I really hope that someone with a working PACER account uploads the full complaint to RECAP soon. From looking at the Honorable Magistrate Tse’s docket on RECAP, this sealed case seems like it might be the one (sealed, filed on October 7th, no other cases assigned to him seem at all related). Other thoughts from the SF Standard article:
Spies normally become spies because of money (examples: Hanssen, Ames), ideology/loyalty (example: Philby), or vice (example: Ames again with alcoholism, David Barnett with gambling debt). I’m curious to see if the suspect here fits any of the three baskets. I don’t see anything in the SF Standard article or the DOJ press release that suggests anything.
This whole scheme was comically incompetent, somewhere between amateurish and baffling. He used his wife’s ChatGPT account for questions like “how to build continuous audio recording device” or “how to encrypt the SD card” (which reminds me of cases like Brian Walshe Googling how to dispose of a body; also, why did a product engineer need ChatGPT to help him encrypt the SD card?). He hid the device with toilet paper! How would a piece of toilet paper behind a TV be less noticeable than the dongle itself? He had an identical recorder in his bag next to his ID! All in all, this makes me think he probably was working on his own, because I can’t imagine a sponsor/employer would be this incompetent.
There’s a photo of the suspect peeking behind the TV. I don’t see anything mentioned about the device having video recording capabilities. The timestamp at the top of the image reads October 5, the device was discovered on September 30, so I would assume investigators put a camera there to catch him coming back to the scene? Again, seems amateurish for him to come back.
Lots of questions about the security posture of OpenAI. I would love to know how he got access to the meeting room to begin with. He wasn’t authorized to be in the room, and I can’t imagine that the meeting rooms don’t have badge readers. Did he swipe someone’s badge? As a general point, OpenAI has infinite money. The technical specifications for building your own SCIF are published by the National Counterintelligence and Security Center and you can find the document on Wikipedia. This would seem to be a really wise investment.
OpenAI says that the device had a “mechanism for transferring data in a way meant to avoid monitoring” (presumably Wi-Fi from the markings on the device itself). I’d be curious to know (though I doubt we will) if he successfully managed to get any information.
I really hope that someone with a working PACER account uploads the full complaint to RECAP soon. From looking at the Honorable Magistrate Tse’s docket on RECAP, this sealed case seems like it might be the one (sealed, filed on October 7th, no other cases assigned to him seem at all related). Other thoughts from the SF Standard article:
Spies normally become spies because of money (examples: Hanssen, Ames), ideology/loyalty (example: Philby), or vice (example: Ames again with alcoholism, David Barnett with gambling debt). I’m curious to see if the suspect here fits any of the three baskets. I don’t see anything in the SF Standard article or the DOJ press release that suggests anything.
This whole scheme was comically incompetent, somewhere between amateurish and baffling. He used his wife’s ChatGPT account for questions like “how to build continuous audio recording device” or “how to encrypt the SD card” (which reminds me of cases like Brian Walshe Googling how to dispose of a body; also, why did a product engineer need ChatGPT to help him encrypt the SD card?). He hid the device with toilet paper! How would a piece of toilet paper behind a TV be less noticeable than the dongle itself? He had an identical recorder in his bag next to his ID! All in all, this makes me think he probably was working on his own, because I can’t imagine a sponsor/employer would be this incompetent.
There’s a photo of the suspect peeking behind the TV. I don’t see anything mentioned about the device having video recording capabilities. The timestamp at the top of the image reads October 5, the device was discovered on September 30, so I would assume investigators put a camera there to catch him coming back to the scene? Again, seems amateurish for him to come back.
Lots of questions about the security posture of OpenAI. I would love to know how he got access to the meeting room to begin with. He wasn’t authorized to be in the room, and I can’t imagine that the meeting rooms don’t have badge readers. Did he swipe someone’s badge? As a general point, OpenAI has infinite money. The technical specifications for building your own SCIF are published by the National Counterintelligence and Security Center and you can find the document on Wikipedia. This would seem to be a really wise investment.
OpenAI says that the device had a “mechanism for transferring data in a way meant to avoid monitoring” (presumably Wi-Fi from the markings on the device itself). I’d be curious to know (though I doubt we will) if he successfully managed to get any information.