kinderpingui

4 ms·

1.
by kinderpingui 11mo ago
LeCun's recent talks and papers on JEPA (Joint-Embedding Predictive Architecture) and energy-based models. His core argument: autoregressive LLMs are fundamentally flawed because they predict in pixel/token space rather than abstr
2.
by kinderpingui 11mo ago
I realize your deep concerns & I want to indicate that I bear full responsibility for it.
3.
by kinderpingui 11mo ago
That graph is not very readable, you are right. Since I don't have a unit to measure the dog's behavior changes, this is reflected in the graph
4.
I accidentally discovered cross-species microbiome colonization in my dog(lightcapai.medium.com)
12 points by kinderpingui 11mo ago|8 comments
5.
Show HN: Interactive Transformer Architecture Designer with Emotion Analysis(wasda.ai)
1 points by kinderpingui 11mo ago|0 comments
6.
Mixture-of-Experts explained with PyTorch implementation(medium.com)
1 points by kinderpingui 11mo ago|1 comments
7.
by kinderpingui 11mo ago
I wrote this to explain how MoE achieves 64x more parameters with only 2x compute cost. The key insight is sparse activation - each token only hits 2 experts out of 8, while all expert weights stay loaded. Includes working PyTorch code for
8.
Show HN: Emotion-Weighting API with GPT-5(lightcapai.medium.com)
1 points by kinderpingui 11mo ago|0 comments
9.
AI Wealth Managers and Smart Contracts = Dead People Controlling Money Forever(lightcapai.medium.com)
2 points by kinderpingui 11mo ago|1 comments
10.
by kinderpingui 11mo ago
Technical deep dive into posthumous autonomous wealth management systems. The stack: Hadoop/Kafka for data processing, LSTM/Deep Q-Learning for trading decisions, Solidity smart contracts for enforcement, and perpetual trust laws