Zetice

93 ms·

1111.
by Zetice 4y ago
Right, and as I’ve been saying, I can throw a stapler at your head, so what?
1112.
by Zetice 4y ago
The point is LLMs are not effective at “hacking” in any “obtaining unauthorized access to computer systems” sense. They can regurgitate information about “hacking”, same as a library, but pointing an LLM at a server will achieve worse resul
1113.
by Zetice 4y ago
That’s a very “is” argument, but I’m saying we “ought” not worry such as I see in this submission’s comments. It’s self defining; whatever people are saying here, I’m saying those comments are overblown. What “it” is I leave up to whoever
1114.
by Zetice 4y ago
On one hand, what kind of monster writes an API that kills people??? On the other hand, we all know it’d be GraphQL…
1115.
by Zetice 4y ago
Innovation doesn’t follow any path; discovery is messy. No matter how much we advance towards smaller chips, we are never going to get to 0nm, for example. There are limits, but even if there weren’t, we’re no closer to AGI today then we we
1116.
by Zetice 4y ago
That’s an easy prediction to make; at worst you’re cautious. And it’s a powerful tool. Even staplers have rules around their use: no stapling people, no hitting people with a stapler, don’t use a staple to pick a lock, etc. But nobody blame
1117.
by Zetice 4y ago
These language models don't know how to hack stuff. They know that certain characters and words strung together can satisfy their training when someone asks them to pretend to hack something. That's wildly different, and a lot
1118.
by Zetice 4y ago
If I set my stapler on my mouse such that it clicks a big ol "Hack stuff" button, my stapler could, too, carry out cyber attacks. This is a very pointless line of thinking.
1119.
by Zetice 4y ago
This is exactly what I'm talking about. None of what you wrote here is anchored in reality. At all. Not even a little. It's pants-on-head silly to think "ChatGPT 2.0" is anything other than, at best, a magpie. If you
1120.
by Zetice 4y ago
No, what? GPT is, very roughly, a set of training data plus a way of associating that data together to answer prompts. It's not "relationships, people, and processes", it's not "our collective intent"; what t
1121.
by Zetice 4y ago
1 is false; it is both accurate and predictive to describe what these systems do as mimicking/regurgitating human output. That's exactly what they're doing. 2 is irrelevant; you can doomsay and speculate all day, but if it&
1122.
by Zetice 4y ago
You can shove staplers into wall sockets too (if you're determined enough), but the consequences are on you, not the stapler. It's just not meaningfully different from our current reality, and is therefore not any scarier.
1123.
by Zetice 4y ago
Can we please stop with this "not aligned with human interests" stuff? It's a computer that's mimicking what it's read. That's it. That's like saying a stapler "isn't aligned with human inte
1124.
by Zetice 4y ago
The market is too cruel to be left to take care of America's children.
1125.
by Zetice 4y ago
The entire article is about how it's not.
1126.
by Zetice 4y ago
Seeing a lot of hate for Spotify's implementation of podcasts; does anyone here regularly use the app for Podcast listening? I do, and it's fine . It doesn't blow me away, and I expect they're trying to, but it's
1127.
by Zetice 4y ago
Ack, YouTube Music did this thing where it was playing random YouTube videos from people who uploaded the song, rather than an "official" release, and I fled wildly. I want nothing to do with YouTube when I load in YouTube Musi
1128.
by Zetice 4y ago
So in other words, no it’s ultimately not important if the idea is good, which Bard seems to be. It seems more indicative of leadership quality that Google had this in their back pocket, ready to go shortly after someone else did the market
1129.
by Zetice 4y ago
We don't know it was an identical response firstly, and secondly if you ask it who the president was in 1845, it will give the "identical" response to that question every time as well, so IMO it is indeed "quite likely
1130.
by Zetice 4y ago
I don't believe that means much; getting the same response twice just means the AI drew the same conclusions from a similar prompt.
1131.
by Zetice 4y ago
Will anyone remember how Bard was released in 5 years, if it delivers on its promises? I doubt it, so the hand wringing over a stumble this early seems overblown.
1132.
by Zetice 4y ago
Couldn’t this potentially be just another completion, where the AI is “playing along” and generating responses as if the user had broken through, even though the user had not? How would we even know?
1133.
by Zetice 4y ago
Those don't really come across as deal breakers though, just preferences. It may not be to your liking but isn't it a bit dramatic to say "worst-in-class" without including the "for me" caveat?
1134.
by Zetice 4y ago
Maybe just schedule the not-intellectually-heavy stuff earlier in the day, if we can't screw up the parents' schedules with a later start time. For example, start school with a study hall (that you can miss if you're so able