HNHacker News
TopNewBestAskShowJobs

Xx_crazy420_xX

115 karma · joined August 20, 2023

meet.hn/city/pl-Katowice

Interests: AI/ML, Blockchain, Cybersecurity, Data Science, Entrepreneurship, Fitness, Hacking, Martial Arts

---

submissionscomments
Xx_crazy420_xX··on OpenAI – vulnerability responsible disclosure
Did you try to ask it to provide data of the company, by explicitly invoking hallucination in the model?

Right now there is no real proof, untill you confirm that the data it provided cannot be hallucinated (which could be not feisable).

Also, acknowledging the response fron OpenAI staff dismissing it, would you mind sharing PoC?

Xx_crazy420_xX··on Curate your shell history
The preference of shell history depends on the use case. In my case it's ever changing projects with different tech stacks that i need to remember set of commands to get SSO to get to container registry etc..

The solution for me was storing bash history separately for each vscode project - https://gist.github.com/Srakai/9f3788cdb07259d65d335ff150c38...

Scripts are cool, but imagine ever evolving script that is controlled by arrows and enter

Xx_crazy420_xX··on The NSA Selector
It would be nice to compare all packets dumped by system and check if there are any unmatching packets running from your computer that were not sent by OS but a closed source hardware inside like Intel ME
Xx_crazy420_xX··on MCP server for Ghidra
Cool to hear you've worked on ARC-AGI — I poked with it too. You’re totally right about the messy traversal space, especially with parallel branches. What feels ambiguous at the token level becomes structured ambiguity in the AST — and that’s progress.

My hunch is that LLMs don’t need to solve the whole traversal space — they just need a clean, abstract interface. Even parallel branches can be normalized into a schema that the model can reason over consistently. And in practice, you rarely need full recursion or a complete tree walk to understand a node — but having that option unlocks deeper comprehension when it counts.

This kind of structural understanding would also massively improve Copilot-style tools, especially for less popular libraries where token-level familiarity breaks down. If models could reason over types and structure instead of guessing based on frequency, completions would be a lot more reliable outside the top 1% of APIs.

Xx_crazy420_xX··on MCP server for Ghidra
For LLMs to solve code I think they should be AST-native. Code is a tree, not a sequence — yet we feed it to models linearly, with no explicit structure. Todays models lack recurrence or true memory, so they can’t reason over hierarchical structures effectively.
Xx_crazy420_xX··on 80% of AI Projects Crash and Burn, Billions Wasted Says Rand Report
80% of AI projects crash, along with this site
Xx_crazy420_xX··on Notebooks Are McDonalds of Code
Notebooks are not IDE, so cannot be compared to ones. You can run notebook in vscode, and keep the best code practices. Experiment with code, incrementally move code chunks to external modules. Another great feature is ability to eaisly switch kernel to remote one without leaving IDE. Downside is that it can be really annoying to manipulate the code.
Xx_crazy420_xX··on [dead]
Hey everyone,

Ever found yourself wondering just how many cells are in a chicken? Or how big the genome of a corn plant is? What about which DNA strain is the most "successful" on Earth? Same here. So, I decided to roll up my sleeves and bring an old domain back to life: dnaleaderboard.com.

Keep in mind, these are estimates, but they offer a fascinating snapshot of life's diversity on this planet.

So head over to dnaleaderboard.com and let me know what you think. Super keen to hear your feedback, and get ideas into improving it (ideally around getting dat from better sources than ChatGPT :))!

Xx_crazy420_xX··on Material UI (MUI) is hot garbage
Mostly everything. If you get to know some MUI components you will know what you want, then just iterate using GPT, for me works surprisingly well. I'm also comboing it with copilot for the dirty work.
Xx_crazy420_xX··on Material UI (MUI) is hot garbage
In my experience, using MUI helps me avoid hitting token limits on ChatGPT, unlike when I use normal CSS, because it adds another abstraction layer.
Xx_crazy420_xX··on [dead]
Hey everyone,

Ever found yourself wondering just how many cells are in a chicken? Or how big the genome of a corn plant is? What about which DNA strain is the most "successful" on Earth? Same here. So, I decided to roll up my sleeves and bring an old domain back to life: dnaleaderboard.com.

Keep in mind, these are estimates, but they offer a fascinating snapshot of life's diversity on this planet.

So head over to dnaleaderboard.com and let me know what you think. Super keen to hear your feedback, and get ideas into improving it (ideally around getting dat from better sources than ChatGPT :))!

← PreviousPage 2 of 2