Hacker News: dhorthy

New comment by dhorthy in "SlopCodeBench 10/2026 – Opus 5.5, Astra wins overall, xhigh considered harmful"

dhorthy — Wed, 21 Oct 2026 03:53:28 +0000

or someone could clone the SCB repo and make a fork that proves it one way or another

New comment by dhorthy in "SlopCodeBench 10/2026 – Opus 5.5, Astra wins overall, xhigh considered harmful"

dhorthy — Tue, 20 Oct 2026 23:42:20 +0000

yep agree

New comment by dhorthy in "SlopCodeBench 10/2026 – Opus 5.5, Astra wins overall, xhigh considered harmful"

dhorthy — Tue, 20 Oct 2026 23:41:59 +0000

they yearn to create content, generate more tokens, etc etc - agree

New comment by dhorthy in "SlopCodeBench 10/2026 – Opus 5.5, Astra wins overall, xhigh considered harmful"

dhorthy — Tue, 20 Oct 2026 23:41:01 +0000

Yeah and I started with subsets although I did not find them representative, eg I picked 6 challenges for a subset and fable scored 33%, while on the full challenge it landed closer to 14-15% range

The part II post linked in my top comment has a list of the challenges I used for that.

To run them, just point your model at the paper, it can find the repo, etc and with a key or sub, get it running

SlopCodeBench 10/2026 – Opus 5.5, Astra wins overall, xhigh considered harmful

dhorthy — Thu, 15 Oct 2026 19:09:34 +0000

Quick update here since a lot of folks asked about what slopcodebench is / where it fits in

Slop code bench is unique among benchmarks in that it makes a model work against a progressively disclosed list of features (checkpoints) that it doesn’t know it advance, inheriting the same codebase through the lifecycle of a single multi-checkpoint challenge

I like this because it much more mirrors how real software dev happens, rather than most benchmarks with one off tasks where slop code doesn’t hinder future progress

The original post[1] and the SOL/fable follow up[2] went into a bit more detail on the benchmark and its challenges

1. https://news.ycombinator.com/item?id=49076391

2. https://github.com/humanlayer/advanced-context-engineering-f...

Comments URL: https://news.ycombinator.com/item?id=50108717

Points: 216

# Comments: 88

New comment by dhorthy in "Etoomanythings? Run Fewer Agents"

dhorthy — Thu, 01 Oct 2026 16:55:27 +0000

great post - the same thing that has always been true of software engineering is still true:

Do fewer things in parallel, do them very well, optimize for getting incremental value out to users, FOCUS

It's been 30 years since the XP and Lean folks said "limit work in progress (50+ if you count the eli goldratt thing) - its still true

The Jagged Frontier of Jev 1.13

dhorthy — Tue, 22 Sep 2026 23:04:07 +0000

Article URL: https://docs.typesafe.ai/model-jaggedness/jev-1.13

Comments URL: https://news.ycombinator.com/item?id=49809509

Points: 3

# Comments: 0

NoCode (2016)

dhorthy — Mon, 31 Aug 2026 22:50:55 +0000

Article URL: https://github.com/kelseyhightower/nocode

Comments URL: https://news.ycombinator.com/item?id=49515808

Points: 4

# Comments: 0

Archil Persistent Sandboxes

dhorthy — Fri, 28 Aug 2026 17:57:15 +0000

Article URL: https://archil.com/post/introducing-persistent-sandboxes

Comments URL: https://news.ycombinator.com/item?id=49482185

Points: 2

# Comments: 0

MicroGPT-C in pure C hits 10M TPS on Apple M5

dhorthy — Tue, 18 Aug 2026 15:46:46 +0000

Article URL: https://github.com/vixhal-baraiya/microgpt-c

Comments URL: https://news.ycombinator.com/item?id=49347477

Points: 141

# Comments: 52

New comment by dhorthy in "AI;DR (AI; Didn't Read)"

dhorthy — Tue, 18 Aug 2026 01:28:07 +0000

> AI-written emails where the commenter would rather see the prompt

this a thousand times this

LLMs are information transformers. if you're trying to take some rough idea and blow it out into something that "feels" substantive, you're just combining your high-signal idea / prompt with a bunch of meaningless bloat and noise from the model weights.

New comment by dhorthy in "Show HN: /show-me: agent skill for compact visual representations"

dhorthy — Thu, 13 Aug 2026 13:19:53 +0000

they seem to be able to do the call-stack diff stuff pretty well still

Show HN: /show-me: agent skill for compact visual representations

dhorthy — Wed, 12 Aug 2026 15:58:22 +0000

Was so sick of reading walls of codex/claude prose, in markdown plans and just in the chat, that I started playing with ideas that force coding agents to display information differently.

The human visual cortex is an amazing thing, and getting coding agents to let me use it has been pretty nice so far.

Been iterating on this a lot internally for the last few months, polishing and mostly removing stuff.

Comments URL: https://news.ycombinator.com/item?id=49274489

Points: 12

# Comments: 3

The Programming Ape (Code Hale, 2012) [video]

dhorthy — Wed, 12 Aug 2026 00:37:44 +0000

Article URL: https://www.youtube.com/watch?v=e_6gkfTomUQ

Comments URL: https://news.ycombinator.com/item?id=49266411

Points: 11

# Comments: 0

Benchmarking Fable, Sol, and Kimi K3 on SlopCodeBench

dhorthy — Tue, 04 Aug 2026 15:08:01 +0000

Article URL: https://github.com/humanlayer/advanced-context-engineering-for-coding-agents/blob/main/benchmarking-sol-fable-kimi-on-slop-code-bench.md

Comments URL: https://news.ycombinator.com/item?id=49170037

Points: 8

# Comments: 1

New comment by dhorthy in "Benchmarking Opus 5 on SlopCodeBench"

dhorthy — Tue, 28 Jul 2026 14:55:05 +0000

so wait is the finding that most of those skills reduce pass rates against SCB? wild

New comment by dhorthy in "Benchmarking Opus 5 on SlopCodeBench"

dhorthy — Tue, 28 Jul 2026 13:31:09 +0000

why would that worry you?

New comment by dhorthy in "Benchmarking Opus 5 on SlopCodeBench"

dhorthy — Tue, 28 Jul 2026 13:30:55 +0000

agree, i think the implication is that low quality code is harder to change in the future

New comment by dhorthy in "Benchmarking Opus 5 on SlopCodeBench"

dhorthy — Tue, 28 Jul 2026 04:52:34 +0000

somebody get this man a curl-pipe-bash stat

New comment by dhorthy in "Benchmarking Opus 5 on SlopCodeBench"

dhorthy — Tue, 28 Jul 2026 04:19:32 +0000

oh i really like the idea of flipping around the order of checkpoints and comparing results. Could be an interesting way to increase/decrease difficulty even

i will look into how easy it would be to zip up some subset of the results without leaking anything...probably doable