fp.
newest
The Launch Directory
https://thelaunchdirectory.com/
1
•
welsenesbros
•
4m ago
•
0 comments
iPhone Duo 1M fold test [video]
https://www.youtube.com/watch?v=vaiHWiExDXg
1
•
worotyns
•
9m ago
•
0 comments
Hindsight Capital's bet on a long, calm voyage pays off (2017)
https://www.ft.com/content/85f35528-e599-11e7-8b99-0191e45377ec
1
•
Betelbuddy
•
10m ago
•
0 comments
Raspberry Pi PicoW + LeRobot SO-ARM100 controlled via a webpage/mqtt/openai
https://github.com/AdrianCX/picohive-robotarm
1
•
AdrianCX
•
11m ago
•
1 comments
Using DocStrip for multiple document variants [pdf]
https://www.tug.org/TUGboat/tb41-3/tb129leingang-docstrip.pdf
1
•
kmstout
•
12m ago
•
0 comments
Struggling to Keep AI Out
https://blog.eunach.scot/struggling-to-keep-ai-out
2
•
latexr
•
16m ago
•
1 comments
Show HN: CaseCite – Free RAG caselaw research and document analysis for lawyers
https://github.com/thomasunise/casecite
13
•
thomasunise
•
16m ago
•
1 comments
Show HN: AgentTrust – Zero-trust security and human approvals for AI agents
https://github.com/SRN0129/Agenttrust
1
•
SRN0129
•
17m ago
•
0 comments
Why I'm Short the AI Buildout Even Though I'm Long Computers
https://unverifiablerewards.substack.com/p/why-im-short-the-ai-buildout-even
2
•
paulbaumgart
•
20m ago
•
0 comments
Don't Hinge Your Career Earnings on Technical Skills Alone
https://leveragethoughts.substack.com/p/dont-hinge-your-career-on-technical
2
•
maverickJ
•
21m ago
•
0 comments
Product Engineers
https://leerob.com/product-engineers
2
•
andsoitis
•
25m ago
•
0 comments
Ask HN: Where is GTAcraft VI?
3
•
tobwen
•
27m ago
•
0 comments
Ask HN: How fast is your CI these days?
3
•
easterncalculus
•
29m ago
•
1 comments
The Factory Outside Phoenix That Could Save American Manufacturing
https://www.nytimes.com/2026/10/23/opinion/ai-computer-chips-semiconductor-arizona-industry.html
2
•
rafaelc
•
30m ago
•
0 comments
The State of Context Engineering
https://www.dbreunig.com/2026/10/24/the-state-of-context-engineering.html
2
•
pchristensen
•
31m ago
•
0 comments
3D Gaussian Splatting for Real-Time Radiance Field Rendering (2023)
https://repo-sam.inria.fr/fungraph/3d-gaussian-splatting/
2
•
gone35
•
32m ago
•
0 comments
Show HN: MTA.rip: a train visualizer that shows congestion
https://mta.rip/?station=127&route=1&view=trackmap
3
•
fatlotus
•
34m ago
•
0 comments
A Comprehensive Introduction and Overview of Gaussian Splats
https://alphapixeldev.com/a-comprehensive-introduction-and-overview-of-gaussian-splats-part-1-int...
2
•
gone35
•
36m ago
•
0 comments
What goes into making an animated film?
https://www.disneyanimation.com/process/
11
•
embedding-shape
•
37m ago
•
3 comments
Information Security North Star – the least a project must know
https://informationsecuritynorthstar.com/?hn
2
•
splintersio
•
38m ago
•
0 comments
Some reasons to work on productivity and velocity
https://danluu.com/productivity-velocity/
2
•
andsoitis
•
40m ago
•
0 comments
I don't want to be an spectator
https://paulocuambe.github.io/posts/spectator/
2
•
gosenx
•
42m ago
•
0 comments
Plan and code review for your agents
https://plannotator.ai/
2
•
mooreds
•
42m ago
•
0 comments
Website to Markdown – keyless HTML→clean Markdown for RAG, pay-per-use
https://apify.com/noniner/website-to-markdown
2
•
fredxhansson
•
44m ago
•
0 comments
Show HN: MuscleMemory – procedural memory for coding agents, proven by tests
https://github.com/TheWariend/MuscleMemory
4
•
maaaz
•
44m ago
•
0 comments
A Stackless Floodfill Automaton
https://t3x.org/files/floodfill.pdf
2
•
andsoitis
•
45m ago
•
0 comments
.NET October 2026 Out-of-Band Security Updates
https://devblogs.microsoft.com/dotnet/dotnet-october-2026-oob-security-updates/
3
•
yread
•
45m ago
•
0 comments
NvALT Back from the Dead
https://brettterpstra.com/2026/10/24/nvalt-back-from-the-dead/
4
•
fork-bomber
•
46m ago
•
0 comments
Focused Company Operating Principles
https://focused.io/operating-principles
2
•
mooreds
•
47m ago
•
0 comments
I Let AI Port an Xbox 360 Game to PC (and It Worked) [video]
https://www.youtube.com/watch?v=rojjkW74d-k
1
•
Maximilian_Bet
•
47m ago
•
0 comments
Load More
Open in hackernews
Kauldron: Modular, scalable library to train ML models
https://github.com/google-research/kauldron
2
•
nateb2022
•
1h ago