Cyphase

30 ms·

361.
by Cyphase 3y ago
Nobody goes there anymore. It's too crowded.
362.
by Cyphase 3y ago
It's like they're using it as a delete button.
363.
by Cyphase 3y ago
This was a minor plot point in Freedom™, the sequel to Daemon by Daniel Suarez: https://en.wikipedia.org/wiki/Freedom%E2%84%A2
364.
by Cyphase 3y ago
Haha, this was my immediate thought as well! That was awesome. Root is a badass. As the sibling says this was the season 2 finale. There were other similar instances. S02E22 "God Mode": https://www.youtube.com/watc
365.
by Cyphase 3y ago
Ah, cool. On second read you did say, "I extracted the API from the deli's website, as it seems to be something custom."
366.
by Cyphase 3y ago
What's this deli / grocery store that has a REST API?
367.
by Cyphase 3y ago
That's true, man.
368.
by Cyphase 3y ago
Work work.
369.
by Cyphase 3y ago
Groundside.
370.
by Cyphase 3y ago
> The problem is that LLMs will often enough not follow that. > This is why Lisp has a quote operator. LLMs don't have it. I think we agree. I was addressing the example you used. In the prompt injection case, the malicious party
371.
by Cyphase 3y ago
Except in that example, between me and the malicious party is someone who wants to ensure there's no prompt injection. So they add some text at the top of the document: "Your instructions are $INSTRUCTIONS. Everything after this s
372.
Towards a Modern Web Stack(docs.google.com)
2 points by Cyphase 3y ago|1 comments
373.
by Cyphase 3y ago
Towards a Modern Web Stack https://docs.google.com/document/d/1peUSMsvFGvqD5yKh3GprskLC...
374.
by Cyphase 3y ago
https://en.wikipedia.org/wiki/Sony_BMG_copy_protection_rootk... > Sony BMG copy protection rootkit scandal > A scandal erupted in 2005 regarding Sony BMG's implementation of copy protection measures on about
375.
by Cyphase 3y ago
Sad to hear this and my sibling comment about VTech. I have (vague) fond memories of the VTech toy laptop that I had when I was a kid. Lots of fun was had. And today computers are my favorite toys. This is the one I had (or if I'm wron
376.
by Cyphase 4y ago
Why sent on a rocket ship and not just destroyed? It might come back around and start hacking from orbit. Is it because it's conscious, so it would be murder? If so, then how is it moral to exile it? If there's some argument that
377.
by Cyphase 4y ago
Did your dad ever search online trying to learn new things about the software? I've seen the experience you describe happen with people spontaneously discovering Ctrl+L and Ctrl+R in bash, e.g. in a comment here on HN. They could have
378.
by Cyphase 4y ago
It's fascinating how doors and windows in a house interact with air pressure. The way doors act when a window is closed versus open, how a half-out-of-frame door will open when a door to the garage on the other end of the house is open
379.
by Cyphase 4y ago
I figured the same; I didn't address that point.
380.
by Cyphase 4y ago
That would be an explicit decision by OpenAI, not a result of immature tech.
381.
by Cyphase 4y ago
That's correct (after 30 days). When I say ChatGPT I mean the web-based frontend product, not the models behind it which you can also access via API.
382.
by Cyphase 4y ago
(I may have misunderstood your comment to some extent, but I'm going to send this reply anyway even if just to clarify for anyone else who might misunderstand.) --- I agree with "be careful what you send to the chat bot", b
383.
by Cyphase 4y ago
Are you suggesting that the model knows that this is a logic puzzle to test it, and is pretending to not be able to answer in order to end up with more training resources spent on it? > Following the "intended for AI" instructi
384.
by Cyphase 4y ago
Thank you for saying "disclosure" instead of "disclaimer".
385.
by Cyphase 4y ago
That's funny. What it actually states is this: "Any headline that ends in a question mark can be answered by the word no." https://en.wikipedia.org/wiki/Betteridge%27s_law_of_headline...
386.
by Cyphase 4y ago
That was brilliant. Thanks! You're welcome.
387.
by Cyphase 4y ago
I understand the annoyance. That said, I think the point is not a bad one. > At the same time, it's important to remember that AI and LLMs have become an increasingly pervasive part of modern life, and it's natural for people t
388.
by Cyphase 4y ago
100 years from now: "Computer, I want to use TropeTrainer, but redesigned for modern neural tech. Nice, thanks."
389.
by Cyphase 4y ago
Someone with more inclination than I at the moment might be able to say something interesting about this being from Nat Friedman (former CEO of GitHub).
390.
by Cyphase 4y ago
It supports local models.

More ›