372: Constellation & Topicus, Search Engine + LLMs, R&D, Tinder Scams, Sony + Honda EV, Leo Szilard, and Dark Matter

"pretty soon science would dry out"

Our reality is similar to the brain-in-a-vat hypothesis; we live in a simulated world, but it is not in a computer—it is in our head. 🧠 –Jeff Hawkins

🤖⛓️📈 One thing to remember about AI capabilities is that it’s a ratchet that only goes up.

Whatever you’re seeing right now is the worst that it’ll ever be.

The *rate* of progress may fluctuate a lot — it certainly has in the past — but the direction is known.

😮‍💨 I’m fried.

The kids have been home since December 23rd, and the oldest is only going back to school tomorrow. Over 2 weeks without much alone-time is a lot for an introvert such as myself. There have been plenty of good family moments, but I’m exhausted. 🪫

It would be interesting to do a survey to find out who comes out of a vacation feeling refreshed and replenished, and who comes out needing to recuperate from the vacation 🤔

I suspect the biggest factors must be extroverts vs introverts (I define these based on whether you recharge your batteries when you’re alone or with other people) and parents of rambunctious young kids vs non-parents/parents of older, low-maintenance kids.

🛀 It's easy to forget how influential Hewlett-Packard (HP) was in its heyday.

👩🏻‍🚀🚀 📖 On Friday, my wife started listening to the audiobook of Project Hail Mary (I wrote my thoughts about it in edition #370). I was a bit nervous that she wouldn't like it — you never know, everyone has different tastes — but so far she loves it!

It’s always fun to share something you like with someone else, it’s a way to relive it through them 🙂

🤖🛀😬 Humans are often cringe-worthy, so when you train AI on human-data, you'll get a lot of cringey stuff — that doesn’t make the AI dumb, it makes it a faithful mimic.

💣🤪🤰🏻📉 Noah Smith has written a good piece explaining why Paul Ehrlich got everything wrong and somehow still gets asked to predict the future — mostly because he tells a segment of the population exactly what they want to hear, and he himself cares more about the attention and prestige than being correct:

• 🐇 Why Paul Ehrlich got everything wrong

Some ideas become extremely viral and it takes decades and decades for them to be purged from most people’s minds. A lot of Ehrlich’s ideas are extremely damaging to humanity and should be debunked vigorously and regularly.

🏋️‍♂️👨🏻‍🍳 Here’s a good video that makes a point about fitness and cooking, but I think it generalizes to other domains too: For some things, many different approaches work and it’s very hard to know which one is better overall, and which one is more suitable for *you*.

There are many factors that are hard to isolate, and even scientific studies can have weak conclusions, poor methodology, or be restricted in what they do (hard to ruin the health of some control group to figure out if something is good/bad for you).

The last segment of the video on “What are you training for?” is also very important.

Don’t put off doing important things in your life to prepare for less important ones!

It may seem obvious when phrased like that, but I think we can all be guilty of that from time to time.

---

🏦 💰 Liberty Capital 💳 💴

🌸 Topicus/TSS Flower Graph of Acquisitions 🌺 + ✨Constellation Software✨ Acquisitions 2005-2022