staunton

5 ms·

1.
by staunton 1y ago
People have been going nuts all throughout recorded history, that's really nothing new. The only scary thing is that they have ever more power to change the world and influence others without being forced to grapple with that responsib
2.
by staunton 1y ago
> you would think they would put your name in the system prompt They probably do, along with "pretend to not have any personal information about the user".
3.
by staunton 1y ago
Don't forget the "nature partner journals", which people also call (and cite as) "Nature XY"...
4.
by staunton 1y ago
> labor is what they most have to offer, so that is what they will trade What will they eat? Whose land might they be allowed to grow their food on?
5.
by staunton 1y ago
Also a lot of errors would be called "typos", not errors. Such as some edge cases missing in the theorem statement which technically makes the theorem false. As long as there's a similar theorem in the same spirit that can be
6.
by staunton 1y ago
Have you ever read the (full) text of any bill that has been passed during the last couple of decades? How about reading all of them? So are you proposing people vote on them without reading them? Or that we write very short bills aimed at
7.
by staunton 1y ago
What does it mean "to make filling out your taxes"??
8.
by staunton 1y ago
> should also be caught in code reviews Assuming they even have code reviews - in your experience, in a situation where the person writing the code didn't check if it already exists, the reviewer will check that and then tell them t
9.
by staunton 1y ago
> models deployed in critical applications such as finance, law, and healthcare. We went really quickly from "obviously noone will ever use these models for important things" to "we will at the first opportunity, so please
10.
by staunton 1y ago
Most musicians (i.e. non-famous ones) get most of their income from teaching students. I don't think such a model makes sense for developers (though who knows, maybe at some time in the future there will be significant numbers of peopl
11.
by staunton 1y ago
Addiction is a matter of degree. There's a bunch of polls where a large majority of people strongly agree that "they spend too much time on social media". Are they addicts? Are they "coosing to use it"? Are they say
12.
by staunton 1y ago
> We've been through this before in the 70s-90s. What do you mean?
13.
by staunton 1y ago
> if people thought social media was bad, they wouldn’t use it. Do you think Heroin is good?
14.
by staunton 1y ago
Would you say Log4j is an internal or an external dependency?
15.
by staunton 1y ago
> interfaces You mean obsolete and subtly wrong comments buried among preprocessor directives in some far-upstream header files?
16.
by staunton 1y ago
I'd guess that's mostly because C has been around long enough to be chosen for it and people who are into that kind of thing don't urgently need more than one contest, regardless of language...
17.
by staunton 1y ago
Agreed. Does any of that contradict something I wrote?
18.
by staunton 1y ago
Objections: - it's possible to transfer knowledge, as demonstrated by the fact that human civilization exists. It's not always easy, doesn't always succeed, and reading is a part of that, but it's possible and happening.
19.
by staunton 1y ago
Well, if you're using GRPC anyway , and have the protocol buffers already ... It's hard to resist the temptation after a few hours of installing broken package versions. Incidentally, I suspect this is the spiritual origin of mi
20.
by staunton 1y ago
Set them up, each in its own container, and use GRPC to communicate between them... (Not even kidding, I've seen people do this)
21.
by staunton 1y ago
This sounds very educated but I don't really see what it has to do with the comment you're responding to (or with AI).
22.
by staunton 1y ago
I think it's important not to be too sure abot what of the future one is "seeing". It's easy to be confidently wrong and one may find countless examples and quotes where people made this mistake. Even if you don't t
23.
by staunton 1y ago
An LLM can even ignore lack of precision and just guess what you wanted, usually correctly, unless what you want is very unusual.
24.
by staunton 1y ago
What standard would that be?
25.
by staunton 1y ago
> you can always detect the AI by text and timing patterns I see no reason why an AI couldn't be trained on human data to fake all of that. If noone has bothered so far, that's because pretty much all commercial applications of
26.
by staunton 1y ago
For this to be plausible, you have to explain why the people controlling the AI would share their wealth. They would either do it voluntarily (and be outcompeted by those who don't?) or be coerced (by who? Someone who doesn't have
27.
by staunton 1y ago
That's just like asking to replace priests by AI...
28.
by staunton 1y ago
However, it's a lot more socially acceptable to say "I missed the email" than "I sat there for 30min while you were talking but didn't actually listen"...
29.
by staunton 1y ago
An interesting aspect of this: Where I work, an email subject line including "important" or "urgent" is a 99% indication for junk...
30.
by staunton 1y ago
Leaving a region means they would give up market share, investments and a lot of staff. This is huge. It's not just about quarterly profit, even if it might sometimes feel like it is. Presumably, before they take such a drastic measure

More ›