Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
BerislavLopac
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
6 ms
·
61.
▲
Interviewing tactics for a post-LLM world
(blog.incrementalforgetting.tech)
2 points
by
BerislavLopac
7mo ago
|
0 comments
62.
▲
Keep Deterministic Work Deterministic
(oreilly.com)
9 points
by
BerislavLopac
7mo ago
|
1 comments
63.
▲
Context Engineering for Coding Agents
(martinfowler.com)
2 points
by
BerislavLopac
7mo ago
|
0 comments
64.
▲
by
BerislavLopac
7mo ago
Python, Linux and most other great open source projects we all use today started as personal projects.
65.
▲
In-Person vs. Remote Engineering Teams
(playfulprogramming.com)
1 points
by
BerislavLopac
7mo ago
|
0 comments
66.
▲
by
BerislavLopac
7mo ago
> A Python4 that actually used typing in the interpreter, had value types, had a comptime phase to allow most metaprogramming to work (like monkey patching for tests) would be great! It would be faster, cleaner, easier to reason about,
67.
▲
by
BerislavLopac
7mo ago
Any program written in Python of any significant size is literally a soup of multiple languages.
68.
▲
by
BerislavLopac
7mo ago
> python code could be so much faster if it didn't have to assume everything could change at any time Definitely, but then it wouldn't be Python. One of the core principles of Python's design is to be extremely dynamic, an
69.
▲
What Is Code Review For?
(blog.glyph.im)
1 points
by
BerislavLopac
7mo ago
|
3 comments
70.
▲
The Structure of Engineering Revolutions
(webdirections.org)
3 points
by
BerislavLopac
7mo ago
|
0 comments
71.
▲
by
BerislavLopac
7mo ago
> only the plot is copyrightable But the plot can't be copyrightable, as the copyright applies only to a tangible representation of an idea (e.g. written text), and not to an idea itself.
72.
▲
Can AI Kill the Venture Capitalist?
(wired.com)
8 points
by
BerislavLopac
7mo ago
|
0 comments
73.
▲
Design-First Collaboration
(martinfowler.com)
1 points
by
BerislavLopac
7mo ago
|
0 comments
74.
▲
by
BerislavLopac
7mo ago
Code is one thing, but what about writing? There is no 100% foolproof way to identify content written by LLMs, and human writing routinely gets incorrectly flagged as such. If I write a book, and a checker says that it's written by LLM
75.
▲
Py2many: Transpiler of Python to many other languages
(github.com)
1 points
by
BerislavLopac
8mo ago
|
0 comments
76.
▲
MKdocs Version 2
(github.com)
2 points
by
BerislavLopac
8mo ago
|
0 comments
77.
▲
The Future of AI
(lucijagregov.com)
152 points
by
BerislavLopac
8mo ago
|
111 comments
78.
▲
Russia soldiers tell BBC they saw fellow troops executed on commanders' orders
(bbc.co.uk)
14 points
by
BerislavLopac
8mo ago
|
1 comments
79.
▲
Are we having the wrong nightmares about generative AI?
(jessicahullman.substack.com)
6 points
by
BerislavLopac
8mo ago
|
2 comments
80.
▲
Who Opened the Door?
(chaosguru.substack.com)
2 points
by
BerislavLopac
8mo ago
|
1 comments
81.
▲
Elvish as She Is Spoke [pdf]
(elvish.org)
72 points
by
BerislavLopac
8mo ago
|
23 comments
82.
▲
Where Tech Leaders and Students Think AI Is Going
(wired.com)
2 points
by
BerislavLopac
9mo ago
|
0 comments
83.
▲
15min City
(whatif.sonycsl.it)
2 points
by
BerislavLopac
9mo ago
|
0 comments
84.
▲
The Next Thing Will Not Be Big
(blog.glyph.im)
1 points
by
BerislavLopac
9mo ago
|
0 comments
85.
▲
The Art of Engineering Management
(taoem.com)
2 points
by
BerislavLopac
9mo ago
|
0 comments
86.
▲
I Still Write Code as an Engineering Manager
(terriblesoftware.org)
2 points
by
BerislavLopac
9mo ago
|
0 comments
87.
▲
by
BerislavLopac
9mo ago
Yeah, I've been using "shit umbrella" for a long time. But the "transparent shit umbrella" is an even more powerful, albeit more disturbing, metaphor.
88.
▲
by
BerislavLopac
9mo ago
https://archive.is/20260125110952/https://www.wired.com/stor...
89.
▲
The Math on AI Agents Doesn't Add Up
(wired.com)
4 points
by
BerislavLopac
9mo ago
|
2 comments
90.
▲
by
BerislavLopac
9mo ago
> What's the source of truth? Nobody knows. Except everyone who ever used standards like OpenAPI and JSON Schema.
More ›