Blog

All ml networks

the macrobean philosophy

we were promised simplicity

Published 2025-12-05
software

Evolution of Internet Congestion Control

The Physics behind your internet speed: A 40-Year Journey

Published Nov 16, 2025
networks

Luminance blending for near-real-time bronchoscopic dehazing (Part 1)

a research note (my work-in-progress)

Published Sep 12, 2025
ml

What happened to xBERT and T5?

We don’t see any scaled-up BERTs anymore. The majority of LLMs (Claude, GPT, Llama..) are decoders. BERTs deprecated in favour of more flexible forms of T5, which are mostly complementary to decoders.

Published Aug 23, 2025
ml

LLMs are Data Compressors

To predict is to compress

Published June 21, 2025
ml

Serialization-not just any 0x01

How data is split into byte stream and my take on obscure binaries

Published June 8, 2025
networks

Network Protocols

Understating networks one packet at a time

Published May 28, 2025
networks

Socket Programming

Talked about the building blocks of computer networking and socket api

Published May 22, 2025
networks