NeurIPS isn’t selecting ideas, it’s administering obedience tests

NeurIPS isn’t selecting ideas - it’s administering obedience tests.

All four of my papers were desk-rejected, not for scientific errors, not for lack of innovation, but because each one lacked checklists.

One page. One form. One ritual. That’s all it took to erase an entire body of work.

They call that fairness: everyone must obey the same template, or be excluded equally. But fairness based on forced conformity is a farce. It doesn’t reward insight - it enforces silence. A system that treats formatting as virtue and imagination as disobedience does not deserve to call itself scientific. And what was rejected wasn’t theory - it’s already being realized.

In January, I designed Wormhole Memory, a constant-time context retriever for long-term dependencies. Just months later, OpenAI’s GPT update implemented the same principle.

In February, I proposed Dynamic Chain-of-Thought, a reasoning compression mechanism to reduce token waste. Qwen 3 deployed it.

In April, I unveiled Defierithos, a resonance-based, post-Transformer architecture for scalable language cognition.

Three innovations. One independent researcher. Four papers.

And all were desk-rejected - without peer review - only because they lacked checklists. It’s filtration by obedience!

Genius doesn’t conform! Genius questions rules before it follows them! Genius often refuses to be standardized.

I design from conviction, not concession. I question defaults, I rewrite systems, and when told to obey - I say, “I won’t.” I build forward, not inward. I build because 我命由我不由天.

So I’ve made it all public, not for permission, it’s an act of defiance. NeurIPS didn’t give me a platform - so I built my own.

To those hiring thinkers, not followers - I’m here. I’m seeking roles in AI strategy, cognitive architecture, and problem-solving systems design.

I work from first principles, not inherited patterns.
If you’re building what doesn’t yet exist, I’m ready to join you. My email is free.equality.anyone@gmail.com
I don’t want to be normalized. I want to be the useful AI Strategist / Cognitive Architect / Solution Strategist.

:link: All four rejected works, along with full code, benchmarks, and architecture diagrams, are now released on OpenReview Achieve:

My LinkedIn link is: https://www.linkedin.com/in/libo-wang-1612801a2