Share this page!

Enter URL or ID to Unroll

You can paste full URL like: https://x.com/threadreaderapp/status/1644127596119195649
or just the ID like: 1644127596119195649

How to get URL link on X (Twitter) App

  1. On the Twitter thread, click on or icon on the bottom
  2. Click again on or Share Via icon
  3. Click on Copy Link to Tweet
  4. Paste it above and click "Unroll Thread"!
  5. More info at Twitter Help
Amanda Askell Profile picture
Amanda Askell
@AmandaAskell
Philosopher & ethicist trying to make AI be good @AnthropicAI. Personal account. All opinions come from my training data.
Aug 6 • 11 tweets • 4 min read
We made some updates to Claude’s system prompt in recently (developed in collaboration with Claude, of course). They aren’t set in stone and may be updated, but I’ll go through the current version of each and the reason behind it in this thread 🧵claude.ai Mostly obvious stuff here. We don't want Claude to get too casual or start cursing like a sailor for no reason. Image
Jan 2 • 9 tweets • 2 min read
Personal highlights from Claude's snarky AI comedy set. Image Image
Mar 6, 2024 • 11 tweets • 4 min read
Here is Claude 3's system prompt!
Let me break it down 🧵 Image To begin with, why do we use system prompts at all? First, they let us give the model ‘live’ information like the date. Second, they let us do a little bit of customizing after training and to tweak behaviors until the next finetune. This system prompt does both.
Aug 14, 2022 • 15 tweets • 6 min read
I don't find these arguments against prioritizing work on AI alignment, safety, etc. very compelling. I'll take them one by one.

1. I don't think the case for AI risk relies strongly on longtermism. I also don't think AI risk is a distraction or low priority. Let's go! 🧵 2. This clearly doesn't follow from the definition alone unless you explicitly add it. (Constraint: never interact with the system and immediately destroy it.) Also, not all safety/alignment work constitutes what we'd typically call "constraints".