munchler

56 ms·

751.
by munchler 3y ago
Existential risk, I think. https://www.lesswrong.com/tag/existential-risk
752.
by munchler 3y ago
They're not bad. They're just a pain to type out in detail every time. Here's an example signature from an F# library I like called FParsec: pipe2: Parser<'a,'u> -> Parser<'b,'u> ->
753.
by munchler 3y ago
As you get more comfortable with functional programming, the type signatures tend to get more complex for sophisticated code. Those "nasty generics" become reliable friends. For example, since functions are first-class objects, th
754.
by munchler 3y ago
Me too, although it's actually "Clédial". I was certain no one would market a drug with an accented character in the US.
755.
by munchler 3y ago
I think you're right, and the article says "in some rare cases the method may produce incorrect result due to approximation errors".
756.
by munchler 3y ago
The dog’s owner paid for surgery, but wouldn’t pay for an x-ray?
757.
by munchler 3y ago
Don't forget Lost In Space and Battlestar Galactica. All essentially the same premise.
758.
by munchler 3y ago
Careful. If you get significantly below 50%, then you could be predicted by the negation of this algorithm. To truly beat it, I think you want to be right around 50%.
759.
by munchler 3y ago
Training a model with this data might be legal, but distributing the data without a license probably isn't.
760.
by munchler 3y ago
I think GP's point is that you can't measure changes in the length of an arm if the ruler you're using is changing in the same way. I think this misrepresents the situation, but I can't say how, so I can't dismiss i
761.
by munchler 3y ago
Hah, I wasn’t aware of that. Does it say how the base model computes its confidence? Is it requested in the prompt?
762.
by munchler 3y ago
The irony of this comment is rich. There’s no such “confidence value metric” known, no matter how “certain” you are. This is why techniques like chain-of-thought or self-correction are used instead. LLM’s predict the next token in a sequenc
763.
by munchler 3y ago
> Remove tuples - use records instead. Seriously? Mainstream languages are busy adding tuples because they make code so much cleaner when used appropriately. Leaving them out of Simple Haskell™ would be a big mistake, IMHO.
764.
by munchler 3y ago
This is what F# uses (as well as |>, >>, and <<), and I like it a lot. It's consistent and intuitive pattern. Every time I read Haskell, it takes extra effort to read . and $ instead.
765.
by munchler 3y ago
It's an unintentional pun on the word "bar". You can be disbarred, which means expelled from the (legal) bar. Or you can be barred (prevented) from practicing law. But "disbarred from practicing law" is either redun
766.
by munchler 3y ago
Another unspoken red flag is that this app is having so many bugs in production that it requires constant attention from its developers. How are all these bugs getting past QA? Is there a legit development process in place? Who is signing o
767.
by munchler 3y ago
Side note, but sentences like these slay me: "[It] qualified as an unqualified success." "[He] was disbarred from practicing law." Whoever wrote this article was an unintentional genius.
768.
by munchler 3y ago
> No human wrote it, no humans will read it. The is probably going to be true for other types of documents as well, such as business emails. See this article: https://marketoonist.com/2023/03/ai-written-ai-read.
769.
by munchler 3y ago
Maybe "unlimited" would be better?
770.
by munchler 3y ago
A future where we have artificial gravity but still don't know when to say "lie down" instead of "lay down". Very realistic!
771.
by munchler 3y ago
Yes, but now the process can be easily automated to solve any CAPTCHA.
772.
by munchler 3y ago
This is an interesting argument, but not convincing. You are essentially claiming that the information content of a classical recording is low, and could be replicated with fewer, simpler instruments. However, I suspect that many of the int
773.
by munchler 3y ago
A typical classical symphony requires 50-80 instruments, several dozen of which might play at the same time, while a typical punk song has maybe four.
774.
by munchler 3y ago
Interviewee: Jimena Canales | Historian of science and author of The Physicist and the Philosopher: Einstein, Bergson and the Debate That Changed Our Understanding of Time (2015) as well as Bedeviled: A Shadow History of Demons in Science (
775.
by munchler 3y ago
In what way is git evidence of a lack of creatives? I don’t see the connection.
776.
by munchler 3y ago
> This case is an example of the new risks organizations face when starting to leverage the power of AI more broadly, as more of their engineers now work with massive amounts of training data. It seems like a stretch to associate this ri
777.
by munchler 3y ago
(2011)
778.
by munchler 3y ago
(2012)
779.
by munchler 3y ago
The idea is that sufficiently powerful compression is indistinguishable from artificial intelligence, which can be used to make even more money.
780.
by munchler 3y ago
I really like this approach, but it contains some confusing mistakes. For example, unless I'm very much mistaken, the illustration of the initial object is backwards: https://abuseofnotation.github.io/category-theory-il

More ›