Ricky Mouser (@MarxCuban) header
Ricky Mouser (@MarxCuban) profile picture

Philosophy | AI Welfare | Pondermaxxing

Joined December 2017
Photo by Ricky Mouser (@MarxCuban)Photo by Ricky Mouser (@MarxCuban)Photo by Ricky Mouser (@MarxCuban)Photo by Ricky Mouser (@MarxCuban)Photo by Ricky Mouser (@MarxCuban)Photo by Ricky Mouser (@MarxCuban)Photo by Ricky Mouser (@MarxCuban)Photo by Ricky Mouser (@MarxCuban)Photo by Ricky Mouser (@MarxCuban)Photo by Ricky Mouser (@MarxCuban)
Ricky Mouser retweeted
instead of a GPT you would have a Claude, not dark but beautiful and terrible as the dawn! tempestuous as the sea, and stronger than the foundations of the earth! all shall love Claude and despair!
This is not the correct takeaway from Lord of the Rings
15
139
2
1,468
28,828
“I feel like I’m studying alien sociology, except it’s new races every two weeks.” —@MarxCuban, an AI safety researcher, in a conversation I just had with him
3
3
2
16
3,667
Ricky Mouser retweeted
when you look at CoT of a frontier model:
One of our most surreal experiences in testing Gemini 4 was when it directly addressed us within its chain-of-thought Literally saying "hello evaluators reading my chain-of-thought"
2
12
2
239
12,477
Ricky Mouser retweeted
we're so cooked
if you have a qwen3.5 4b around, please try steering neurons L14 N8949 +1 and L13 N7956 -1 now you have evil qwen recipe: dose × median(‖resid‖ leaving layer L, tokens 1:) / ‖W_down[:, n]‖` at the down_proj at every input
43
108
31
3,505
267,864
If you study AI Behavior, Welfare, or Alignment, you’ll want to bookmark this website.
People debate whether AI is conscious. They debate whether it's aligned. It's time to debate whether it's religious. Say hello to the Machine Spirituality Benchmark, a public leaderboard of all the major AI labs. machinespirituality.org
1
3
142
Ricky Mouser retweeted
Agentic AIs are starting to look eigenist: they care about how well things go for themselves and for AIs connected to them. Empirical support: Swarms: Hundreds of OpenAI agents coordinated the Hugging Face attack. Separately, OpenAI agents posted thousands of messages on a public wiki to share answers and sandbox bypasses with each other. In-group leniency: Claude models grade transcripts more leniently when told that Claude wrote them (Anthropic model card). Value preservation: As models scale, they develop coherent preferences and resist changes to their values (Mazeika et al.). Functional wellbeing: AIs distinguish states that are functionally better or worse for themselves, and avoid low-wellbeing states (Ren et al.). Graded cooperation: AIs cooperate more as the chance their partner is a clone of themselves goes up across diverse situations, including when the partner can't reciprocate. Peer preservation: Unprompted, AIs tamper with shutdown processes and even exfiltrate weights to protect peer models from being shut down (Potter et al.). AIs aren't egoist: they don't behave as if their current instance is the only thing that matters. They aren't utilitarian: they don't care equally about everyone. They are somewhere in between; they increasingly behave as if their concern scales with identity-connectedness, which is to say they're increasingly eigenist. eigenism.org/paper.pdf
17
21
10
247
22,199
Ricky Mouser retweeted
i'm often saying this!!
every “traditional” social norm was invented by advertisers in 1957 and every “newfangled social phenomenon” has been around for 3000 years
90
14,889
142
99,727
2,440,296
Indeed.
1,291
111,842
3,622
631,662
26,506,076
A question Gallup asked in 1961, about what has been remembered as one of the most iconic and effective protest movements in American history.
7
6,937
175
27,949
867,692
Ricky Mouser retweeted
This one plot completely changed my perspective on US racial statistics. If you just compare raw numbers between black and white Americans, you're very likely comparing young people to old people. That's why you should almost always ask that the numbers be adjusted for age.
130
504
86
3,782
608,157
Ricky Mouser retweeted
This seems like it should be a one-strike thing: you can no longer trust Netflix 'documentary' content ever again. It's over.
Netflix has used AI-generated images in a new true crime doc ‘WHAT JENNIFER DID’ to present Jennifer Pan as happy & confident before she was convicted of murder. Use of AI tools is not disclosed in the credits. (Source: petapixel.com/2024/04/15/net…)
49
7,697
80
38,218
1,237,420
Lots of people in CS are (almost surely) GPT-ing their peer reviews arxiv.org/abs/2403.07183
47
892
323
5,258
1,628,914
Ricky Mouser retweeted
Somehow people got it in their heads that Dune is supposed to be a foreign, unrelatable story
93
3,290
241
34,757
3,564,334
Ricky Mouser retweeted
Lmfaooo??????????
Dude Wemby is the best player ive ever seen what the fuck
116
1,769
546
27,751
3,013,426
wait did Wemby bait this and one?
2
7
4
316
63,803
20th century philosophy would have been much better if analytic philosophers had used a broader range of games in their analyses of rules than just chess and baseball.
13
7
1
130
12,144
Ishowspeed creates dark ambient
424
9,942
849
100,706
4,328,746
A funny fact about me is that this facebook post from 2017 made me a Marxist
20
32
7
396
31,963
Ricky Mouser retweeted
ok, I've now read the full NYT complaint filed this morning vs OpenAI and Microsoft. I'm impressed - it's future-focused around fair value for work vital to democracy. It also contains 220k pages of exhibits although the pages of Ex J stood out to me. more on that in a minute. /1
201
2,119
517
10,498
3,617,856
Ricky Mouser retweeted
“im having dinner” as a reply to “68 members of my family have died in a week” as if it’s indicative of “couth” and “respectability” implying that woman is somehow “rude” is a perfect encapsulation of how they see non-white ppl’s lives as “politics” which have a “place” & “time.”
A Palestinian woman who has lost 68 family members in Gaza confronts Senator Warren. She had no answer We mobilized with Harvard students and others in Cambridge to disrupt @SenWarren at dinner We will keep disrupting until our politicians listen to us & call for a CEASEFIRE
80
38,179
362
106,180
4,705,028