Johngibb

I am actually asking this question in good faith: are we certain that there's no way to write a useful AI agent that's perfectly defended against injection just like SQL injection is a solved problem?

Is there potentially a way to implement out-of-band signaling in the LLM world, just as we have in telephones (i.e. to prevent phreaking) and SQL (i.e. to prevent SQL injection)? Is there any active research in this area?

We've built ways to demarcate memory as executable or not to effectively transform something in-band (RAM storing instructions and data) to out of band. Could we not do the same with LLMs?

We've got a start by separating the system prompt and the user prompt. Is there another step further we could go that would treat the "unsafe" data differently than the safe data, in a very similar way that we do with SQL queries?

If this isn't an active area of research, I'd bet there's a lot of money to be made waiting to see who gets into it first and starts making successful demos…

This reads like an infomercial... like someone was paid to write it.

If we're going to grant that a machine can have infinite ram, surely we can grant that a spec of C can have infinite stack allocated arrays... :)

Can I have one please?

I think your criticism is well founded.

That said, the author did put work into something, document it, and share it freely with the community - and for that reason, I'd rather not see someone hurt their feelings calling their work an 'abomination.'

I just think with a slightly different tone you could have made the same point in a way that would show the OP some brutal honesty but without discouraging them.

Please, be constructive instead of just dismissive. A nasty tone doesn't help anyone nor does it encourage trying new things.

I bet the variation in the size of people's fingers is greater than the difference in the size of touch targets between original and mini iPads... :)

Deprecated, but not removed yet. It "may be removed in a relatively distant future"

That's true, and I think assigning malice to this situation is unfounded. But maybe this is a problem worth solving?

Come on, don't reject it so flippantly. Sure, for the case of a woman getting married and taking a new last name, I doubt it's that big of a deal - you can change your name for future commits, but your old name will exist for historical commits. However, there _are_ cases where you might not want your former name around (transgender, or even something like privacy / witness protection). Right now, these folk are being sort of excluded (however inadvertently) and it's worth discussing ways to fix that.

That stood out to me too... is there an inside joke? What's the white have to do with it?

I know you'd miss the syncing, but why not let me just cc [email protected] instead of having to gain access to my google account? It would be usable to people on exchange as well that way...

Yea that pesky tablet revolution wasn't a thing either, and neither was the GUI actually. I'm still using lynx.

(No offense to lynx users.)

Snark aside, it seems like high dpi will inevitably become standard in the next few years.

I didn't know it had been confirmed yet, can you add a link?

That doesn't mean that people don't care. I certainly wouldn't able to remember the birthday of all my closest friends, but I do care and am glad to be reminded. On the other hand, there are of course a ton of bday wishes from people who probably don't really care, but is that really a bad thing?