Top all-time · #3002
Can.ac
Personal blog · blog.can.acsecuritysystemsai
Can.ac is a personal blog at blog.can.ac about security, systems and ai. Since 2018, 24 different Hacker News readers have linked it 26 times, most often “Speculating the entire x86-64 instruction set in seconds with one weird trick”. It is still writing: the last post was on 2026-08-03, and you can follow it by feed.
Too few sampled comments (13) to say how often HN recommends it.?
Why HN links it
HN readers link Can Bölük's posts on AI coding-agent tooling, especially the hash-anchored edit approach for LLM agents. Other linked posts cover x86 instruction counting and PatchGuard.
Summarised by Claude from a sample of 13 HN comments that link here. Each comment is marked recommending, critical or neutral (citing without judging); how this works.
Reverse-engineering and whatnot
Most-linked posts?
- /2026/02/12/the-harness-problem/8 people · 8 mentions · first linked 2026-02-18
- Speculating the entire x86-64 instruction set in seconds with one weird trick8 people · 8 mentions · first linked 2021-03-23
- Snapcompact: SoTA Compaction – Instant, Local, Free. Pick 35 people · 6 mentions · first linked 2026-06-11
What Hacker News says?
On x86. The linked article loaded very, very slow. Original (?) bloppost: https://blog.can.ac/2021/03/22/speculating-x86-64-isa-with-o...
Comment linking /2021/03/22/speculating-x86-64-isa-with-one-weird-trick/
on “Clever Hack Finds Mystery CPU Instructions”Submitted Speculating the entire x86-64 instruction set in seconds with one weird trick as a story · 208 points
HN discussionAlso see https://blog.can.ac/2021/03/22/speculating-x86-64-isa-with-o...
Comment linking /2021/03/22/speculating-x86-64-isa-with-one-weird-trick/
on “How many x86 instructions are there? (2016)”Submitted Speculating the entire x86-64 instruction set in seconds (2021) as a story · 104 points
HN discussionAlso vision can be used for "compaction" https://blog.can.ac/2026/06/10/snapcompact/
Comment linking /2026/06/10/snapcompact/
on “DeepSeek Introduces Vision”Submitted Improving 15 LLMs at Coding in One Afternoon. Only the Harness Changed as a story · 10 points
HN discussionFor the hash-anchored edits, sharing here Can Bölük's original post about the idea https://blog.can.ac/2026/02/12/the-harness-problem/
Comment linking /2026/02/12/the-harness-problem/
on “Show HN: OSS Agent I built topped the TerminalBench on Gemini-3-flash-preview”Submitted Speculating the Entire x86-64 Instruction Set in Seconds as a story · 7 points
HN discussionThere will always be a gap, but what's interesting is that because new models are constantly coming out, we as an industry never spend any time extracting the maximal value out of an existing model. What if there are techniques, and harness workflows that could be optimized for a singular model end to end? How far can that push the state of the art. An example is https://blog.can.ac/2026/02/12/the-harness-problem/ f…
Comment linking /2026/02/12/the-harness-problem/
on “DeepSeek 4 Flash local inference engine for Metal”Submitted Speculating the Entire x86-64 Instruction Set in Seconds as a story · 5 points
HN discussion> Fun observation - almost every coding harness (claude code, cursor, codex) uses a find/replace tool as the primary way of interacting with code. This requires the agent to fully type out the code it's trying to edit, including several lines of context around the edit. This is really inefficient, token wise! Why does it work this way? Because the LLMs are really bad at counting lines, or using other ways of describ…
Comment linking /2026/02/12/the-harness-problem/
on “I don't think AGI is imminent”Submitted Speculating the Entire x86-64 Instruction Set with This One Weird Trick as a story · 4 points
HN discussionFor the edit tool, you should consider implementing a hash-based approach where each line of code is hashed and referenced by it when doing replacements. You can read up on the approach here: https://blog.can.ac/2026/02/12/the-harness-problem/ I didn't do much benchmarking, but anecdotally, I found it to be making less edit errors. YMMV
Comment linking /2026/02/12/the-harness-problem/
on “Ask HN: Has anyone replaced Claude/GPT with a local model for daily coding?”Submitted Snapcompact: SoTA Compaction – Instant, Local, Free. Pick 3 as a story · 2 points
HN discussionPrior art: https://blog.can.ac/2026/02/12/the-harness-problem/
Comment linking /2026/02/12/the-harness-problem/
on “A new EDIT tool for LLM agents”Submitted Snapcompact, SoTA Compaction: Instant, Local, Free. Pick 3 as a story · 2 points
HN discussionNote there is a fork oh-my-pi: https://github.com/can1357/oh-my-pi of https://blog.can.ac/2026/02/12/the-harness-problem/ fame. I use it as a daily driver but I also love pi.
Comment linking /2026/02/12/the-harness-problem/
on “Pi – a minimal terminal coding harness”Submitted Reverse Engineering Hyperliquid as a story · 2 points
HN discussionhttps://blog.can.ac/2026/02/12/the-harness-problem/ It's worse than I expected.
Comment linking /2026/02/12/the-harness-problem/
on “Benchmarking Coding Agents on Databricks' Multi-Million Line Codebase”Submitted Speculating the Entire x86-64 Instruction Set in Seconds as a story · 2 points
HN discussionRelated: https://blog.can.ac/2026/06/10/snapcompact/
Comment linking /2026/06/10/snapcompact/
on “60% Fable cost cut by converting code to images and having the model OCR it”Submitted Snapcompact: Rendering context as pixel-font bitmaps cuts input cost 3x as a story · 1 points
HN discussion> it just couldnt write the changes to file 9/10 times when using it with opencode You might want to give this a try, it dramatically improves Edit tool accuracy without changing the model: https://blog.can.ac/2026/02/12/the-harness-problem/
Comment linking /2026/02/12/the-harness-problem/
on “Apple's accidental moat: How the "AI Loser" may end up winning”Submitted Snapcompact: SoTA Compaction – Instant, Local, Free as a story · 1 points
HN discussion2024 and PatchGuard is still a meme: https://blog.can.ac/2024/06/28/pgc-garbage-collecting-patchg...
Comment linking /2024/06/28/pgc-garbage-collecting-patchguard/
on “Bypassing PatchGuard on Windows x64 (2005)”Submitted Arbitrary Code Execution at Ring 0 Using CVE-2018-8897 as a story · 1 points
HN discussion
