awful.systems in c/techtakes@awful.systems
[–] dgerard@awful.systems 5 points 1 day ago

christ these people are so bad at reading things

Bezos loves the Culture too, and you can be sure he and Musk identify with the Minds

Stubsack: weekly thread for sneers not worth an entire post, week ending 23rd November 2025 - awful.systems in c/techtakes@awful.systems
[–] dgerard@awful.systems 4 points 1 day ago (3 children)

subtlety is wasted on most readers

12
IQ maximalism is going great - ooops, turns out Raven's Progressive Matrices was just a personality test (pmc.ncbi.nlm.nih.gov)
Vibe nuclear — let’s use AI shortcuts on reactor safety! in c/techtakes@awful.systems
[–] dgerard@awful.systems 4 points 1 day ago

have you ever met actual anti nuclear activists

35
Vibe nuclear — let’s use AI shortcuts on reactor safety! (pivot-to-ai.com)
 

‘GenAI for Nuclear Licensing’

https://www.youtube.com/watch?v=PW9lusiwMz8&list=UU9rJrMVgcXTfa8xuMnbhAEA - video
https://pivottoai.libsyn.com/20251118-vibe-nuclear-lets-use-ai-shortcuts-on-reactor-safety - podcast

time: 6 min 24 sec

78
Data Center Watch worries that anti-AI activism is working (pivot-to-ai.com)
 

making AI haters seem very cool and effective

(send this to an AI data centre hater you know!)

https://www.youtube.com/watch?v=nPbTXEOJEVM&list=UU9rJrMVgcXTfa8xuMnbhAEA - video
https://pivottoai.libsyn.com/20251117-data-center-watch-worries-anti-ai-activism-is-working - podcast

time: 6 min 29 sec

28
On Incomputable Language: An Essay on AI by Elizabeth Sandifer (www.eruditorumpress.com)
 

much more sneerclub than techtakes

28
Yudkowsky denies the accusations! several thousand words in, and ten years after they were made (x.com)
 

yes, that's his high-volume account, linked from @ESYudkowsky

18
How Lightcone fucked up returning an FTX donation: they asked a chatbot (awful.systems)
submitted 5 days ago* (last edited 5 days ago) by dgerard@awful.systems to c/buttcoin@awful.systems
14 comments fedilink
 

This could have been buttcoin, sneerclub or techtakes, but y'know pick one.

‪Lightcone is the rationalist org that runs LessWrong these days and runs totally not race science conferences.

SBF donated $6m to Lightcone. Then FTX in ch 11 wrote demanding a clawback (cos it was stolen customer funds). Lightcone ignored it. This went badly for Lightcone. They had to sell a building and their funding has dried right up.

this is Oliver Habryka, CEO of Lightcone:

"Oh man, we are being sued by FTX. I should hire a bankruptcy lawyer to prepare our defense." WRONG. I mean yes, of course go ahead and hire a bankruptcy lawyer to prepare the defense. But in-parallel use language models to prepare a defense yourself, then run the defense by the lawyer you hired until you think you understand reasonably well what the core constraints are. Then work together with the bankruptcy lawyer on the defense.

Habryka wrote this a few years after Lightcone completely fucked this one. Well done, Oliver.

https://www.lesswrong.com/s/cegJvyiP2dBkhDh2K/p/rSCxviHtiWrG5pudv

32
Habryka posts a NEW OFFICIAL LESSWRONG ENEMIES LIST. Guess who's #1, go on, guess (awful.systems)
submitted 5 days ago* (last edited 5 days ago) by dgerard@awful.systems to c/sneerclub@awful.systems
40 comments fedilink
 

https://www.lesswrong.com/posts/Hun4EaiSQnNmB9xkd/tell-people-as-early-as-possible-it-s-not-going-to-work-out

archive: https://archive.is/NSVXR

Oliver wrote an internal Lightcone Infrastructure memo that lists the top enemies of the Rationality movement. He saw fit to post his Enemies List to the site, because that's a very normal thing to do.

no. 2 is a neoreactionary troll who ran a downvote bot in 2013-2014.

Emile Torres is only #3, sorry Emile some of us are just better at increasing existential risk

no. 4 is Ziz. I am officially considered worse than the literally murderous death cult.

what can i say some of us have just got it

also I trounce complete pikers like (checks notes) Peter Thiel

LessWrong used to call themselves a "phyg" in the hope that the word "cult" would not show up in Google so much as being associated with them

16
Anthropic: Chinese AI hackers are after you! Security researchers call BS (pivot-to-ai.com)
 

Be VERY SCARED, okay?

https://www.youtube.com/watch?v=UTbyGFW0new&list=UU9rJrMVgcXTfa8xuMnbhAEA - video
https://pivottoai.libsyn.com/20251114-anthropic-claims-chinese-ai-hackers-security-researchers-call-bs - podcast

time: 5 min 40 sec

77
No, fake AI music bought onto a minor chart is not actually popular (pivot-to-ai.com)
 

the market for AI bands is gullible journalists

https://www.youtube.com/watch?v=jrvenVdkOYA&list=UU9rJrMVgcXTfa8xuMnbhAEA - video
https://pivottoai.libsyn.com/20251113-no-fake-ai-music-bought-onto-a-minor-chart-is-not-popular - podcast

time: 6 min 02 sec

32
GEMA wins against OpenAI on copying song lyrics (pivot-to-ai.com)
 

yo ho ho, rum, etc

https://www.youtube.com/watch?v=JPTYrU73Z9o&list=UU9rJrMVgcXTfa8xuMnbhAEA - video
https://pivottoai.libsyn.com/20251112-gema-wins-against-openai-on-copying-song-lyrics -podcast

time: 4 min 47 sec

19
Getty Images loses UK suit against Stable Diffusion (pivot-to-ai.com)
 

Getty insists this is a win, actually

https://www.youtube.com/watch?v=ZWctD7wl1Xg&list=UU9rJrMVgcXTfa8xuMnbhAEA - video
https://pivottoai.libsyn.com/20251111-getty-images-loses-uk-suit-against-stable-diffusion - podcast

time: 4 min 26 sec

view more: next ›