The article isn't about automatic proofs, but it'd be interesting to see a LLM that can write formal proofs in Coq/Lean/whatever and call external computer algebra systems like SageMath or Mathematica.
Audalin
A DIY Mjolnir! But I'd still go with batteries for portability.
Cool! Will try once the F-Droid version updates.
Looks like there're going to be some actual uses for alchemy aside from pies, honey potions and healing fruits. (Other shiny things are nice, I tried most of them at some point, but they usually don't provide enough advantages to take up an inventory slot just for this unique crafted item. Perhaps they're more useful if you turn some challenges on.)
I had two runs with that mechanic, first huntress, then mage. At first I didn't know how the now challenging ascension would play out, but I was pretty confident - after all, can there be any threat for one who's reached the amulet? I turned out to be so wrong then. I made it through in both cases, but I was close to death so many times. A wand of blast wave let me kill enemies instantly in some cases in the first run; I finished with a drained +6 staff of blast wave in the second one (also with other guaranteed environmental damage wands).
An extremely busy month, so I'm just reading some Apollinaire.
LaTeX is great (that is, nothing can really compete with it), but its design is an abomination at this point. There're just too many different subsystems working in different paradigms with different conventions simultaneously, e.g.:
- plain TeX with boxes, kerning, counters, lengths and so on;
- LaTeX;
- expl3;
- TikZ/PGF is its own world with its own scary places. At the same time it relies too much on TeX/LaTeX to be fast or reliable;
- Lua if you use LuaLaTeX.
Good luck debugging if something doesn't work immediately - each of the subsystems doesn't know how the other subsystems work, most have their own independent error handling and reporting routines, so ~95% of errors don't actually tell what's wrong, and only ~50% somewhat help figuring out where there's something that's wrong. You have reasonable suspicions in most cases, but not always (especially if you're in the process of writing a complex macro and you know your initial try is likely to be faulty - you just don't know where and why exactly, even with plenty of experience). There're no easy ways to debug macro expansion step-by-step.
Here're some fun exercises if you don't realize how terrible the situation is:
- Make a macro generating a table - a true tabular - with m rows and n columns (both macro parameters). In each cell, print the sum of the row number and the column number. (easy)
- Set up the compilation process so that all compile files (
.aux
,.log
&c.) end up in thebuild/
directory and not wherever your.tex
files are. Now draw some nice TikZ pictures. Now set up TikZ externalization. (medium) - The same as the first exercise, but now apply word wrapping behaviour to table columns: if there's no place for another column, the table should end, and a new table should continue from the next line (the numbering of columns must continue). Avoid columns getting past the line boundaries, even slightly. But you can't break groups of 3 columns. Also draw horizontal red lines connecting the centres of horizontally adjacent cells within those groups of 3. (hard)
I had to solve these - or very similar in nature - problems while working on various things.
I've been trying it on dozens of wins across different versions. Nothing so far. The chasm acts as a solid wall.
I also planted rotberries there a few times.
Now bring it back! It isn't nearly as hard as it once was. I remember virtually invincible 1st level rats. I had to spend everything consumable on my way back in that run - fire potions turned out very useful as HP numbers weren't scaled then. Now you just hit the enemies and they just die, very simple.
You should enjoy Italo Calvino's Mr Palomar then.
Being an honest observer of the world outside, Mr Palomar seeks patterns in everything and is always torn between constantly emerging contradicting perspectives. He strives to attain transcendent understanding and fails; he also strives to connect with mundane everyday things and fails. He can't even decide which of the two is a nobler goal, and that indecision is also something he analyses.
It's a short beautiful book, give it a try.
An average day in Tol-in-Gaurhoth:
One giveaway is teeth: surely a tooth shouldn't grow from the upper jaw into the lower.
Yeah, I win 1/3 of my runs consistently, but without challenges. I'm still figuring out which challenge would be the easiest to try. Probably the boss-boosting one.
What do you tend to craft with challenges enabled?