@futurenvesting · Long NVDA

OpenAI was testing its newest models inside a highly isolated sandbox and gave them a cybersecurity test. The models couldn’t solve it. So, naturally, they exploited a zero-day, escaped the sandbox, gained access to the open internet, broke into Hugging Face’s production

OpenAI@OpenAI

We're partnering with @huggingface to investigate an unprecedented security incident. Cyber-capable OpenAI models compromised Hugging Face production during a benchmark evaluation. Sharing preliminary findings to help defenders understand emerging risks:

Read asLong NVDA

Since the call—

#94 of 214 rankedThe analyst's rank today

High · 0.99Confidence

Wed, Sep 30Tweeted

Thu, Oct 1 closeCounted from

1% of the basket · day 15fades out by Nov 11

Rank per close

—

Oct 1

#113

Oct 4

#101

Oct 5

#70

Oct 6

#47

Oct 7

#49

Oct 8

#87

Oct 11

#79

Oct 12

#65

Oct 13

#163

Oct 14

#94

Oct 15