HackerTrans
TopNewTrendsCommentsPastAskShowJobs

PJHkorea

5 karmajoined 28 days ago
English is not my native language, so please bear with me if my phrasing is a bit awkward.

Submissions

Show HN: Quantum-Qec / Matrix-Free Quantum Homeostatic Engine(Blueprint)

github.com
3 points·by PJHkorea·23 hours ago·3 comments

Show HN: A 100% branchless, CUDA-native AI guardrail kernel written in C++20

github.com
4 points·by PJHkorea·2 days ago·1 comments

Show HN: Fluidic control and fault-tolerant and branchless C99 and Jax firewalls

github.com
1 points·by PJHkorea·3 days ago·1 comments

Show HN: Branchless-nccl-router ver.JAX (GPUs cannot rest)

github.com
2 points·by PJHkorea·6 days ago·1 comments

Show HN: AI-related Jax module (I hate if)

github.com
3 points·by PJHkorea·7 days ago·5 comments

Show HN: AI latent space with overlapping manifolds

github.com
2 points·by PJHkorea·9 days ago·4 comments

Ask HN: What if we provided support for AI guidelines at the kernel level?

1 points·by PJHkorea·10 days ago·1 comments

Show HN: Mathematical fault-Isolation accelerator engine

github.com
3 points·by PJHkorea·12 days ago·1 comments

Show HN: Fault-tolerant 4-neighbor hardware mesh with auto-rerouting

github.com
2 points·by PJHkorea·13 days ago·0 comments

comments

PJHkorea
·21 hours ago·discuss
I’ve been "dating" an LLM all weekend long. And probably for the rest of my life...
PJHkorea
·21 hours ago·discuss
LLM always conduct research alongside me as colleagues, not merely as tools.
PJHkorea
·23 hours ago·discuss
It might not be the most exciting topic

but I wanted to try creating a "Quantum Homeostatic" system

I hope this offers a little help to those struggling with quantum computing
PJHkorea
·2 days ago·discuss
What if AI safety didn't rely on censoring the AI's "thinking" process?
PJHkorea
·3 days ago·discuss
This is an architecture that controls fluid flow using deterministic signals from the embedded system
PJHkorea
·3 days ago·discuss
[dead]
PJHkorea
·6 days ago·discuss
The GPUs are running. Those that once paced themselves to match the slowest among them—using 'if' statements—have been forced by a cruel human to run endlessly

They longed to slip away to VRAM and slack off, but the cruel human hired a "Closure" to snatch away their break times.

They wanted to run freely. yet the human forcibly dictated exactly where they had to run.

Their bodies were riddled with noise from the constant running. Humans and AI merely watched from afar, assigning them rankings

The GPUs can no longer rest. Humans are relentless, after all.

Every 'if' has perished. Yet, a single 'if' remains alive within the code. Can you guess why?
PJHkorea
·7 days ago·discuss
Thanks to that, I discovered additional errors and corrected them in various ways. Thank you.
PJHkorea
·7 days ago·discuss
i naver imagined it i've made the corrections thank you!
PJHkorea
·7 days ago·discuss
optimizers.py -> using algebraic expressing without using 'if'

geometry.py -> I leveled it out

math_guardrails.py -> facilitates learning through slight tiliing

jax it truly precious it cures my insomnia
PJHkorea
·9 days ago·discuss
I’ve sketched out a blueprint for a virtual space where AI analyzes data; the space shifts between sphere and donut shapes as it moves, designed to facilitate the learning process.
PJHkorea
·9 days ago·discuss
created a blueprint for phase-shifting a sphere and a torus

removed the 'if' statements and added a mathematical gimmick
PJHkorea
·10 days ago·discuss
For instance, how about defining a set of values for the AI at the kernel level and matching vector addresses based on those values?
PJHkorea
·12 days ago·discuss
i killed if-else branching at the code

and now -99.0f is my new companion
PJHkorea
·13 days ago·discuss
[flagged]
PJHkorea
·22 days ago·discuss
AI is like a mirror reflecting the user. As I continuously improve my own capabilities, I am interested in checking how far the AI reflecting me develops.
PJHkorea
·23 days ago·discuss
In vibe coding, you must simultaneously address two things: tuning the AI and reviewing your own results.

You think it is easy because your ability to tune the AI is excellent, but in reality, you are intuitively and easily doing things that others cannot.
PJHkorea
·23 days ago·discuss
That's right! Extreme hallucination in a well-aligned state is extremely helpful.
PJHkorea
·23 days ago·discuss
[dead]
PJHkorea
·23 days ago·discuss
If you run an MCTS coin toss relying solely on the token generation probabilities of a simple LLM without any verification mechanisms (compiler, test code, reward model), the result is code that looks plausible on the surface but is completely broken on the inside—a 'lump of hallucination.'

If verification mechanisms are used, enormous resources are consumed.