Hacker News

How do you trace your own mental loops?

I keep catching myself in recursive thought patterns—debugging production issues at 2am while my brain debugs hypothetical conversations from 2019.

I've started treating these like performance bugs: tracing the call stack, finding the root trigger, watching the loop unwind.

Before I build anything: How do you "debug" your own cognition? Do you journal, use frameworks, white-knuckle it, or just accept the overhead?

Genuine curiosity. If there's a pattern in the responses, I'll share what I've been experimenting with.

Comments URL: https://news.ycombinator.com/item?id=46972710

Points: 1

# Comments: 0

Hi HN,

We are a team of independent researchers from Germany working on ARC AGI 2 since last summer. The general opinion on open-weight models is that they are too weak for this fairly difficult benchmark and score at near noise levels. We found that GPT OSS 120B is actually much more capable than previously thought, once the interleaved thinking regime is stabilized. We basically let the model use a stateful IPython based REPL via function calling and patched vLLM so that the model can reliably do interleaved thinking. The score jumped more than 400%.

Technical write-up: https://pivotools.github.io/posts/agentic_coding_arc_agi/ Code: https://github.com/gutfeeling/arc-agi-2-submission Data: https://huggingface.co/datasets/arcagi2/arcagi2-agentic-codi...

For safety, we support sandboxed execution using IPyBox (local Docker) and Daytona (cloud), so others can reproduce this without running untrusted code locally.

It gets more interesting: the effect seems to be general and translates seamlessly to other models without even changing prompts. We are not sure why agentic coding is so powerful in ARC AGI 2, which isn't traditionally thought of as an agentic benchmark. Perhaps code execution provides a stronger form of verification than COT, or perhaps it encourages a qualitatively different form of thinking.

We will be around for a while and would be happy to hear ideas / feedback and discuss infra issues / interleaved thinking / GPT OSS / ARC AGI 2.

Comments URL: https://news.ycombinator.com/item?id=46972709

Points: 1

# Comments: 0

Article URL: https://intentcode.dev

Comments URL: https://news.ycombinator.com/item?id=46972702

Points: 1

# Comments: 1

Article URL: https://www.sandordargo.com/blog/2026/02/11/deferred-map-initialization

Comments URL: https://news.ycombinator.com/item?id=46972688

Points: 1

# Comments: 0

Article URL: https://matharena.ai/?view=problem&comp=aime--aime_2026

Comments URL: https://news.ycombinator.com/item?id=46972626

Points: 2

# Comments: 1

Article URL: https://fortune.com/2026/02/10/government-borrowing-cbo-report-deicit-lending-interest/

Comments URL: https://news.ycombinator.com/item?id=46972617

Points: 1

# Comments: 0

Article URL: https://www.nytimes.com/2026/02/11/us/faa-el-paso-flight-restrictions.html

Comments URL: https://news.ycombinator.com/item?id=46972610

Points: 7

# Comments: 0

Article URL: https://blog.dougbelshaw.com/polycrisis-ebook/

Comments URL: https://news.ycombinator.com/item?id=46972604

Points: 1

# Comments: 1

Article URL: https://webtil.es/

Comments URL: https://news.ycombinator.com/item?id=46972570

Points: 1

# Comments: 0

Article URL: https://fosdem.org/2026/schedule/event/HVJRNV-how_to_level_up_the_fediverse/

Comments URL: https://news.ycombinator.com/item?id=46972569

Points: 1

# Comments: 0

Article URL: https://www.youtube.com/watch?v=l03Os5uwWmk

Comments URL: https://news.ycombinator.com/item?id=46972551

Points: 1

# Comments: 0

Article URL: https://www.thehumanlineproject.org/

Comments URL: https://news.ycombinator.com/item?id=46972549

Points: 1

# Comments: 0

Article URL: https://github.com/sara8086/PostgresBooks

Comments URL: https://news.ycombinator.com/item?id=46972547

Points: 1

# Comments: 0

Article URL: https://blogs.ubc.ca/networked/2011/02/02/google-docs-more-than-just-documents/

Comments URL: https://news.ycombinator.com/item?id=46972539

Points: 2

# Comments: 0

Article URL: https://vasilis.nl/nerd/2026/rip-marijn/

Comments URL: https://news.ycombinator.com/item?id=46972508

Points: 1

# Comments: 0

Article URL: https://www.forbes.com/sites/conormurray/2026/02/09/anthropic-ai-safety-researcher-warns-of-world-in-peril-in-resignation/

Comments URL: https://news.ycombinator.com/item?id=46972496

Points: 2

# Comments: 1

As you know, we can install the desired skill using the following command:

npx skills add https://github.com/anthropics/skills --skill frontend-design

Is there a way to achieve the same setup in an on-premise environment without exposing the repository publicly?

Comments URL: https://news.ycombinator.com/item?id=46972491

Points: 1

# Comments: 0

Article URL: https://github.com/onurkanbakirci/skills-gateway

Comments URL: https://news.ycombinator.com/item?id=46972473

Points: 1

# Comments: 0

Article URL: https://astro.theoj.org/article/156033-a-cosmic-miracle-a-remarkably-luminous-galaxy-at-_z_-sub-spec-sub-14-44-confirmed-with-jwst

Comments URL: https://news.ycombinator.com/item?id=46972471

Points: 2

# Comments: 0

Article URL: https://blog.fluxer.app/how-i-built-fluxer-a-discord-like-chat-app/

Comments URL: https://news.ycombinator.com/item?id=46972226

Points: 1

# Comments: 0