another reason is that they don't offload work to a calculator. They generate a program *using inference* to compute it. So where "12 * 15" is going to be about 2 tokens of inference to hallucinate an answer directly, offloading it to a deterministic solver requires generating that whole program
October 2, 2026 at 4:02 PM
another reason is that they don't offload work to a calculator. They generate a program *using inference* to compute it. So where "12 * 15" is going to be about 2 tokens of inference to hallucinate an answer directly, offloading it to a deterministic solver requires generating that whole program
(She's not talking about GPS as in global positioning systems here - she's talking about G.P.S. as in the highly influential 1957 "Global Problem Solver" computer program).
en.wikipedia.org/wiki/General...
en.wikipedia.org/wiki/General...
General Problem Solver - Wikipedia
en.wikipedia.org
September 3, 2025 at 5:01 PM
(She's not talking about GPS as in global positioning systems here - she's talking about G.P.S. as in the highly influential 1957 "Global Problem Solver" computer program).
en.wikipedia.org/wiki/General...
en.wikipedia.org/wiki/General...
Craziest thing about lean.py? You can roll your own (almost) feature-parity successor to Dafny.
Python programs. Lean obligations. Grind as a solver.
Proof complete. Program verified. Python code. Real.
Python programs. Lean obligations. Grind as a solver.
Proof complete. Program verified. Python code. Real.
May 20, 2026 at 5:52 AM
Craziest thing about lean.py? You can roll your own (almost) feature-parity successor to Dafny.
Python programs. Lean obligations. Grind as a solver.
Proof complete. Program verified. Python code. Real.
Python programs. Lean obligations. Grind as a solver.
Proof complete. Program verified. Python code. Real.
Come help us design, build and buy the trains and buses that move millions! We’re hiring for the special projects team in the rolling stock program.
We’re looking for a fast and strategic thinker, problem solver and consensus builder who’ll assist w/‘hot potatoes’ — careers.mta.org/jobs/1827678...
We’re looking for a fast and strategic thinker, problem solver and consensus builder who’ll assist w/‘hot potatoes’ — careers.mta.org/jobs/1827678...
September 24, 2026 at 9:42 PM
Come help us design, build and buy the trains and buses that move millions! We’re hiring for the special projects team in the rolling stock program.
We’re looking for a fast and strategic thinker, problem solver and consensus builder who’ll assist w/‘hot potatoes’ — careers.mta.org/jobs/1827678...
We’re looking for a fast and strategic thinker, problem solver and consensus builder who’ll assist w/‘hot potatoes’ — careers.mta.org/jobs/1827678...
Here's a solver for outcomes of 3d6 + 1d4
anydice.com/program/a5ec
(I would write this as a dynamic solver.)
"16 or better" results in
8.22 + 6.02 + 3.94 + 2.31 + 1.16 + 0.46 + 0.12
= 22.23%
~ 2/9
Taking rounding into account, it is plausible that the answer is precisely 2/9.
2/
anydice.com/program/a5ec
(I would write this as a dynamic solver.)
"16 or better" results in
8.22 + 6.02 + 3.94 + 2.31 + 1.16 + 0.46 + 0.12
= 22.23%
~ 2/9
Taking rounding into account, it is plausible that the answer is precisely 2/9.
2/
AnyDice
AnyDice is an advanced dice probability calculator, available online. It is created with roleplaying games in mind.
anydice.com
November 23, 2024 at 1:23 AM
Here's a solver for outcomes of 3d6 + 1d4
anydice.com/program/a5ec
(I would write this as a dynamic solver.)
"16 or better" results in
8.22 + 6.02 + 3.94 + 2.31 + 1.16 + 0.46 + 0.12
= 22.23%
~ 2/9
Taking rounding into account, it is plausible that the answer is precisely 2/9.
2/
anydice.com/program/a5ec
(I would write this as a dynamic solver.)
"16 or better" results in
8.22 + 6.02 + 3.94 + 2.31 + 1.16 + 0.46 + 0.12
= 22.23%
~ 2/9
Taking rounding into account, it is plausible that the answer is precisely 2/9.
2/
The program for DigiPro 2026 has been posted! There's a lot of great talks this year; I'm really looking forward to Alexey Stomakhin's talk on WetaFX's new fire solver/system. If you're familiar with Alexey's work, you already know it's going to be mind-blowing.
digiproconf.org/program/
digiproconf.org/program/
June 17, 2026 at 5:34 AM
The program for DigiPro 2026 has been posted! There's a lot of great talks this year; I'm really looking forward to Alexey Stomakhin's talk on WetaFX's new fire solver/system. If you're familiar with Alexey's work, you already know it's going to be mind-blowing.
digiproconf.org/program/
digiproconf.org/program/
The absolute solver can be a pretty unpredictable and corrupting program so best be careful using it or else~
Art by @lunabara.bsky.social
#roboErdno #Erdno #murderdronesOC #macro
Art by @lunabara.bsky.social
#roboErdno #Erdno #murderdronesOC #macro
January 14, 2025 at 3:12 PM
The absolute solver can be a pretty unpredictable and corrupting program so best be careful using it or else~
Art by @lunabara.bsky.social
#roboErdno #Erdno #murderdronesOC #macro
Art by @lunabara.bsky.social
#roboErdno #Erdno #murderdronesOC #macro
sat solver based obfuscation where it mathematically constructs a program that is equivalent to yours
June 27, 2025 at 1:42 AM
sat solver based obfuscation where it mathematically constructs a program that is equivalent to yours
Amy Paris was hired to help modernize the nation's donor organ network, which ensures critically-ill people get the organs they need.
She was also let go.
"We had alignment from Democrats and Republicans on the Hill, we had funding, and they were hiring more of us.”
She was also let go.
"We had alignment from Democrats and Republicans on the Hill, we had funding, and they were hiring more of us.”
February 22, 2025 at 2:09 PM
Amy Paris was hired to help modernize the nation's donor organ network, which ensures critically-ill people get the organs they need.
She was also let go.
"We had alignment from Democrats and Republicans on the Hill, we had funding, and they were hiring more of us.”
She was also let go.
"We had alignment from Democrats and Republicans on the Hill, we had funding, and they were hiring more of us.”
Portland Problem Solver meetings give Portlanders, particularly business owners and neighborhood association leaders, a biweekly chance to talk to city staff who are directed to work fast on their behalf. The program was launched by former Mayor Wheeler in 2022 to address “livability issues.”
August 13, 2025 at 4:16 PM
Portland Problem Solver meetings give Portlanders, particularly business owners and neighborhood association leaders, a biweekly chance to talk to city staff who are directed to work fast on their behalf. The program was launched by former Mayor Wheeler in 2022 to address “livability issues.”
Just for fun, I created a Mixed Integer Linear Program for compiling crosswords www.solvermax.com/blog/crosswo...
Solver Max - Crossword MILP - Model 1
We use a mixed integer linear program to compile crossword grids
www.solvermax.com
December 22, 2024 at 3:07 AM
Just for fun, I created a Mixed Integer Linear Program for compiling crosswords www.solvermax.com/blog/crosswo...
Some Portlanders have been meeting frequently with city staff for more than three years through the Problem Solver program, meant to quickly address city livability issues. But the program has not been widely publicized, and basic information about the meetings is hard to come by.
Portland's Most Effective Channel For Addressing Community Concerns is a Back Door
Portland’s new form of government, which went into effect this year, was designed to make it more accessible for all residents to participate in city policy discussions. Still, most Portlanders strugg...
www.portlandmercury.com
August 13, 2025 at 4:53 PM
Some Portlanders have been meeting frequently with city staff for more than three years through the Problem Solver program, meant to quickly address city livability issues. But the program has not been widely publicized, and basic information about the meetings is hard to come by.
AVBD turned out to be much harder to program compared to XPBD. I suspect there still are lots of bugs in my solver, but the biggest issue right now seems to be stable contact generation/maintenance... #gamedev #indiegamedev #gamephysics
February 25, 2026 at 2:26 PM
AVBD turned out to be much harder to program compared to XPBD. I suspect there still are lots of bugs in my solver, but the biggest issue right now seems to be stable contact generation/maintenance... #gamedev #indiegamedev #gamephysics
The Absolute Solver Creator, a man who knows his program. #murderdrones #murderdronesoc #absolutesolver #ocart #digitalart
March 19, 2026 at 8:10 PM
The Absolute Solver Creator, a man who knows his program. #murderdrones #murderdronesoc #absolutesolver #ocart #digitalart
Spent the last couple of days porting my program verification class from Dafny to Lean via Loom/Velvet, and it just works!
Whenever the SMT solver can’t fully prove a program correct, Lean’s aesop and grind take care of the remaining goals.
Whenever the SMT solver can’t fully prove a program correct, Lean’s aesop and grind take care of the remaining goals.
October 30, 2025 at 11:59 AM
Spent the last couple of days porting my program verification class from Dafny to Lean via Loom/Velvet, and it just works!
Whenever the SMT solver can’t fully prove a program correct, Lean’s aesop and grind take care of the remaining goals.
Whenever the SMT solver can’t fully prove a program correct, Lean’s aesop and grind take care of the remaining goals.
Anybody know any good jokes? Waiting for something..
February 1, 2026 at 7:46 PM
Anybody know any good jokes? Waiting for something..
I wrote my first real Julia program (as in, a program that actually does something useful) a few weeks ago. It was fun! I needed a constraint solver interface and JuMP was very nice for that. Two lines of code (generated by LLM) I don't understand, the rest was so clean.
github.com/shriram/lab-...
github.com/shriram/lab-...
November 25, 2024 at 10:50 PM
I wrote my first real Julia program (as in, a program that actually does something useful) a few weeks ago. It was fun! I needed a constraint solver interface and JuMP was very nice for that. Two lines of code (generated by LLM) I don't understand, the rest was so clean.
github.com/shriram/lab-...
github.com/shriram/lab-...
Sudoku solver program using exact cover?
December 5, 2024 at 7:06 PM
Sudoku solver program using exact cover?
Some people like the meetings and have submitted glowing testimonials about how effective they’ve been in getting their concerns addressed. Again, the city staff is expected to work very quickly with Problem Solver attendees—on a scale most residents can’t relate to.
August 13, 2025 at 4:29 PM
Some people like the meetings and have submitted glowing testimonials about how effective they’ve been in getting their concerns addressed. Again, the city staff is expected to work very quickly with Problem Solver attendees—on a scale most residents can’t relate to.
oh wow! what does the solver program do?
February 22, 2026 at 5:54 PM
oh wow! what does the solver program do?
possibly *the* foundational problem in program analysis is complexity cliffs: usability is not linear. once your program surpasses a threshold of complexity, everything goes to hell.
- type errors become useless
- program slicers miss dependencies
- SMT solver heuristics give way to O(2^n)
- etc
- type errors become useless
- program slicers miss dependencies
- SMT solver heuristics give way to O(2^n)
- etc
July 21, 2025 at 9:17 PM
possibly *the* foundational problem in program analysis is complexity cliffs: usability is not linear. once your program surpasses a threshold of complexity, everything goes to hell.
- type errors become useless
- program slicers miss dependencies
- SMT solver heuristics give way to O(2^n)
- etc
- type errors become useless
- program slicers miss dependencies
- SMT solver heuristics give way to O(2^n)
- etc
some more interesting stuff:
"It seems that we need an "automatic programmer" as a component of an extensible, efficient problem solver. Many people working on automatic program synthesis expect that a system can be built which will, given a description of the problem to be ...
"It seems that we need an "automatic programmer" as a component of an extensible, efficient problem solver. Many people working on automatic program synthesis expect that a system can be built which will, given a description of the problem to be ...
September 24, 2026 at 6:03 PM
some more interesting stuff:
"It seems that we need an "automatic programmer" as a component of an extensible, efficient problem solver. Many people working on automatic program synthesis expect that a system can be built which will, given a description of the problem to be ...
"It seems that we need an "automatic programmer" as a component of an extensible, efficient problem solver. Many people working on automatic program synthesis expect that a system can be built which will, given a description of the problem to be ...
i guess it's kinda funny that i'm like so i made a wordle solver back when it was new and instead of intuitively picking the correct data structure for the search problem i just took the most naive approach and knocked out entries in a list and that's proof i just don't know how to program at all
December 1, 2025 at 8:14 AM
i guess it's kinda funny that i'm like so i made a wordle solver back when it was new and instead of intuitively picking the correct data structure for the search problem i just took the most naive approach and knocked out entries in a list and that's proof i just don't know how to program at all
I am just reading up on the “Problem Solver Network” and fuck fucking ted wheeler right in the fucking ear, of course his priority was a backdoor anti-homeless program to connect biz owners with city officials weekly. This fucking city sometimes, i swear.
August 13, 2025 at 6:30 PM
I am just reading up on the “Problem Solver Network” and fuck fucking ted wheeler right in the fucking ear, of course his priority was a backdoor anti-homeless program to connect biz owners with city officials weekly. This fucking city sometimes, i swear.
🦊finished up ashe's C version of our nonogram solver (or as much as I feel like finishing up lmao) seems like it IS faster than the C++ version, but only when you don't have -O1 on. C seems really cool and bits of C in an overall C++ program seems neat. A large scale C app though, I dont know lol
January 10, 2026 at 10:56 PM
🦊finished up ashe's C version of our nonogram solver (or as much as I feel like finishing up lmao) seems like it IS faster than the C++ version, but only when you don't have -O1 on. C seems really cool and bits of C in an overall C++ program seems neat. A large scale C app though, I dont know lol