2021#23 Readings
5 minutes read | 974 words by Ruben BerenguelAs usual, skipping an edition means a longer collection later on.
📯 Concept Maps
This is one from me, I hope you like it. Concept maps were originally introduced as a teaching/learning tool (particularly for kids) by Joseph Novak, and have been later used by the cognitive task analysis people (cf. Gary Klein) to establish baselines for expertise.
Data & Data Engineering — the past, present, and future
From Linkedin’s data influencer Zach Wilson. There’s many things missing in this story (like, the origin of spreadsheets, or Apache Pig) but it covers many years in a reasonable space.
📚 Memory Craft
This is one of the best memory books I have read. Not only for the techniques (which I knew), but the whole historical context. Techniques used all over the world since ancient times.
Sustainable coffee grown in Finland
This sounds out of science fiction movies.
How Lessonly, Bumble and HubSpot fight burnout
The time off politics of the US are pretty much crazy. What is oh-so-awesome in thi article is basically law here.
How I make CSS Art
Very creative!
Python as a build tool
It was for a Java project. It makes sense, if you can’t bend the usual tools (maven, gradle, make) to do your bidding, roll your own. I’m surprised bazel wasn’t tried, from my (extremely) low experience with bazel I’d expect it to support solving this problem.
Airbyte — Worth the hype?
It seems like an excellent tool for initial setups. We do have something similar internally, but built by us and with all the bells, whistles and corner cases that we need. Because in the end, corner cases are where the magic sauce is.
Red Hot: The 2021 Machine Learning, AI and Data (MAD) Landscape
Quite long but very interesting. Probably the point that caught my attention the most is Neo4j, a graph database provider, raised $325M at a more than $2B valuation. That’s the same valuation as Redis Labs (which, support, of course, Redis).
Anaconda Announces Partnership with Snowflake to Bring…
There’s not a lot of detail, but would seem as if Snowflake wants to have ML built-in to prevent Databricks from eating them in that area. Not sure about this, snakes don’t like cold weather.
Fast number parsing in Scala
Use toIntOption in Scala 2.13+!
Generics in Go
Yet another post about Go generics. I’m thinking of problems I’d want to solve with Go using them so I can polish a bit my “modern Go”.
It takes a PhD to develop that
Or not.
The state of the AWK
I have mentioned that I like AWK a lot. It is far superior to bash in similar situations, since it actually looks like a programming language.
Haskell for all: The “return a command” trick
I prefer an initial encoding to a final encoding. I want to have the intermediate structure to log it, introspect it, analyze it, etc.
The Emotional Needs of Dashboarding
This seems to match very well my experience. At my previous work we had a very small team so any requirement for a new kind of report having to look back eons or a dashboard for a really complex set of data was heavily pushed back, trying to figure out why it was needed. And quite often we could convert a significant amount of work into basically an alert email, or a ping when something changed behaviours. Or a one-off request instead of a recurrent, expensive job.
Working From Orbit. VR Productivity
I tried working in VR as a gimmick soon after getting a Quest and found the headset too heavy for comfort, even with the elite strap, which distributes weight better. Also, the Quest LED panels feel too bright for me. But of course the screen real state you get is freaking awesome.
The Fastest Path to the CEO Job, According to a 10-Year Study
I don’t find the results surprising at all. You learn way more from experience than by credentials.
How to Execute Pandas Workloads in a Distributed Manner With Apache Spark
I didn’t think project Zen/Koalas were going to pan out. Of course I didn’t think that Databrick needed it to grow its marketshare.
And You Will Know Us by the Company We Keep
How graphs shape behaviours (particularly online).
Speed matters
This reminds me of Wardley mapping in a way. Once something becomes a commodity (for example, deep learning with TPUs/GPUs) new possibilities open, like self-driving cars or dog-or-muffin. Higher execution speed is thus similar.
Always do Extra
Extra is what gets you to… 10x? Maybe.
Some reasons to work on productivity and velocity
There are many interesting ideas here.
Every Leader Has Flaws. Don’t Let Yours Derail Your Strategy
The problem is then identifying what kind you are. Which is kind of the hardest point, after all.
Our $16M Series A - Hex Blog
I learnt about Hex in Christophe Blefari’s newsletter and the animated GIF midway this post is intriguing and fascinating.
Harvesting ‘true cinnamon’: The story of the Ceylon spice
The part about the colonies time is horrible.
