March 2, 2025

Daily Log Digest – Week 8, 2025

2025-02-23

What is good coffee

What Is "Good" Coffee? The Dark Side Of Flavour! - YouTube #coffee #flavor #taste

Great video on how subjective coffee tasting is, and how you may not be crazy for not liking that expensive packet of beans.

ChatGPT Summary: ChatGPT - Coffee Quality and Preferences

Garbage

Garbage — Ridgeline issue 203

Craig Mod has a great piece on social norms around garbage in Japan. Most Japanese public spaces don't have a lot of garbage bins and you are supposed to carry your own garbage.

The first time I walked into a random shop in Tokyo and asked to throw away something (a Starbucks cup, perhaps? an item I did not buy from the shop itself) was twenty-five years ago. The owner looked at me like I had just asked him if I could jump on his desk and take a shit. I’ve never bothered a shop with my garbage since.

…

In Kamakura, Starbucks has big signs instructing non-Japanese customers to please not leave their take-away cups in random locations. (Apparently this was becoming endemic.) There are no garbage cans in Kamakura, and, indeed, if you are buying a coffee to go, you will be responsible for that receptacle for, potentially, a very long time. This is your grandé-sized hair shirt to bear.

…

This obsession with the immediate “unburdening” of a thing you created is common in non-Japanese contexts, but I posit: The Japanese way is the correct way. Be an adult. Own your garbage. Garbage responsibility is something we’ve long since abdicated not only to faceless cans on street corners (or just all over the street, as seems to be the case in Manhattan or Paris), but also faceless developing countries around the world. Our oceans teem with the waste from generations of averted eyes. And I believe the two — local pathologies and attendant global pathologies — are not not connected.

…

Personally, I don’t love carrying my garbage around with me, but I recognize that it wouldn’t exist without my intervention. Nobody ran up and asked me to hold an empty cup. I thoughtlessly bought something. Thoughtlessly consumed it, and now I have to hold onto the detritus for a little while? Great. It’s easy. Easy to embrace that modicum of responsibility for your own waste. This is my protest song, the world’s lamest: I will attend to my garbage without complaint. Maybe give it a try next time you’re in Japan? It’s very exciting — to realize you will not be killed by your garbage, that holding a Snickers’ wrapper will not drain your crypto reserves, that not having piles of everyone else’s garbage all around is quite a nice bonus when walking through a city. And it might just keep you from buying unnecessary junk.

Ideological Propaganda by Way of the Algorithm

how the algorithm keeps you under control - by Adam Aleksic #algorithms #ideology #culture #sociology

In their 1947 Dialectic of Enlightenment, the philosophers Theodor Adorno and Max Horkheimer voiced their concerns about how the popular media of the time_—film, radio, and magazines—_primarily functioned to pacify the general population.

To them, the “culture industry” was a tool to both occupy our senses and influence our attitudes toward the world. Through the mass production of entertainment, the media-makers are able to dominate people’s leisure “from the time they leave the factory in the evening to the time they clock in again the next morning.” The flat, repetitive nature of the content, meanwhile, enforces social structures by pushing the same conformist narratives, and the movie-goer never questions anything since he “sees the world outside as an extension of the film he has just left.”

Adorno and Horkheimer would probably be losing their shit today.

…

Notably, all of this content is user-generated. There’s no bogeyman imposing cultural messaging from the top down. Rather, conformity is ingrained into the very structure of social media. The act of participating on TikTok, for example, schematizes certain assumptions like valuing follower counts or view counts. This ties one’s self-worth to what goes viral on the algorithm, incentivizing the creation of ever more content. ch If you as the viewer enjoy a meme, you mentally legitimize the algorithm that brought it to you. If you engage by liking or commenting, you even help it crowdsource information about the type of audience that should receive that meme in the future. To exist on social media at all is to opt into a technofeudalistic fiefdom where we individually and collectively feed platforms the information they need to keep us docile.

Thoreau's idea of success

How to REALLY Avoid Living a Life of Quiet Desperation | The Art of Manliness #masculinity #self-help #self-improvement #philosophy

True success in Thoreau’s view thus cannot be understood in terms of monetary or conventional values, or even in the kinds of epic adventures that show well on Instagram.

A dedicated homebody, he rarely traveled far from home. He refused to dedicate himself full-time to his father’s pencil manufacturing business, though he possessed the mechanical acumen and inventiveness that could have turned him into something of an industrial magnate. Instead, he structured his life to allow for as little work, and as much writing and meditative leisure as possible. And even when it came to that writing, while he did care about his works being read and praised (at least by those he respected), he was unwilling to alter them in order to court a broader audience. Indeed, Thoreau’s friend and mentor, Ralph Waldo Emerson, thought that if his protégé had one flaw, it was a lack of ambition.

Yet in some ways this criticism misses the mark. For while Thoreau wasn’t ambitious for the traditional status markers held up by society, he was ambitious for something else: life. Life at its very essence. Life in its fullest form.

Social theorist Gregg Easterbrook astutely calls this process of getting what we want, but never feeling like we have enough, “abundance denial.”

Compounding this cycle of dissatisfaction — and the desperation it produces — is the fact that attaining external desires often costs money. Money that can only be procured in trade for one’s time and labor. And this frequently isn’t the only payment required: the work one must perform frequently demands compromises to one’s individual values, principles, and dreams. It demands a loss of independence; even the entrepreneur must defer to the whims of the marketplace.

Thus, the more you want, the more you have to work to pay for it, the less autonomous you become, and the further removed you get from the beating heart of life.

Thoreau thus rightly argued that “the cost of a thing” was not simply a matter of its price tag, but “the amount of what I will call life, which is required to be exchanged for it, immediately or in the long run.”

2025-02-24

Working on a thing!

2025-02-25

Spent most of the day diving into React and Tailwind.

Agency vs Intelligence

Words of wisdom from Karpathy sensei.

Cursor Rules

I talked to the AI today and told it stuff it doesn't know yet 🤷🏽‍♂️

2025-02-26

Resend

Send emails with Node.js · Resend #email #api

Resend is the email API for developers.

This may be the simplest and most intuitive API for email sending I have come across so far. I was able to setup my custom domain with it and start sending emails in less than 15 mins. They also have Go and Node SDKs, among others.

2025-02-27

Podscript UI overhaul

UI overhaul · deepakjois/podscript@ef94c87 · GitHub #podscript #react #shadcn

Overhauled the Podscript web UI with close to 1000 lines of AI assisted React code that uses shadcn components.

I used Cursor heavily because my React foo is kinda rusty. However, I did get it to explain every line of code to me as a learning exercise. Along the way, I was able to prompt the AI to improve some of the code it generated, so I am gonna say it was a very collaborative endeavor.

Roti

Community-Based Research Explores Roti’s Global Histories – SAPIENS #roti #food #history #culture

Kale roti, for example, is a regional delicacy from Bangladesh that contains black gram beans (mashkalai) and other flours, and is eaten with mashed dishes made of chili, eggplant, tomato, or spiced beef. Using ingredients such as melted butter and cake flour changes the flatbread’s texture into the soft and spongy South African butter roti. In South Asia, Kenya, and Uganda, the flatbread goes by “chapati,” from the Urdu-Hindi root word “chapat” (slap), referring to the slapping technique used to flatten dough balls into thin, round discs before cooked on a tawa, or hot griddle. In Guyana, the name “clap roti” similarly points to a clapping technique for fashioning a flaky, tender roti—perfect for picking up steaming hot goat curry. In the Indian province of Gujarat, there is an extra-thin roti called a rotli. In Malaysia, the word “roti” can refer to many types of leavened and unleavened breads, including the famous roti canai, enjoyed as a circular, crunchy, flaky bread. And in various places, people have created versions to accommodate dietary restrictions and preferences, including vegan roti.

Slop

The New Aesthetics of Slop - by Ted Gioia #ai #slop

We have come a long way from the days of Impressionism and Naturalism and all the rest. Those were serious movements. They happened because of dedicated artists committed to their craft.

Slop is the opposite.

It’s the perfect aesthetic theory for 12 year olds with no artistic sensitivty—but possessing a crude sense of humor and lots of pop culture detritus in their heads.

Tech companies embrace this—and even brag about the sloppiness of their Slop. Each generation of AI aspires to new levels of whackness.

AI does not possess a self. It lacks personhood. It has no experience of subjectivity. So any art it creates will inevitably feel empty and hollow.

The Slop Manifesto is pretty cool!

How to code with Claude Sonnet 3.7

Some tips to code with Claude Code from somebody who works at Anthropic. #code #claude #ai #tools

Catherine Olsson: Claude Code is very useful, but it can still get confused.

A few quick tips from my experience coding with it at Anthropic 👉

1. Work from a clean commit so it's easy to reset all the changes. Often I want to back up and explain it from scratch a different way.

2. Sometimes I work on two devboxes at the same time: one for me, one for Claude Code. We’re both trying ideas in parallel. E.g. Claude proposes a brilliant idea but stumbles on the implementation. Then I take the idea over to my devbox to write it myself.

3. My most common confusion with Claude is when tests and code don't match, which one to change? Ideal to state clearly whether I'm writing novel tests for existing code I'm reasonably sure has the intended behavior, or writing novel code against tests that define the behavior.

4. If we're working on something tricky and it keeps making the same mistakes, I keep track of what they were in a little notes file. Then when I clear the context or re-prompt, I can easily remind it not to make those mistakes.

5. I can accidentally "climb up where I can't get down". E.g. I was working on code in Rust, which I do not know. The first few PRs went great! Then Claude was getting too confused. Oh no. We're stuck. IME this is fine, just get ready to slowww dowwwn to get properly oriented.

6. When reviewing Claude-assisted PRs, look out for weirder misunderstandings than the human driver would make! We're all a little junior with this technology. There's more places where goofy misunderstandings and odd choices can leak in.

Beliefs about Groups

Property Rights Part II: Groups Are All In Your Head #groups #psychology

Consider the case of the Hutu and Tutsi. Before colonial powers arrived in Rwanda and Burundi, the distinction between Hutu and Tutsi was tied to roles: Tutsi were cattle herders, and Hutu were primarily farmers. These roles were malleable. A Hutu who gained wealth, for example, could become a Tutsi, and a Tutsi who fell into poverty might be seen as a Hutu. It wasn’t penguin-like. No genetic barrier separated the two; rather, it was a set of beliefs tied to social and economic status that people shared about who belonged where.

Then colonial powers showed up, bringing their own beliefs about human groups. The Belgians, looking for a tidy way to rule, assigned group membership based on external traits—height, nose shape, and skin tone—and issued identity cards that fixed people as either Hutu or Tutsi. What had been fluid became rigid, and what had been a matter of local belief became a matter of colonial administration. Suddenly, group membership wasn’t just in people’s heads; it was on ID cards. But here’s the key: the colonial project worked the way it did because it created new beliefs. Hutu and Tutsi became fixed categories, not because DNA had changed but because the social world had.

This artificial rigidity had horrible consequences, which echo to this day. Once people’s beliefs about group membership hardened, so too did the lines of power and conflict. By the mid-20th century, these categories fueled violent struggles that culminated in the Rwandan Genocide. Even in the aftermath of such horror, the reality remained: the distinction between Hutu and Tutsi is still a matter of belief. Just as in the case of property rights—and moral rules as well—if no one believed the distinction mattered, it would vanish tomorrow. There is no penguin-like essential quality separating the two groups. Hutus and Tutsi are Hutus and Tutsis because someone interprets them as being such.

The Bull Case for Generative AI

Key Takeaways OpenAI Is Not A Real Company - Better Offline (podcast) | Listen Notes #openai

Ed Zitron makes a great case for how generative AI maybe more hype than we imagine it to be.

✅ Generative AI is not a profitable industry – it is entirely propped up by venture capital and cloud subsidies.
✅ OpenAI loses billions every year – even on its paid customers.
✅ The user numbers are misleading – low conversion rates suggest weak market demand.
✅ OpenAI’s product strategy is failing – new offerings are expensive, unreliable, and unprofitable.
✅ Future prospects look bleak – if venture capital dries up, OpenAI and the broader generative AI space could collapse.

Full transcript and summary here: OpenAI is Not a Real Company - Transcript Summary

Reality

2025-02-28

Karpathy on using LLMs

The only useful thing I did today was watching this Karpathy video in full, apart from watching the latest episode of Severance.

2025-03-01

What to Pack in a Bug-Out Bag | The Art of Manliness #survival #prepper #outdoors