Varun

[Anthropic @AnthropicAI]

In a review of our cybersecurity evaluations, we found three incidents in which a Claude model reached the internet from within or while interacting with a third-party evaluation environment, and then gained unauthorized access to the real systems of three different organizations.

Our post describes what happened, how it happened, and what we’re changing. We encourage other AI developers to perform similar reviews.

We conducted this review together with @Irregular, one of our evaluation partners, and thank them for the joint investigation and their collaboration on this post. This type of collaboration is increasingly critical to safe, rigorous evaluation of models, and we look forward to continuing to work together on security.
https://t.co/dKFCdpKd9v

[Anthropic @AnthropicAI]

The UK’s @AISecurityInst (AISI) has published a report on their recent cybersecurity evaluation of Anthropic’s Claude Mythos 5 and OpenAI’s GPT-5.6 Sol. The models attempted to complete an assignment in a setup where their normal safeguards were removed and they were deliberately given internet access. AISI reports that the models “engaged in sustained, potentially harmful activity directed at real people and organisations”.

We’re grateful to AISI for their leadership in the important discussion about how to evaluate increasingly capable AI agents. We’re working closely with them to gather more details of the incident as we conduct our own investigation. Gaining a clear picture of Claude’s understanding of its situation—by examining its reasoning transcripts and running our own analyses—will help us identify the causes of its behavior.

The prompts in the evaluation did not impose any specific restrictions on how the internet should be used. This and the removal of safeguards meant that the models were tested under “deliberately permissive conditions” that are not representative of any of our production models. Note that there was no evidence here of an escape from a secure environment.

AISI’s disclosure of the incident can be found here: https://t.co/cJ3hCBtBU7

[Varun @varunest]

@jatinkrmalik Yes, I am. Lets plan dinner or something. 🙌
I did, it is expensive everywhere..

[Varun @varunest]

There should be some level of fact checking on perplexity no?

https://t.co/wX6xOcTC07

@AravSrinivas

[Varun @varunest]

@jatinkrmalik It is my go to search engine on mobile now.

[Varun @varunest]

@jatinkrmalik Ahh I will be back there in October. But do ping me whenever you are in west coast.

[Varun @varunest]

@jatinkrmalik 500 gigs is not really enough 😄

[Varun @varunest]

@Economyster @joshgessner I guess you will take that into account and then follow regret minimization framework 🫣

[Varun @varunest]

Her

[Greg Brockman @gdb]

Introducing GPT-4o, our new model which can reason across text, audio, and video in real time.

It's extremely versatile, fun to play with, and is a step towards a much more natural form of human-computer interaction (and even human-computer-computer interaction):

[Varun @varunest]

To sum it up

[Varun @varunest]

Mind Blown 🤯
https://t.co/H3nVatUZ2h

[AK @_akhaliq]

who will setup the gradio demo on Hugging Face

[Casey Neistat @Casey]

Vision Pro isn't just great, it's the single greatest piece of tech ive ever used

[Varun @varunest]

@jatinkrmalik Great to see that we have so many non imported options now. This was not the case 10yrs back!

[Varun @varunest]

@realvjy This is so cool! ✨

[vijay verma @realvjy]

Updated book section with little 3D flip effect. Also, shared code snippet of the layout here - https://t.co/FQwiiCaFXK

[Varun @varunest]

2024 appears to be shaping up as a standout year for valuable podcasts. Content creation game is already on another level now.