GoKawiil
1.
2.
Run Kimi K3 using 29 GB of RAM at 0.50 tok/s
(news.ycombinator.com)
3.
Show HN: Open-source engine running Gemma 4 26B in 2 GB RAM on any M-series Mac
(news.ycombinator.com)
4.
Non-Gaussian statistics of the order parameter across a phase transition
(feeds.nature.com)
5.
Ring-Zero: Scaling Zero RL to a Trillion Parameters for Emergent Reasoning
(news.ycombinator.com)
6.
Show HN: I implemented a neural network in SQL
(news.ycombinator.com)
7.
Reinforcement learning control of quantum error correction
(feeds.nature.com)
8.
9.
Puzzling Success of Overparameterization: Lottery Tickets or Escape Dimensions?
(news.ycombinator.com)
11.
I keep tripping over "true, false, true"
(news.ycombinator.com)
12.
I’ve banned query strings
(news.ycombinator.com)
13.
A Theory of Deep Learning
(news.ycombinator.com)
14.
Consequences of passing too few register parameters to a C function
(news.ycombinator.com)
15.
Which one is more important: more parameters or more computation? (2021)
(news.ycombinator.com)
16.
Ternary Bonsai: Top Intelligence at 1.58 Bits
(news.ycombinator.com)
17.
Are the costs of AI agents also rising exponentially? (2025)
(news.ycombinator.com)
18.
MegaTrain: Full Precision Training of 100B+ Parameter LLMs on a Single GPU
(news.ycombinator.com)
19.
Weight Transfer for RL Post-Training in under 2 seconds
(news.ycombinator.com)
20.
A single click mounted a covert, multistage attack against Copilot
(arstechnica.com)
21.
22.
Reproducing DeepSeek's MHC: When Residual Connections Explode
(news.ycombinator.com)
23.
24.
LLMs contain a LOT of parameters. But what’s a parameter?
(technologyreview.com)
25.
Titan’s strong tidal dissipation precludes a subsurface ocean
(feeds.nature.com)
26.
Post-transformer inference: 224× compression of Llama-70B with improved accuracy
(news.ycombinator.com)
27.
Show HN: Mirror_bridge – C++ Reflection powered Python binding generation
(news.ycombinator.com)
28.
Checked-size array parameters in C
(news.ycombinator.com)
30.