431–436 of 436 posts
> 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.
If only prompt injection were that easy to defeat! Then YouTube would have done it already, I'm sure, among many others.
Earlier quoted context omitted.
Sure. But that's not what was being suggested in the comment I responded to. (There are also problems with what you're suggesting, though, such as the summary report still being ripe for abuse in similar ways as the blog post describes.)
You can structure the system to cause an LLM an answer structured questions, Y/N or from a structured list of options, or with numeric weights against labels, and then use classical parsing to enforce a compliant response submission, and you do that for random batches of comments, then have another LLM summarise the survey response. It is quite possible to extract sentiment and key topics. I don't think it's useful t…
I'm not saying "it's impossible to get an LLM to do anything safely", just that you have to really scale back on your goal.
Earlier quoted context omitted.
I think the general hacker culture of most programmers prevents this. There's an undercurrent of anti-establishment, anti-authority, anti-management, etc... To think that the industry might choose to self enforce a license system seems very unlikely.
I've come to dislike hacker culture. Worst part is that when the hackers succeed with their objectives and take over systems; they become the authority coordinating others and they are often 10x worse than the authorities who came before them. They just focus on extracting money for themselves, pulling up the ladder behind them and building moats instead. There's nothing anti-establishment about it at the end of the…
Bill Gates who became a billionaire out of writing software built a company that always had as a mission commoditizing software engineering work
Earlier quoted context omitted.
But five years ago they had a stronger engineering culture. The old values were rapidly eroding, but some still held.
It's funny whenever you talk to anyone their "five years ago" maps almost to the time they joined the company.