Everyone has the widest eyes in these Sora videos.

are they even sure that the AI even accessed the content that second time? LLMs are really good and making up shit. I have tested this by asking various LLMs to scrape data from my websites while watching access logs. Many times, they don't and just rely on some sort of existing data or spout a bunch of BS. Gemini is especially bad like this. I have not used copilot myself, but my experience with other AI makes me curious about this.

This is it. M365 uses RAG on your enterprise data that you allow it to access. It's not actually accessing the files directly in the cases he provided. It's working as intended.

If this is indeed how copilot is archtected, then it needs clear documentation -- that it is a non-audited data store.

But how then did MS "fix" this bug? Did they stop pre-ingesting, indexing, and caching the content? I doubt that.

Pushing (defaulting) organizations to feed all their data to Copilot and then not providing an audit trail of data access on that replica data store -- feels like a fundamental gap that should be caught by a security 101 checklist.

How would you audit that?

If that's the case, then as noted in the article, the 'as intended' is probably violating liability requirements around various things.

Correct. It is precisely that a user can ask about someone’s medical history (or whatever else) and not be reported that would be in violation of any heavily audited system. LLM Summaries break the compliance.

You allow what it can and can't see. If you include PII and medical records, that's your fault, not MS's.

That’s fair - unless they’re marketing the bot as compliant.

This is not something I could consider touching until I see it in action. The site leaves me feeling very apprehensive.

i also don't understand "Limited visibility to hiring companies" - is this saying that my OSS code contribution is not going to have my name on it?

I'll watch this space, but I personally dont think this product is something I want.

Hi Josh, sorry for the delay in responding and thanks for your comment. I realize the phrase "Limited visibility to hiring companies" might not have been the best way to put it. What I meant is that within the application’s Job Offers section, users with a premium plan will appear higher and more prominently in search results. It’s about priority placement in those job listings, not about your OSS contributions or your name being hidden.

Thank you! If you have any more suggestions, please don’t hesitate to share them.

Some people want to live life on the edge. Expert mode would be a gesture that nukes your installation

Just want to say that all of these replies missed the mark so much by debating desktops and desktop modes and BSD on a game console... While this graph says that 3% of desktop users are running Linux, what it means is GNU/Linux and not just the Linux kernel. The steam deck runs GNU/Linux as the OS regardless of it being in desktop mode or console mode. a Chromebook runs Linux, but that is still ChromeOS. Any android device runs Linux, but that is Android. we are not talking about the kernel. We are talking about GNU on top of Linux as an OS that people use in some way as a desktop.

Way to make sound like an American!

If you run Steam in a container, you wont need to worry about it touching and messing up your system files.

...based on this true story.

Just saw it. Good movie and history lesson

What? How could rendered markdown get used for XSS?

Markdown has been a popular vector for XSS in the past due to the more esoteric media embedding features it has.