peterhull90
peterhull90 on Hacker News
Can anyone explain (or link to an article which explains) why telling Claude things like not to cheat or that there is a holdout set makes the output better? People say that an LLM is just a statistical next token predictor. I've read the article by Stephen Wolfram [0] on this, which makes sense to me, but then I can't see how 'more complex' behavior like cheating or not cheating can arise.
[0] : https://writings.stephenwolfram.com/2023/02/what-is-chatgpt-...
In the HN comment that the article discusses [0] is the conclusion that commenter a1369209993 is correct (there are as many between 0 & 1 as 1 & +INF) and llm_trw is not correct? I got a bit confused.
Also, the article links to a blog by Daniel Lemire [1] in which he says (with regard to producing an unbiased random float) "picking an integer in [0,2^32) at random and dividing it by 2^32, was equivalent to picking a number at random in [0,1)" is incorrect and there is a ratio of up to 257:1 in the distribution so obtained. Not wanting to disagree with Daniel Lemire but I can't see why, and a quick experiment in Python didn't give this ratio.
[0]: https://news.ycombinator.com/item?id=41112688
[1]: https://lemire.me/blog/2017/02/28/how-many-floating-point-nu...
You can put an extra 'level' in, to make both options return the same 0,1...9
def loop():
for number in range(10):
def outer(n):
def inner():
return n
return inner
yield outer(number)
Is there a neater way?
Currently reading "Byzantium: The Surprising Life of a Medieval Empire" by Judith Herrin[0]
It's very interesting, a sort of missing piece of history from my point of view. At school those centuries was just skipped as 'the dark ages.'
[0]: https://en.wikipedia.org/wiki/Judith_Herrin
It may be different on newer MacOS but on mine you will need ping6
> ping6 ff02::1%en1
for example.
I wasn't able to get it to work on Windows (I can get the command to work in WSL but it seems to sit alone in an internal network). Any advice welcome.
When I started my first job in 1995, the old hands were still using DEC ALL-IN-1 [0] from terminals on their desks but it was in the process of being phased out (for Microsoft Mail or something on Windows for Workgroups), so I never got to use it myself.
[0]: https://en.wikipedia.org/wiki/ALL-IN-1
In my youth I had a ZX spectrum which had BASIC with line numbers and no renumber command. Sometimes when adding code I'd simply run out of line numbers so had to GO TO an unused block of line numbers, put the new code there and GO TO just after the original code. I've never quite recovered from that.
It's a consequence of being block-based as mentioned elsewhere, but interesting to note that cat'ing together bzip2 files gives a valid bzip2 file. That's the basis of pbzip2 [0] - it breaks the input file into chunks of 900K by default, compresses each chunk and then concatenates the compressed chunks. The individual chunks can be compressed in parallel if hardware allows.
[0]: https://man.freebsd.org/cgi/man.cgi?query=pbzip2&apropos=0&s...
Also Brian Anderson (brson) was/is a significant rust contributor
And could one swap between the two backends with the same VM image (.vbox +.vdi) to see which one gave the better performance?
Might depend on what kind of book it is - do you have a lot of images, tables, cross-references, ... or is it mostly plain text?
Can anyone just clarify for me what these AI / Machine learning chips are? As far as I can tell they are general purpose microprocessors with some added instructions which accelerate operations that are commonly used (matrix multiplication possibly?) but there's a lot I don't quite understand because some of the info is marketing and some of it it heavy technical stuff and my knowledge falls in between!
Bravo! Was there anyone used the 'Briefcase' back in the day? I'd forgotten about that.
In fact I got so in the zone that I pressed ALT-F4 to close Win98 notepad and shut down my whole web browser instead!
> Popovertarget passes every choice made by the player, for example: x1-o2-x3 means X chooses the first position, O chooses the second and X chooses the third, and so on.
I think this is going to lead to a lot of redundant states, isn't it, because there are multiple paths to get to a particular board layout?
"Physics of Semiconductor Devices" was a very useful and clearly-written book I read in the 90s - I must still have my copy somewhere. I didn't know the author had made such groundbreaking discoveries.
I've been using X to view graphical applications running on the Pi from my desktop - will this still be possible?
So if your revenue (not profit) exceeds $5000, you need to pay $5600 for Delphi. Surely that can't be right?
[edit] I can only get the prices in GBP but for 'professional' it's £1,140 (Perpetual License) + £399 (Year one Maintenance) so not quite as bad as I thought but still a sizeable portion for year 1.
Does anyone know why they picked frequencies quite close together for the in/out tones (2525Hz/2475Hz) when the pass band for telephony was much wider?
In the mid-90s I did some playing round with something called MINDY from Carnegie Mellon. It was (IIRC) a Dylan interpreter (the acronym was 'MINDY Is Not Dylan Yet') and was very interesting both to experiment with Dylan and also to see how MINDY itself was implemented.
As I remember, a cool thing about the Dylan concept was that it had a sliding scale of dynamic-ness - in other words you could start off being quite relaxed about types but, where needed, you could narrow down the types, seal generic methods etc. and the compiler would then be able to optimise better, for example by avoiding dynamic dispatch if it could prove it knew the specific method at compile time.