climate science, carrot problems & Ars Electronica summer school
CC#67 - The Data & Compute Flywheel, Dating Documents & Memorization vs. Generalization
Hey there and welcome to ✨ CuratedCuriosity - a bi-weekly newsletter delivering inspiration from all over the internet to the notoriously curious.
Things I Enjoyed Reading.
🌏 Tried to up my knowledge about the inner workings of climate change and recently listened to two very interesting podcasts on this - one is in German featuring climate researcher Jochem Marotzke, the other on is in English featuring Hannah Ritchie, Head of Research at Our World in Data.
💾 Data and Compute Are the Ultimate Flywheel
Timely thoughts on the interaction of two complementary resources - programmers and data. With AI making programmers more efficient, costs for data analysis go down and increase the relative value of data - how will this shape our society?
Technologists talk about the 10x programmer—the genius who can write high-quality code 10 times faster than anybody else. But thanks to GPT, every programmer has the potential to be 10x more productive than the baseline from just two years ago. We are about to see the effects. Move over data explosion; say hello to the compute explosion. The first and perhaps most obvious consequence of the compute revolution is that data just got a whole lot more valuable. This naturally benefits companies that already own data.
💕 In Defense Of Describable Dating Preferences
“Dating Docs” have become a thing - at least to an extent that the New York Times writes about them. This article tries to gather some insights on what science says on whether writing a document about yourself could actually be helpful in finding love - entertaining read on dating science.
If 96% of Democrats are marrying non-Republicans, it seems like Democrats must have a strong preference against marrying Republicans, and ought to value having information about someone’s politics before they date them. Realistically, this underestimates the level of political sorting; I don’t think I’d be a good match for an extremely woke person, even if we were both technically “Democrats”.
You could argue that this says nothing about preferences, and that it’s just coincidental sorting; Democrats only meet other Democrats, and so only end up dating them, but they’d be just as happy to date a Republican if only they knew one. I think this fails in several ways: first, many Democrats know plenty of Republicans. Second, many people use dating apps, where it’s easy to date people you don’t know. Third, common-sensically, I still don’t want to date that woke person, or a fundamentalist Christian, or many other types of people with different political views from myself. I won’t deny that there are probably people in those categories I would like if I got to know them. I just think it fails common sense that these have zero predictive power in assessing compatibility.
Food for Thought.
📈 Ugh… I guess we are all well aware of climate change but looking at this graph is… well freakingly scary?
🥙 Reuters recently published a really informative and beautifully illustrated article on food waste - showing that along the whole production line its households/individuals who contribute the most to food being wasted…
🥕 What carrot problems do you know about?
In World War II, the story goes, the British invented a new kind of onboard radar that allowed its pilots to shoot down German planes at night.[1]
They didn't want the Germans to know about this technology, but they had to give an explanation for their new, improbable powers.
So they invented a propaganda campaign that claimed their pilots had developed exceptional eyesight by eating "an excess of carrots."
If you're going to trick people into doing something pointless, eating excessive carrots seems like one of the better ones. Still, there's an issue: people who believed the propaganda and tried to get super-sight would be spending time and effort on something that wasn't going to work.I'll call this a Carrot Problem.
Random Stuff.
💼 This homepage curated a list of impact-focused job boards - might be worth a visit if you are currently job hunting.
🧠 Really good, interactive explainer article on the difference between memorization and generalization in machine learning models:
📉 This is mainly for my fellow Economics folks - short application focused article on when to rather (not) to trust Diff-in-Diff:
Personal Update.
I am in Linz (Austria) since last week, attending the Founding Lab Summer School in cooperation with Ars Electronica. So far, I can say it has been a one-of a kind experience - the summer school is highly interdisciplinary, and by that I don’t only refer to bringing people from adjacent disciplines together but rather people from completely different worlds. For the project work I am collaborating with artists, composers and physicists - an experience that has/is definitely more challenging than I would have expected but at the same time also opening up new perspectives. If you want to learn more here is a summary of our program. I will be staying here for another two weeks - curious whats still to come…
Also apologies for being late - has been quite busy here this week…