fp.
news
Open in hackernews
"The Matilda Effect": How Pioneering Women Are Written Out of Science History
https://www.openculture.com/2025/12/matilda-effect.html
6
•
coloneltcb
•
1h ago
Comments
ekjhgkejhgk
•
1h ago
I would've been funnier if they'd given this effect a male name.
AMD GPU Debugger
https://thegeeko.me/blog/amd-gpu-debugging/
127
•
ibobev
•
2h ago
•
8 comments
Strong earthquake hits northern Japan, tsunami warning issued
https://www3.nhk.or.jp/nhkworld/en/news/20251209_02/
139
•
lattis
•
4h ago
•
77 comments
Let's put Tailscale on a jailbroken Kindle
https://tailscale.com/blog/tailscale-jailbroken-kindle
87
•
Quizzical4230
•
2h ago
•
20 comments
Quanta to Publish Popular Math and Physics Titles by Terence Tao and David Tong
https://www.simonsfoundation.org/2025/12/08/quanta-books-to-publish-popular-math-and-physics-titl...
38
•
digital55
•
1h ago
•
4 comments
Hunting for North Korean Fiber Optic Cables
https://nkinternet.com/2025/12/08/hunting-for-north-korean-fiber-optic-cables/
106
•
Bezod
•
2h ago
•
9 comments
Launch HN: Nia (YC S25) – Give better context to coding agents
https://www.trynia.ai/
37
•
jellyotsiro
•
1h ago
•
28 comments
Flow: Actor-based language for C++, used by FoundationDB
https://github.com/apple/foundationdb/tree/main/flow
126
•
SchwKatze
•
5h ago
•
33 comments
Show HN: DuckDB for Kafka Stream Processing
https://sql-flow.com/docs/tutorials/intro/
17
•
dm03514
•
1h ago
•
5 comments
A series of tricks and techniques I learned doing tiny GLSL demos
https://blog.pkh.me/p/48-a-series-of-tricks-and-techniques-i-learned-doing-tiny-glsl-demos.html
20
•
ibobev
•
2h ago
•
0 comments
Legion Health (YC S21) is hiring a founding engineer (SF, in-person)
1
•
the_danny_g
•
1h ago
Microsoft Download Center Archive
https://legacyupdate.net/download-center/
13
•
luu
•
2d ago
•
0 comments
Nova Programming Language
https://nova-lang.net
46
•
surprisetalk
•
3h ago
•
26 comments
RIP Tetsu Yamauchi (Former Free and Faces Bassist)
https://www.loudersound.com/bands-artists/former-free-and-faces-bassist-tetsu-yamauchi-dead-at-79
15
•
pauseandplay
•
1h ago
•
3 comments
Microsoft has a problem: nobody wants to buy or use its shoddy AI products
https://www.windowscentral.com/artificial-intelligence/microsoft-has-a-problem-nobody-wants-to-bu...
242
•
mohi-kalantari
•
2h ago
•
189 comments
Colors of Growth
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5804462
38
•
mhb
•
5h ago
•
16 comments
The "confident idiot" problem: Why AI needs hard rules, not vibe checks
https://steerlabs.substack.com/p/confident-idiot-problem
253
•
steerlabs
•
3d ago
•
288 comments
IBM to Acquire Confluent
https://www.confluent.io/blog/ibm-to-acquire-confluent/
224
•
abd12
•
5h ago
•
181 comments
Turtletoy
https://turtletoy.net/
290
•
ustad
•
5d ago
•
52 comments
Twelve Days of Shell
https://12days.cmdchallenge.com
209
•
zoidb
•
8h ago
•
68 comments
Damn Small Linux
https://www.damnsmalllinux.org/
214
•
grubbs
•
17h ago
•
56 comments
Emacs is my new window manager (2015)
https://www.howardism.org/Technical/Emacs/new-window-manager.html
202
•
gpi
•
3d ago
•
76 comments
Berkshire Hathaway Announces Leadership Appointments [pdf]
https://berkshirehathaway.com/news/dec0825.pdf
57
•
kamaraju
•
3h ago
•
24 comments
How the Creator Economy Destroyed the Internet
https://www.theverge.com/cs/features/810002/influencers-creator-economy-special-series
41
•
ecliptik
•
2h ago
•
11 comments
I wasted years of my life in crypto
https://twitter.com/kenchangh/status/1994854381267947640
585
•
Anon84
•
1d ago
•
808 comments
Client-side GPU load balancing with Redis and Lua
https://galileo.ai/blog/how-we-boosted-gpu-utilization-by-40-with-redis-lua
39
•
lneiman
•
6d ago
•
6 comments
Bag of words, have mercy on us
https://www.experimental-history.com/p/bag-of-words-have-mercy-on-us
291
•
ntnbr
•
20h ago
•
317 comments
Show HN: Lockenv – Simple encrypted secrets storage for Git
https://github.com/illarion/lockenv
84
•
shoemann
•
11h ago
•
26 comments
Apex: Universal Markdown Processor
https://brettterpstra.com/2025/12/06/introducing-apex-universal-markdown-processor/
14
•
zdw
•
1d ago
•
4 comments
AI should only run as fast as we can catch up
https://higashi.blog/2025/12/07/ai-verification/
39
•
yuedongze
•
1h ago
•
46 comments
The C++ standard for the F-35 Fighter Jet [video]
https://www.youtube.com/watch?v=Gv4sDL9Ljww
316
•
AareyBaba
•
1d ago
•
391 comments
Load More
ekjhgkejhgk•1h ago