A bit of time has passed since my last post. Many interesting things have happened in the tech world.
Gathered some of the most interesting resources that I read or watched in the last 2 months, although the list is not comprehensive and I might have not captured all of them.

Reading

Videos

Chatting with Robert Xiao from PPP | 023 CTF Radiooo

I haven't played CTFs seriously for the last 4 years or so (although still very interested in them, but no much free time anymore), but seeing the title of the video it was a no brainer that I want to watch it!
Robert Xiao is a well-known and very successful CTF player.

Much ado about noping - JF Bastien - NDC TechTown 2025

Another great talk from JF in his usual pleasant style. The subject is unique and I like the parallels he draws with other topics. Liked everything about nothing.

Modern Architecture 101 for New Engineers & Forgetful Experts

The talk is a great introduction to modern architecture and its principles. The presentation style is engaging and easy to follow.

On a side note, after watching this talk I realized that BigQuery (which is quite fast) uses columnar storage.
The underlying components of BigQuery are designed to efficiently handle large volumes of data and provide fast query performance. These components (high level) are:

• Dremel: The Execution Engine

  Dremel is the "brain" of BigQuery. It is a large-scale multi-tenant cluster that executes your SQL queries.

  Tree Architecture: Dremel turns your SQL query into an execution tree.

  Slots: The "leaves" of this tree are called slots. When you run a query, BigQuery can spin up thousands of these slots (compute units) in parallel to process your data.

  Mixers: The "branches" of the tree are called mixers, which aggregate the results from the slots and pass them up to the root to give you the final answer.

• Colossus: The Storage Layer

  BigQuery doesn't store data on the same machines that do the computing. Instead, it uses Colossus, Google’s global distributed file system.

  Durability: It handles data replication, recovery, and management across multiple data centers.

  Capacitor Format: Within Colossus, data is stored in a proprietary columnar format called Capacitor. By storing data in columns instead of rows, BigQuery only reads the specific columns your query asks for, which is why it can scan terabytes of data in seconds.

• Jupiter: The Network

  Since storage (Colossus) and compute (Dremel) are on different physical hardware, they need a way to talk to each other incredibly fast. Jupiter is Google’s petabit-scale internal network.

  Speed: It provides enough bandwidth to move data from storage to compute at rates exceeding 1 terabyte per second. This eliminates the "bottleneck" usually found in distributed systems when data moves across a network.

• Borg: The Orchestrator

39C3 - Asahi Linux - Porting Linux to Apple Silicon

39C3 - Hacking washing machines

BlueHat IL 2017 - John Lambert - Cyber in a World of Cloud

BlueHat India 2024: Day 1 Keynote with John Lambert

BlueHat IL 2018 - John Lambert - The New Paradigm of Security Controls

All three great talks by John Lambert. I like rewatching them. Highly recommending to read John Lambert's blog too: https://medium.com/@johnlatwc