|
|
|
|
|
|
“We do not learn from experience; we learn from reflecting on experience.”
—John Dewey
|
|
|
|
|
|
|
If you enjoy or get value from The Interesting Times, I'd really appreciate it if you would support it by forwarding it to a friend or sharing it
wherever you typically share this sort of thing - (Twitter, LinkedIn, Slack groups, etc.)
|
|
|
|
|
Paul Graham
A classic Paul Graham essay I found myself re-reading. Some bits that stuck out to me reading this reading:
On how your professional choices impact your kids:
“If you take a boring job to give your family a high standard of living, as so many people do, you risk infecting your kids with the idea that work is boring. Maybe it would be better for kids in this one
case if parents were not so unselfish. A parent who set an example of loving their work might help their kids more than an expensive house.”
On what to do when you’re not sure what you want to do:
“Another test you can use is: always produce. For example, if you have a day job you don't take seriously because you plan to be a novelist, are you producing? Are you writing pages of fiction, however
bad? As long as you're producing, you'll know you're not merely using the hazy vision of the grand novel you plan to write one day as an opiate. The view of it will be obstructed by the all too palpably flawed one you're actually writing.
"Always produce"is also a heuristic for finding the work you love. If you subject yourself to that constraint, it will automatically push you away from things you think you're supposed to work on, toward
things you actually like. "Always produce"will discover your life's work the way water, with the aid of gravity, finds the hole in your roof.”
I find the "always produce"heuristic is a clever diagnostic tool. If you claim to be an aspiring novelist but never write, you're probably using the identity as an escape rather than pursuing the work. Production forces honesty about what you actually enjoy doing versus what you enjoy
imagining yourself doing.
|
|
|
Dwarkesh Podcast
The fundamental problem with LLMs today is that they don't get better over time the way a human would. Current AI systems are like extremely competent consultants who forget everything between meetings. Yes, you can build custom prompts and keep updating them and that is a material improvement but fundamentally has some limits.
Consider the analogy of learning to play the saxophone. Human learning happens through iteration, feedback, and
gradual refinement. Training AI works like sending away each failed student and refining instructions for the next one. No matter how well honed your prompt is, no kid is just going to learn how to play saxophone from reading your ten thousand word prompt.
“LLMs don’t get better over time the way a human would. The lack of continual learning is a huge huge problem. The LLM baseline at many tasks might be higher than an average human's. But there’s no way to give a model high level feedback. You’re stuck with the abilities you get out of the box. You can keep messing around with the system prompt. In practice this just doesn’t produce anything even close to the kind of learning and improvement that human employees experience.
The reason humans are so useful is not mainly their raw intelligence. It’s their ability to build up context, interrogate their own failures, and pick up small improvements and efficiencies as they practice a task.”
For me, the current models feel like powerful engines that lack a transmission system to convert power into useful work. They are amazing, but still require quite a lot of babysitting in my experience so far.
I don’t have enough knowledge to understand how this limit could be broken through, but it seems significant and there is some question about what type of training data would be needed for this and if we have it.
“We don’t have a large pre-training corpus of multimodal computer use data. … For the past decade of scaling, we’ve been spoiled by the enormous amount of internet data that was freely available for us to use. This was enough for cracking natural language processing, but not for getting models to become reliable, competent agents. Imagine trying to train GPT-4 on all the text data available in 1980—the data would be nowhere near enough, even if we had the necessary compute.”
To some extent, technological revolutions are limited not by peak performance but by consistent operation. The steam engine existed for decades before it became economically transformative. This wasn’t because engineers couldn't build powerful engines, but because they couldn't build reliable ones.
AI is amazing for tasks where being 80-90% accurate is good enough. This is a large corpus of tasks but few people would stay in a job long if they fucked it up 10-20% of the time and never improved.
Perhaps we're not waiting for smarter AI, we're waiting for AI that gets smarter.
AQR
Most of what investors perceive as American corporate superiority is actually just valuation expansion dressed up as fundamental strength.
Since 1990, roughly 4% of the annual U.S. outperformance over international markets came from relative repricing rather than actual earnings growth. By December 2024, U.S. market valuations had reached nearly twice those of non-U.S. developed markets—a historically extreme divergence that would require a 45%
drop in U.S. prices (or commensurate rise in international markets) just to return to parity.
When U.S. stocks outperform, it feels like confirmation of American innovation and entrepreneurial superiority. The Magnificent Seven's dominance over European markets becomes evidence of tech leadership rather than what it partly is: a re-rating of future expectations. Investors often seem to mistake the result of higher valuations (better returns) for the cause of those valuations (better fundamentals).
Valuation-based predictions have actually proven more reliable than growth-story extrapolations over longer time horizons. The correlation between relative valuations and subsequent decade-long performance is modest (+0.5) but meaningful given the noise in financial markets.
To be sure, the US growth story isn't illusory. The U.S. has maintained roughly a 1% annual earnings growth advantage over international markets across long time periods. But,
markets appear to be pricing in something closer to a 2.2% annual growth edge based on current valuations. In essence, markets are not merely pricing historical levels of US exceptionalism but EVEN MOAR exceptionalism than has historically been the case. For the current valuation gap to persist, it is not enough to ‘Make America Great Again’ - it must become even greater than ever.
The Intimate Mirror
A beautiful write-up on what constitutes healing: Healing happens when suffering receives "loving attunement"in the presence of safety.
"At the heart of every movement of healing lies this truth: when that which suffers receives loving attunement, healing naturally follows. It's a fundamental pattern that emerges consistently across contexts and approaches. The movement toward
wholeness depends on just two essential ingredients: sufficient safety for what's been fragmented or disowned to emerge into awareness, and contact with the quality of loving presence that is, in fact, our fundamental nature. When these two elements come together, transformation isn't something we do—it's something that happens through us, as naturally as a river finding its way to the sea or a flower opening to the sun."
The failure to receive this attunement is the root of suffering:
“...suffering emerges from a failure of love—or perhaps more precisely, a failure of attunement. Suffering occurs when some movement of reality fails to receive the loving attunement it requires to be integrated with the whole.
This is a lived reality we can observe everywhere. A child's tantrum represents their attempt to receive attunement for an emotional state that feels overwhelming. Addiction masks a desperate attempt to soothe parts of ourselves that never received adequate care. Even physical symptoms can reflect aspects of our being crying out for the attention and love they've been denied.”
This framework explains the inefficiency of our analytical approach to healing.
"when faced with suffering, it immediately wants to solve, to fix, to deploy the right strategy. We want healing to be complex because complexity is what our analytical minds know how to work with.
Chris Arnade on Walking Cities [Podcast]
Conversations with Tyler
By far, my favorite activity when visiting a new city is to go for a walk. Most ways of understanding places are not as good as simply moving
slowly through space and letting strangers tell their stories.
Arnade spent his first career as a bond trader on Wall Street, which gave him "the view from the Ramada Inn"—flying in, staying in wealthy neighborhoods that are variations on the same theme everywhere. This is a certain top-down view of the world. He has made a second career as a journalist, going on walks nowhere near the Ramada Inn with a certain bottom-up view.
On where the food is actually good (I especially endorse Vietnam!)
“...one of the dirty secrets is the food you get in a place is not that much better than you can get in the U.S.
The exceptions being France, Japan, Korea, and Italy. … I get better—I got better Indonesian food in Netherlands than I did in Indonesia by far. Vietnam, I'd also put up there as a place where the food is exceptionally better relative than what you get outside of Vietnam.
And on why El Paso is the most optimistic city in America:
“Tyler: Why do you like El Paso so much?
Chris: The optimism. It the American dream. So I think the American dream is very much alive in the working class Mexican-American community. And you see that in El Paso. Like you don't have, you know, when I was doing my project on addiction and poverty, El Paso is just fundamentally different. You know, you don't have the despair that you have in places. And a low crime rate, too. Extraordinarily. In some senses, Mexico acts as a roach motel. If you're going to do crime, go over to Juarez. So consequently, there's no crime in El Paso. But it's one of the most optimistic cities in the United States. It has amazing food, by the way.”
|
|
|
|
|
As always, if you're enjoying The Interesting Times, I'd love it if you shared it with a friend (or three). You can send them here to sign up. I try to make it one of the best emails you get every week and I'm always open to feedback on how to better do that.
If you'd like to see everything I'm reading, you can follow me on Twitter or LinkedIn for articles and podcasts. I'm on Goodreads for books. Finally, if you read anything interesting this week, please hit reply and send it over!
|
|
|
|
|
The Interesting Times is a short note to help you better invest your time and money in an uncertain world as well as a digest of
the most interesting things I find on the internet, centered around antifragility, complex systems, investing, technology, and decision making. Past editions are available here.
|
|
|
|
|
Here are a few more things you
might find interesting:
Interesting Essays: Read my best, free essays on topics like bitcoin, investing, decision making and marketing.
Consulting & Advising: Are you looking for help with making decisions around scaling your company from $500k to $5 million? I’ve been working
with authors, entrepreneurs, and startups for half a decade to help them get more out of their businesses.
Internet Business Toolkit: An exhaustive list of all the online tools I use to be more productive.
|
|
|
|
|
