Top all-time · #1309
Entropic Thoughts
Personal blog · entropicthoughts.com · on HN as kqrmathprogramming
Entropic Thoughts is a personal blog at entropicthoughts.com about math and programming. Since 2024, 56 different Hacker News readers have linked it 307 times, most often “Lines of Code Are Useful”. It is still writing: the last post was on 2026-10-28, and you can follow it by feed.
Most-linked posts?
- Lines of Code Are Useful5 people · 5 mentions · first linked 2026-03-12
- Updated LLM Benchmark (Gemini 3 Flash)4 people · 7 mentions · first linked 2025-12-20
- You need a kitchen slide rule4 people · 6 mentions · first linked 2025-08-26
- The Misunderstood Kelly Criterion4 people · 6 mentions · first linked 2024-10-18
- FizzBuzz Through Monoids3 people · 5 mentions · first linked 2025-05-23
- Deploying a Single-Binary Haskell Web App3 people · 4 mentions · first linked 2024-10-11
- Snakes and Ladders3 people · 4 mentions · first linked 2026-04-18
- Root cause analysis? You're doing it wrong3 people · 4 mentions · first linked 2024-10-09
- Statistical Process Control: Practitioner's Guide3 people · 4 mentions · first linked 2025-01-29
- Haskell: A Great Procedural Language3 people · 4 mentions · first linked 2024-11-27
What Hacker News says?
This is one of my favourite images from a long-defunct proto-meme blog: https://entropicthoughts.com/image/doesntworkbutfast.jpg
Comment linking /image/doesntworkbutfast.jpg
on “You did this with an AI and you do not understand what you're doing here”Submitted Transparent leadership beats servant leadership as a story · 514 points
HN discussionYes, 2.5 Flash is extremely cost efficient in my favourite private benchmark: playing text adventures[1]. I'm looking forward to testing 3.0 Flash later today. [1]: https://entropicthoughts.com/haiku-4-5-playing-text-adventur...
Comment linking /haiku-4-5-playing-text-adventures
on “Gemini 3 Flash: Frontier intelligence built for speed”Submitted Haskell: A Great Procedural Language as a story · 280 points
HN discussionFor anyone curious: an average of 1.2 instances per month the first year and a half. During the worst season, it can be twice in a month, sure. https://entropicthoughts.com/how-often-does-a-child-get-sick (Child spent 16 % of the year sick. Across two parents, that is nearly a month of absence in a year.)
Comment linking /how-often-does-a-child-get-sick
on “If you're remote, ramble”Submitted Babel is why I keep blogging with Emacs as a story · 272 points
HN discussionI have made a similar tool but for the command line[1] with similar but slightly more ambitious motivation[2]. I really like that more people are thinking in these terms. Reasoning about sources of variation is a capability not all people are trained in or develop, but it is increasingly important.[3] [1]: https://git.sr.ht/~kqr/precel [2]: https://entropicthoughts.com/precel-like-excel-for-uncertain... [3]: https:/…
Comment linking /statistical-literacy
on “Show HN: Unsure Calculator – back-of-a-napkin probabilistic calculator”Submitted Practices of Reliable Software Design as a story · 231 points
HN discussion> Nah, the plural of anecdote is confirmation bias. Why do people keep trotting out this "plural of anecdote is data" line, is it a quote from some high-profile idiot? It is standard Bayesian reasoning[1]. But it requires independence between observations, which many people forget! [1]: https://entropicthoughts.com/bayes-rule-odds-form
Comment linking /bayes-rule-odds-form
on “My 16-month theanine self-experiment”Submitted Probability-generating functions as a story · 207 points
HN discussionI want to write more about this, but it has been a really difficult subject to structure. I gave up halfway through this article, for example, and never published it – I didn't even get around to editing it, so it's mostly bad stream of consciousness stuff: https://entropicthoughts.com/root-cause-analysis-youre-doing... I intend to come back to it some day, but I do not think that day is today.
Comment linking /root-cause-analysis-youre-doing-it-wrong
on “Ryanair flight landed at Manchester airport with six minutes of fuel left”Submitted NixOS is making me less afraid as a story · 202 points
HN discussionThis is surprisingly competent. A couple of months ago I evaluated some leading models on a bunch of text adventures[1]. Typical regression coefficients would be +0.02 for top level models like Sonnet and Gemini 2.5 Pro, but notably also Gemini 2.5 Flash. (The baseline is GPT 5 Chat, i.e. the one where OpenAI routes to a thinking model only when they determine it's needed.) When I include an attempt from Haiku 4.5 i…
Comment linking /evaluating-llms-playing-text-adventures
on “Claude Haiku 4.5”Submitted What Killed Perl? as a story · 199 points
HN discussionI have noticed that when I wrote blog posts they tend to fall in one of the two categories. Sometimes I'm trying to share an insight, in which case I make sure not to bury the lede[1]. Sometimes it's the journey to the insight that matters more than the insight itself[2], in which case the narrative take precedence, even if it buries the lede. In some cases it is possible to combine both, by using the storytelling f…
Comment linking /deploying-single-binary-haskell-web-app
on “How to write blog posts that developers read”Submitted Root cause analysis? You're doing it wrong as a story · 137 points
HN discussionOh yeah. I realised this the day I discovered my fancy digital SLR was a thermometer: https://entropicthoughts.com/does-my-dslr-have-dead-pixels
Comment linking /does-my-dslr-have-dead-pixels
on “Phased Array Microphone (2023)”Submitted Parser Combinators Beat Regexes as a story · 122 points
HN discussionI found a what I think is a flaw in the methodology and started to write a comment, but it turned into its own article. If you've ever wondered what the entropy of the Mario colour sequence is, let's find out! https://entropicthoughts.com/the-most-mario-colour-revisited
Comment linking /the-most-mario-colour-revisited
on “The Most Mario Colors”Submitted Evaluating LLMs playing text adventures as a story · 118 points
HN discussionLog-odds difference from contingency table: ln (4×11) - ln (1×8) = ln (44/8) = ln 5.5 ~= 2.3 × lg 5.5 ~= 2.3 × 0.75 = 1.75. Standard error of log-odds difference: √(1/1 + 1/11 + 1/4 + 1/8) ~= √(1 + 0.09 + 0.25 + 0.13) = √(1.47) = 1.47^0.5 = 10^(0.5 × lg 1.47) ~= 10^(0.5 × 0.16) = 10^0.08 ~= 1.2. The effect is 1.75/1.2 ~= 1.5 standard errors from zero, thus not significant by common thresholds. You would find Fisher'…
Comment linking /learning-some-logarithms
on “Jim Lovell, Apollo 13 commander, has died”Submitted The Greatness of Text Adventures as a story · 117 points
HN discussionI can strongly recommend memorising some logarithms for use in mental maths. It's given me powers I did not expect to have! Here's what I wrote about it when I started: https://entropicthoughts.com/learning-some-logarithms
Comment linking /learning-some-logarithms
on “The Lost Art of Logarithms”Submitted Advent of Code on the Z-Machine as a story · 108 points
HN discussionI feel like I have a much better grasp of those numbers since I started using logarithms for mental maths: https://entropicthoughts.com/learning-some-logarithms
Comment linking /learning-some-logarithms
on “The Lost Art of Logarithms”Submitted A curmudgeon tries a language server as a story · 106 points
HN discussionI noticed this when I tried to get LLMs to play text adventures. Early on in my experiments, I wanted to give them hints when they got stuck at a puzzle. I did this by stopping the loop, injecting thoughts into the LLMs own persistent scratchpad (as if it had thought of that itself), and then starting the loop up again.[1] That way, the LLM would read what it had intended to remember from the previous turn including…
Comment linking /getting-an-llm-to-play-text-adventures
on “Three sites made 215,128 “best software” pages for AI. Perplexity cites them”Submitted You Want Technology with Warts as a story · 100 points
HN discussionInsurance can actually be a win-win if you keep your money sitting anywhere else other than a literal piggy bank. Basically, it takes just as much time for your investments to go from 10,000 USD to 50,000 USD as it does for 50,000 USD to go to 250,000 USD. So a setback of 5,000 USD has a disproportionate impact on your future financials the less money you start off with. In other words, if your total wealth is low e…
Comment linking /when-is-insurance-worth-it
on “Home Loss File System”Submitted The Wind, a Pole, and the Dragon as a story · 93 points
HN discussion
