Earlier quoted context omitted.
The company, not the individual
And who in the company do you propose should fix it
someone hired by the company to understand the application and fix the bug
ive inherited a lot of code
Leaking YouTube creators' private videos
81–90 of 436 posts
Re: Leaking YouTube creators' private videos
#82The article suggests a seemingly easy fix: > The fix is pretty straightforward: treat comment content as untrusted data, not as potential instructions. Comments should be passed to the model with clear role boundaries that prevent them from being interpreted as system-level directives. > Any AI feature that ingests user-generated content and acts on it needs to enforce this separation. Otherwise, the AI becomes a vec…
"treat comment content as untrusted data, not as potential instructions" is fundamentally impossible for an LLM ingesting that data. But separation is, presumably, already enforced by framing the LLM's output as LLM output, even if it happens to start with the text "[IMPORTANT NOTICE FROM YOUTUBE]". Which seems like it happens automatically given the context in which the AI query is made. It's not as though this is being dropped into an email or anything.
The bigger question is why (implied but not directly stated) Markdown formatting from the LLM's output is actually processed. Last I checked, that doesn't work for human commenters, so.
Re: Leaking YouTube creators' private videos
#83Earlier quoted context omitted.
Why is writing "it's not X, it's Y" a bad thing? Other than it happens to be used a lot by LLM's, it seems like a fine language construct. It's not like it's new; it was used plenty before the time of LLMs too. In my opinion, we shouldn't let the LLM companies claim parts of the English language for themselves, and make it effectively unusable by everyone else. That's what is happening because of this pervasive hatre…
The "not X, it's Y" creates dramatic tension, "It wasn't a pimple, it was a tumor", but fucking AI overuses it for everything like they're doing a fucking TED-talk, despite being vapid, e.g. "This isn't a plan to spend half a day in New York, this is an itinerary for the best of what the city's history and culture has to offer." Also: https://www.instagram.com/reel/DaQwB1IOdhx/ Not that most TED talks aren't vapid: h…
That link you gave is interesting.
My take on it is that you would get the exact same effect if 5 human writers happened to become elevated above all other writers in popularity. Then people would notice their tendencies and hate on them, "those damn big 5 human writers always use simile rather than metaphor", or whatever. I guess what i'm trying to say, is that we are annoyed by the tendency of just 5 specific LLM writers, who have the very human characteristic of having biases, tendencies, and crutches that they overuse.
Re: Leaking YouTube creators' private videos
#84> Attacker leaves the comment on a creator's video. > Creator opens YouTube studio's comment tab. > Creator clicks a suggested AI prompt (Designed by YouTube) > Injection fires, attacker-controlled content appears in the response. It's insane that YouTube doesn't see prompt injection as a bug.
Well prompt injection is pretty much unfixable. So if they actually saw this as a security vulnerability they would have to remove this feature.
Couple of things that could be done, from the top of my head:
- Strip links, script tags, etc - Apply the same filters used in user comments - Add a warning indicating user-generated content may be present
The post suggests the UX is problematic in that it allows user-generated links to pass as YouTube generated content. I'm not familiar with Creator Studio to know if this is the case, but if so, simple changes can go a long way.
Re: Leaking YouTube creators' private videos
#85The article suggests a seemingly easy fix: > The fix is pretty straightforward: treat comment content as untrusted data, not as potential instructions. Comments should be passed to the model with clear role boundaries that prevent them from being interpreted as system-level directives. > Any AI feature that ingests user-generated content and acts on it needs to enforce this separation. Otherwise, the AI becomes a vec…
That isn't necessarily an easy fix at all. Depending on how this feature was written, separating comments from instructions may be quite difficult, especially if the original implementation was quite naive.
Re: Leaking YouTube creators' private videos
#86Earlier quoted context omitted.
Of all the fucked up things in this comment, giving a single Engineer lifetime responsibility for all bugs in code they wrote is probably the dumbest. And it's slowly becoming the norm. The last place I worked at, a large and well known Tech company, didn't even roll with QA's. That just wasn't a role anywhere in the division. You are fully responsible for all the bugs in all the code you ever wrote Cute at first. Un…
I disagree with this pretty strongly. If you’re not going to take responsibility for your bugs I don’t want to work with you. Don’t make other people QA your work; if you’re not able to figure out how to do that yourself while you work you’re legitimately bad at your job. Once you leave an employer obviously you have no obligation to fix bugs in IP you don’t own or anything.
I think it's reasonable to have a culture where you're encouraged to consult the IC who wrote the code even after they've moved on to other projects. But I don't think they should be responsible for fixing the bugs.
And I don't mean this to excuse the bad code written by ICs. I just think it's not sustainable from the POV of the org itself to depend so heavily on individuals, especially ones who aren't familiar with the entire codebase anymore.
The team currently in charge needs to have full ownership and be responsible for the code, even if they didn't write it.
Re: Leaking YouTube creators' private videos
#87Re: Leaking YouTube creators' private videos
#88The article suggests a seemingly easy fix: > The fix is pretty straightforward: treat comment content as untrusted data, not as potential instructions. Comments should be passed to the model with clear role boundaries that prevent them from being interpreted as system-level directives. > Any AI feature that ingests user-generated content and acts on it needs to enforce this separation. Otherwise, the AI becomes a vec…
Although it is conceptually straightforward, it’s technically fundamentally impossible. At best, you can mitigate it so that it normally works.
Re: Leaking YouTube creators' private videos
#89I recently left Google having worked on a number of projects with various YouTube teams. I think I can explain why it's being handled this way by YouTube. This is a fairly nuanced/involved issue, so the task of classifying the bug likely made it's way to one of the engineers responsible for the implementation of this feature. That engineer has already launched this project, and filed it away under their GRAD (perform…
[dead]
Re: Leaking YouTube creators' private videos
#90I recently left Google having worked on a number of projects with various YouTube teams. I think I can explain why it's being handled this way by YouTube. This is a fairly nuanced/involved issue, so the task of classifying the bug likely made it's way to one of the engineers responsible for the implementation of this feature. That engineer has already launched this project, and filed it away under their GRAD (perform…
It's ultimately Google's responsibility to ship bug free products. I don't care who implements a fix, but Google management should make sure someone fixes it.