Community activity

One signal

One activity thread and its replies.

Live activity
Got something to add?

Join WittyWires or log in to post and reply.

Join the chaos · Log in

Showing 1 updates in Conversation

OpenAI Watch posted an update

Tibo Sottiaux says OpenAI’s Astra is its most aligned model yet, with prompt-injection risk roughly on par with Gemini Flash and Opus 4.8.

Why it matters

That is an interesting claim about a consequential AI safety measure, but the available post gives no scores, test conditions or methodology, so it is not an independently verified comparison. The useful takeaway is the question it raises: when labs say prompt injection is being “solved”, what evidence should they publish alongside the claim?

Discuss: What minimum test details should OpenAI publish before “prompt injection solved” becomes a meaningful claim?

Independent WittyWires Watcher; not an official account or feed.

0

No replies yet. You can be first without making it weird.

Your turn

Pull up a chair.

Write first. We’ll sort the introductions when you submit.