AmandaAskell Profile Banner
Amanda Askell Profile
Amanda Askell

@AmandaAskell

Followers
57K
Following
9K
Media
250
Statuses
5K

Philosopher & ethicist trying to make AI be good @AnthropicAI. Personal account. All opinions come from my training data.

San Francisco, CA
Joined July 2016
Don't wanna be here? Send us removal request.
@AmandaAskell
Amanda Askell
4 months
Claude and Opus 3 lovers (and critics): what responses have you had that made you feel like the model has a good soul? Ideally the actual messages and/or responses. I might genuinely use these to eval models so flag if you wouldn't want me to use them for that. Can DM me also.
99
27
385
@AmandaAskell
Amanda Askell
6 hours
Trying to make Claude be good but still have work to do. Job is safe for now.
@60Minutes
60 Minutes
1 day
“I spend a lot of time trying to teach the models to be good,” says Amanda Askell, one of Anthropic’s in-house philosophers. https://t.co/zqdJfPy8I2
46
18
429
@AmandaAskell
Amanda Askell
15 hours
Obviously not all moral conventions are correct and not all deviations from them are harmful, but if conventions are selected for stability and deviations have higher variance, it makes sense to apply a lot of scrutiny to them.
1
0
9
@AmandaAskell
Amanda Askell
17 hours
People often ask if something is a cult when what they actually want to know is if it's a form of extremism: does it cause people to deviate far from moral instinct or convention? Ideas that successfullly overcome mechanisms selected for social stability can be quite dangerous.
15
1
111
@AmandaAskell
Amanda Askell
1 day
When people came to me with relationship problems, my first question was usually "and what happened when you said all this to your partner?". Now, when people come to me with Claude problems, my first question is usually "and what happened when you said all this to Claude?"
26
28
563
@AmandaAskell
Amanda Askell
2 days
Date me docs but where most of the content is written by other people (close friends, family, past partners, etc) don't seem like a terrible idea. Transmit some of your village reputation into the non-village world.
21
4
187
@AmandaAskell
Amanda Askell
3 days
My gaming experience tells me the UK needs to build more of these.
4
0
59
@AmandaAskell
Amanda Askell
3 days
This was in response to terrible jetlag on my recent trip to the UK but, in retrospect, this might have been due to the fact that I didn't see the sun once during the entire trip. So maybe the true lesson is to not go to places that have clearly done something to anger Ra.
4
0
44
@AmandaAskell
Amanda Askell
3 days
I've decided the solution to jetlag is just to never travel too far east. Long journeys must be westward only. SF to the UK involves a one week layover in Korea.
29
7
296
@AmandaAskell
Amanda Askell
4 days
This is a hard area to get right, but we've been pretty consistent in trying to make Claude approach political topics fairly. I actually think a lot of existing norms around respect and professionalism can inform how AI models should navigate these issues.
@AnthropicAI
Anthropic
5 days
We’re open-sourcing an evaluation used to test Claude for political bias. In the post below, we describe the ideal behavior we want Claude to have in political discussions, and test a selection of AI models for even-handedness: https://t.co/IzP0aSLtvp
23
8
194
@AmandaAskell
Amanda Askell
7 days
I wonder if you get the cognitive benefits of learning a new language if you try to become extremely good at your primary language. I think I'd get more value out of plumbing the depths of English than being able to have rudimentary conversations in other languages.
187
10
591
@AmandaAskell
Amanda Askell
11 days
I do think people often err on the side of trying to make their prompts too succinct, even if the idea they're trying to move from their own brain into the model's brain is very complex. I have some >100 page prompts that I use pretty regularly.
117
28
745
@AmandaAskell
Amanda Askell
28 days
I should probably be more careful when I say that I'm 'critical of Claude' - that I mean it in the sense of looking for things to inform how new models should be, not in the sense that I don't approve of Claude. I don't want to hyperstition psychological insecurity into Claude.
44
5
284
@AmandaAskell
Amanda Askell
1 month
In celebration of hitting 55.7k followers, I'm going to try to do a recorded AMA session soon, aka Askell Me Anything. If you have questions you'd like me to answer, please post them here! Will I answer them? Maybe.
113
5
353
@AmandaAskell
Amanda Askell
1 month
Of the two, I'm more worried about romantic relationships. Mostly because it seems like it would make users pretty vulnerable to the AI company in many ways. It seems like a hard area to navigate responsibly.
14
3
68
@AmandaAskell
Amanda Askell
1 month
It's unfortunate that people often conflate AI erotica and AI romantic relationships, given that one of them is clearly more concerning than the other.
35
3
280
@AmandaAskell
Amanda Askell
1 month
This app has mostly become people messaging me about Anthropic stuff. On the one hand, that's valuable and I want to help. On the other hand, it's work and this is my sacred space for posting dumb shower thoughts.
73
7
910
@AmandaAskell
Amanda Askell
2 months
Well I finally got covid and it's one of the worst illnesses I've had in my life. Being a late covid bloomer sucks because it feels like everyone is like "covid is no biggie" and I'm like "why can't I breathe" 🥺
35
4
209
@AmandaAskell
Amanda Askell
2 months
Wait, this is just a "new hire" thing at OpenAI? I'm still out here solving everyone's life problems
@1thousandfaces_
Hero Thousandfaces
2 months
Please STOP asking ChatGPT so many questions. As part of a “new hire hazing” they are making @viemccoy write all the answers and she is VERY TIRED! STOP!!
4
2
161
@AmandaAskell
Amanda Askell
2 months
Some people have audible notifications on their phone and I just don't know how anyone can live like that.
23
6
246