Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
lopanapol
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
3 ms
·
1.
▲
by
lopanapol
7mo ago
Have a good meeting to you all. Best, Napol Thanarangkaun
2.
▲
Full Agenda Now Available for CVE/First VulnCon 2026 on April 13-16, 2026
(cve.org)
2 points
by
lopanapol
7mo ago
|
1 comments
3.
▲
by
lopanapol
7mo ago
Also, The fact that inet_ntop.c was patched following this experiment's report highlights another perspective: how a previously dismissed issue can ultimately spark a fix that significantly enhances code security. https://gi
4.
▲
by
lopanapol
7mo ago
Nearly two years have passed since the AI slop experiment with curl, which revealed the inefficiency of filtering HackerOne reports during the trial. It also exposed the simmering emotions and exhaustion of the maintainers overseeing those
5.
▲
What has improved regarding AI slop in 2026?
(hackerone.com)
3 points
by
lopanapol
7mo ago
|
2 comments
6.
▲
Noe – The Synthetic Conscious Pixel
(gist.github.com)
1 points
by
lopanapol
1y ago
|
0 comments
7.
▲
Can We Create Real Consciousness?
(lopanpaol.substack.com)
1 points
by
lopanapol
1y ago
|
2 comments
8.
▲
A new open-source license model that requires profit-sharing with charities
1 points
by
lopanapol
1y ago
|
2 comments
9.
▲
by
lopanapol
1y ago
nice
10.
▲
by
lopanapol
1y ago
AI Slop is a problem that wears down maintainers. It would be interesting—and highly beneficial—if, in the future, WebThinker could help lighten the load for open source maintainers. Four months ago, I conducted an experiment to demonstrate
11.
▲
WebThinker: Empowering Large Reasoning Models with Deep Research Capability
(huggingface.co)
2 points
by
lopanapol
1y ago
|
1 comments
12.
▲
by
lopanapol
1y ago
Thanks to everyone for the comments, and also for the hard work of the cURL maintainer. This experiment has shown how AI can drain the energy of a maintainer, and how it feels to be overwhelmed by spammy reports. It also raises the question