Hacker News: fractorial

New comment by fractorial in "Gemini 4 Argon"

fractorial — Thu, 01 Oct 2026 00:40:30 +0000

How is Anthropic different? You still cannot use a subscription in a custom harness in the same way Codex allows you to.

And, no, wrapping claude -p or any other “allowed” use wherein you don’t control the agent loop isn’t the same.

New comment by fractorial in "I am often wrong"

fractorial — Mon, 21 Sep 2026 17:40:39 +0000

I am in the process of open sourcing it! Unfortunately, my employer requires a (hopefully) cursory legal review.

New comment by fractorial in "I am often wrong"

fractorial — Sun, 20 Sep 2026 20:44:46 +0000

I’ve rolled my own harness in Go. I have a rule to not let the production LoC exceed 50k lines. It is _very_ nice for my use cases. DeepSeek Flash V4.1 often performs at the level of GPT 5.6 Sol for non-orchestration tasks (programming and maths).

I add features and tweak it all the time: You just need to get comfortable with spending 3 hours in a chat session every 3 weeks to think hard about how to simplify whatever slop you didn’t simplify the last time you did this :)

New comment by fractorial in "Show HN: Art – draw one stroke, let symmetry complete it"

fractorial — Thu, 10 Sep 2026 13:17:45 +0000

As a programmer, maths researcher, and painter, there’s certainly something satisfying (to me) when an artistically symmetrical piece observes no mathematical symmetry: The very asymmetry in the language used to describe it is, in and of itself, beautiful.

New comment by fractorial in "Mathematicians want proof OpenAI didn't use their work"

fractorial — Thu, 10 Sep 2026 11:42:08 +0000

Only under gross negligence would it be unprovable: Did you use a model whose training set included user data? Did the transcripts of any of the agents include a tool call whose result including user data?

New comment by fractorial in "Claude, change the “Add to Cart” button to blue"

fractorial — Wed, 09 Sep 2026 13:30:05 +0000

I’m impressed you had the patience to even make it that far!

New comment by fractorial in "Claude, change the "Add to Cart" button to blue"

fractorial — Wed, 09 Sep 2026 13:09:52 +0000

Brilliant. Precisely the reason I stopped using Anthropic's products.

New comment by fractorial in "Can AI design circuit boards yet?"

fractorial — Sat, 05 Sep 2026 00:56:21 +0000

Sir, I think you are lost.

New comment by fractorial in "Google Antigravity TOS: 3rd party usage can get Google account suspended"

fractorial — Thu, 03 Sep 2026 12:35:39 +0000

> Because Codex without codex is fantastic.

Seriously: I roll my own harness over OpenRouter and GPT 5.6, and it is genuinely better than CC+Fable for my maths and programming use cases.

The one thing Fable 5.1 still tops out on is the ability to take a simple "Let's engage in an interactive complexity-removing session" and materialize abstractions and semantics quickly.

New comment by fractorial in "OpenAI restores 5-hour Codex and Work limits for ChatGPT Plus users"

fractorial — Sat, 29 Aug 2026 03:12:52 +0000

`pathological` is being used here in the security context: For example, _a parser contains a vulnerability when pathological inputs cause a program to crash._

New comment by fractorial in "Agent Lightning v1.0"

fractorial — Tue, 25 Aug 2026 01:53:24 +0000

How does a _multi-trillion dollar company_ think this is good presentation?

New comment by fractorial in "Fences, Not Sandboxes"

fractorial — Tue, 25 Aug 2026 00:11:08 +0000

This is ludicrous; also, why publicly admit that blatant violation of ToS?

New comment by fractorial in "Anthropic appears to be A/B testing reduced effort levels in Claude Code"

fractorial — Sat, 22 Aug 2026 18:15:08 +0000

Roll your own harness or use an open source harness with a Codex subscription.

I maintain a Claude subscription for Fable but seldom use it.

New comment by fractorial in "Stop Making TUIs"

fractorial — Sat, 22 Aug 2026 13:01:55 +0000

Damn, what an epic library name.

New comment by fractorial in "Palomar: A registry of Lean verified mathematics"

fractorial — Wed, 19 Aug 2026 13:02:29 +0000

Palomar requires the Comparator/formalization.yaml challenge stack. It also has an LLM-based vacuity gate backed by Tao.

Theoremdb is just a slop farm.

New comment by fractorial in "MathCode, Mathematical Coding Agent"

fractorial — Sun, 16 Aug 2026 21:58:40 +0000

To be clear, I am deep into auto-research, but hooking up slop to slop is just unlikely to produce anything valuable.

Value is in how maths is communicated: The process, frustrations, triumphs, etc.

We have to able to take generated formalizations from “it compiles” to “it is correct” before crystallizing them.

New comment by fractorial in "Show HN: ThoughtDAG – An editable context graph for LLM conversations"

fractorial — Sat, 15 Aug 2026 14:26:54 +0000

Irrespective of opinions about a project or its authors, disclosing a potential security issue in a public forum is markedly not responsible disclosure.

New comment by fractorial in "AI Model Atlas – visualizing populations of ML models as interconnected 3D graph"

fractorial — Fri, 14 Aug 2026 15:38:43 +0000

This is, plainly stated, awesome.

New comment by fractorial in "The Conceptual Reasoning Index"

fractorial — Thu, 13 Aug 2026 14:43:29 +0000

I would love for someone to give me a coherent argument as to how this isn’t tone-deaf, vacuous garbage.

New comment by fractorial in "Go is an ideal language for AI-assisted software engineering"

fractorial — Wed, 12 Aug 2026 02:24:05 +0000

I can’t say much for a Chief Evangelist, but the Product Manager certainly has quite the C.V.