2020#13 Readings
4 minutes read | 678 words by Ruben BerenguelThis week is Data+AI summit week.
I got a new iPad Pro 11 with M1 chip. It flies compared with my Mini 5. And the magnetic pencil is so nice.
đź“Ż Modelling data pipelines with Alloy
I showed how you can use 30-60 minutes of modelling in Alloy to find the stupid mistakes you could find anyway if you stopped and thought everything through. The difference? One rarely stops and thinks because this is easy, won’t fail of course.
🔊 Streetlights and Shadows
I found this book outstanding, with a lot of food for thought into how decision making is never easy.
🍿 Jazz piano explained in 20 minutes
I was aware of all of it already, but it’s done in 20 minutes. Impressive!
How Gravity Is a Double Copy of Other Forces
No Lie groups were harmed in reading or writing this post.
The Mortifying Ordeal of Pairing All Day
Luckily we only pair on occasion, when it seems to fit or we feel like it. The brain drain of pairing looks extreme in this case.
Game theory as an engine for large-scale data analysis
How PCA (Principal Component Analysis) can be modelled as the Nash equilibrium of a multi-agent game. This is the summary blog post, I’m extremely tempted to read the paper, even if my days as data scientist are far in the past.
Explaining the mechanics of Spark caching
This is a very in-depth explanation of caching. I learnt stuff from here, and to be fair, it’s not that common with Spark internals and usage combos.
Databricks is playing the long game in its battle against Snowflake
I wrote a bit around this here. I think Databricks will eventually have an advantage, but will depend on worldwide data needs: Databricks is strong in real-time, for example. If more real-time big scale processing is needed, Databricks has an upper hand on Snowflake.
Why Bad CEOs Fear Remote Work
I like the direct language of this piece.
How to (Actually) Save Time When You’re Working Remotely
On the contrary, data we collected from 12,000 people across the U.S. and Europe during the pandemic show that the additional time is often burned on unproductive work and unsatisfying leisure activities.
uxn virtual machine
These people are amazing.
How Black Pepper Won Europe From a Tastier Pepper
No particular reason why. By the way Spanish friends, I found out black pepper in the UK tastes far better than what we have here. Black pepper we can find in supermarkets is crap. I found a good one on Amazon Spain I could recommend.
The Evolution of Trust
This is a brilliant animated simulation of game theory scenarios that anybody can follow.
How HashiCorp Makes Writing a Priority
I’m not a fan of writing internal documentation (because very often it becomes obsolete faster than you can finish and is rarely used), but I’m a fan of writing to keep track of what has happened, why and the results. That’s invaluable.
Efficient SQL on Pandas with DuckDB
This sounds like the best approch for “SQL on Pandas” I have seen. DuckDBs performance is excellent.
The Lisperati1000 Is a Cyberdeck Terminal Dedicated to Lisp Programming
Given the large amount of tech gadges I own and do not use already, I want one of these too.
- Alloy
- Formal Methods
- Python
- Data science
- Data engineering
- Music
- Apache Spark
- Databricks
- Management
- ReadingsOfTheWeek
- Readings