This is a list of the best media I consumed last year.
AI and the limits of language – Yann LeCun & Jacob Browning
With the hype around large language models, the authors make a compelling argument about the relationship between thought and language. Language is not thought or sentience, and can never be human-like in any real sense.
For more about how large language models like GPT work, check out this paper for a detailed explainer and a comparison between Humans and Large Language Models. Also see this rebuttal to the idea that Google’s large language model is sentient.
Abandoning the view that all knowledge is linguistic permits us to realize how much of our knowledge is nonlinguistic. While books contain a lot of information we can decompress and use, so do many other objects: IKEA instructions don’t even bother writing out instructions alongside its drawings; AI researchers often look at the diagrams in a paper first, grasp the network architecture and only then glance through the text; visitors can navigate NYC by following the red or green lines on a map.
What does a Batsman See? – SB Tang
This article is ostensibly about cricket (It’s in a magazine called Cricket Monthly!). However, I read it as a metaphor for life. Experts and amateurs look at the same things and see different things. What does a professional batsman see? I used to think that reflexes were about quick responses after observing an action. In reality, at the highest level, the fastest reflexes are about action after accurate prediction. It’s predictive, not reactive. They don’t have faster muscles, they have better ability to predict what’s going to happen, faster and quicker than anyone else, and act on that information.
Well, science now tells us that elite batsmen aren’t much different: they know where the ball is going to be before it gets there and saccade their vision to that point. That’s how they appear to have such quick reflexes.
If you build products for a living, you should read this book. In the interviews leading up to the release of the book, Fadell said how this is a collection of answers to the questions he gets asked the most. And boy, does it deliver. Tony Fadell has strong opinions about everything – hiring, the integration of product development and marketing, curating and managing a comprehensive product experience and much more. You won’t always agree with everything, but you’ll have to think carefully why you disagree with someone who helped create the iPod and the wildly successful Nest line of products.
Reversion to the mean – the real long Covid – John Luttig
A thoughtful note in May 2022 that foreshadowed the wave of technology layoffs through the end of the year. Despite Sequoia’s predictions of doom, the pandemic ended up being a good thing for founders and entrepreneurs, at least in the short term. More people were buying things online, and valuations were up and to the right to the new normal. However, by May 2022, most of those gains had been erased.
I’ve included my favorite chart from the article below.
There’s so much more in the article about what this means for investors and entrepreneurs. It appears, only a small subset of these long-term trends are permanent, if any.
- Sandy Kempner on Fintech One-On-One podcast: This podcast opened my mind to the amazing world of working capital financing. The more credit is available, the more valuable “niche” solutions can be. I’m also fascinated because the product has its own growth loop embedded in it.
- So You Want to be the next Warren Buffet, How’s your Writing? : Investing is hard, we’re doomed to suck at it. And good work only comes from doing something consistently over a long period of time. And luck.
- My First Impressions of Web3: One of the more balanced takes on Web3 in a long time. You don’t have to agree with everything, but participants in this ecosystem must have a point of view about the points raised otherwise, they’re choosing to have their heads in the sand.