• We know that we can live in the current climate, we don’t know that about any other climate.

• We know that society functions with current laws and norms, we don’t know that about any other set of laws.

Often, we are adapted to the status quo and the status quo is adapted to us.

• We and the plants and animals we depend on evolved to live in the current the global climate. Many of our city are built on the coast, our population is distributed in a way that can be supported by agriculture.

• Our laws are the result of millenia of trial and error.

Often, the status quo is a compromise between different groups, individuals or even desires within an indivdual.

The test for status quo bias is reversal. If we oppose a change in one direction, do we also oppose change in the other direction? We’re concerned about global warming, and we were also concerned about global cooling. Any change in the climate has unpredictable consequences.

But what does vertical integration mean exactly and why is it so important?

A common way to think of vertical integration is to have hardware and software that works well together, each adapted to the other.

But if this were all, open source should work just as well. Android OEMs are free to adapt android to their phones, and so android phones should be just as vertically integrated as iPhones.

In fact, the key to vertical integration is that software is closed source and that the only way to use it is to buy certain hardware.

Economically, software and hardware excell at different things:

• Software locks you in. When you are used to certain software, using anything else is painful.

• Hardware is something you’re willing to spend money on.

The moment-to-moment experience of using a device is determined much more by the hardware then by the operating system. How does the keyboard feel when you type? Do the characters appear immidiately? Is syntax highlighting and intellisense immidiate? Is the device getting noisy or hot?

But hardware is a commodity. Even if you are satisfied with your laptop, you will not pay much of a premium to buy a laptop from the same OEM and the same CPU manufacturer again.

This limits the resources that hardware companies can invest in research.

If Intel or AMD make a better chip, they can only charge a premium until the other has caught up. They have to earn the money they invested back very fast.

If Apple makes a better chip, more people buy Apple devices and lock themselves (and even their friends and family) into the Apple ecosystem forever.

The most common devices today do not have this problem. They are light enough not to need a handle. The phone is easy to put in one’s pocket, the laptop is easy to put in a bag and even easier to fold, which is often enough to do something else. The phone can be held in one’s hand, the laptop needs a table, but an empty table looks normal.

But there is still one device that dominates the room like the TV once did, that makes it as hard to do anything else and that is the desktop PC.

There’s only one equivalent to the closet that Rams hoped people would put the TV into, and that is to treat the entire room as a closet, to leave it, to close the door and to go to other rooms to do other things. But what if there are no other rooms?

The FT and Bloomberg rely on “open” paywalls to fund themselves, paywalls that let you read a few free articles per month. These paywalls ensure both resources — because people have to subscribe to read more — and reach — because people can read a few articles before subscribing, so that articles published by these outlets can go viral, everyone can read them and discuss them.

But they’re easy to bypass with certain browsers extensions and web services, which most people in tech use. That only allows for two kinds of tech media:

• Free outlets that have reach, but not the resources to do serious journalism.

• Paywalled outlets like The Information that have the resources to do serious journalism, but no reach, because their paywalls don’t allow any free articles, otherwise everyone in tech would bypass it.

Why not have some articles always free and others always paywalled? Ars Technica does this, but the problem with it becomes apparent when you ask yourself which articles to paywall and which articles to make available for free. If the best articles are paywalled, they can’t go viral, if they’re free, why would anyone pay? The reach/resources dilemma exists for every article, just as it exists for every publication.

YouTube was invented two years before the iPhone, by people who grew up with laptops, and these people run it today, when the first device for most people is a phone, which often remains their most powerful device even when they get a laptop — they might have an iPhone as their main device and a $200 chromebook for school.

Phones are used in portrait mode by default. I’d love to see data on this, but I’d estimate that most people have auto-rotate turned off, because it constantly happens by mistake, for example when you try to use your phone while lying on your side in bed.

Tapping the fullscreen icon on mobile YouTube rotates the video, and this irritates me, when I use fullscreen I just want to focus in the video without seeing anything else.

The most successful mobile games are in portrait mode. For a long time people made mobile games in landscape mode, like PC games, but even when they play a game, when they choose to enter a different world, people don’t want to rotate their phone anymore.

TikTok is successful because people would rather use a different app to watch video then rotate their phone.

Why did it take a chinese company to realize this? In China, almost everyone had a phone as their first device. The generational divide between laptop first and mobile first people doesn’t exist. Also, ByteDance as a company is five years younger then YouTube. The chinese tech industry started later and that is an advantage here, it means that the decision makers on every level are younger.

The name “Shorts” shows that to YouTube the important thing about TikTok is the length. These “Shorts” are in portrait mode because every aspect of the “Shorts” tab is a clone of TikTok, but they’re also pushed on desktop.

YouTube should figure out longform video in portrait mode.

tragedy n. the story of a hero brought down by a fatal flaw

Twitter started with two unique features:

• Thoughts without titles. A blog post has to have a title. This not only a requirement of blogging software, which would be easy to remove, but also of feed readers and link aggregators. A thought doesn’t naturally come with a title.

• Asymmetric relationships. On twitter you can follow someone, and they can follow you back or not. On facebook you have to send a friend request. This makes the platform impotent, in the sense that it can only replicate real world relationships but not create its own. There would be no friendships if one had to send a friendship request before talking to someone.

Twitters users developed conventions, like the RT, the hashtag and the @mention and twitter adapted to these conventions.

Twitter added replies. Today, every social media platform has replies, but twitter implemented them differently and better:

• Replies are just tweets. They can be liked and retweeted like other tweets.

• Replies can be infinitely nested, and this with a flat interface that works on mobile.

• When you reply to a tweet, the reply is only shown to people who follow both you and the person you’re replying to, so that you naturally write for an audience that is sympathetic to both of you. This also means that replying to a tweet doesn’t give additional attention to the tweet, the reply is only seen by people who have seen the tweet, so that there’s no point in trolling for replies.

There was one thing many twitter users did that was almost too simple to be called a convention, and that is to link to images. Twitter adapted to this too, embedding images into the app.

Soon, people started tweeting screenshots of tweets, rather then retweeting them. This had two advantages:

• It was a way to “steal” the tweet, to ensure that any following retweets “belonged” to you rather then original tweeter. (When you see a retweet, you only see the person who retweeted it into your timeline and the person who originally tweeted it, but not the intermediary steps it took for the tweet to reach you.)

• It allowed you to add a comment, usually either some stupid expression of approval, to not make the “steal” motivation too obvious, or some stupid expression of disapproval, of outrage and anger. (There’s an asymmetry here, which is that the approving screenshot comes off as a steal, whereas the disapproving screenshot comes off as smart, as not giving attention to the person you disapprove of.)

Seeing that people were now mainly using images for screenshots of tweets, twitter should have dropped the embed feature. Instead they invented the quote RT.

The quote RT differed from the screenshot tweet in two ways:

• People could click on the quoted tweet and follow the author, so that trolling people into quote RTing became a way of getting followers.

• Quote RTing is so much easier then screenshotting, and almost as easy as repying. And while a reply is shown only to people who follow both you and the person you reply to, the quote RT is shown to all of your followers, giving you a homefield advantage. Thus quote RTs quickly became the main mode of engagement for people with enough followers.

Together, these two things created a kind of bootlegger-and-baptist relationship between between outrageous tweets and outraged quote retweets.

“The solution to bad speech is more speech.” This is often interpreted as a call to engage with bad speech, or to set up a dichotomy between engaging with or censoring bad speech.

I see a lot of bad speech scrolling through twitter, speech that would not be worth engaging with in any context and speech that is not worth engaging in a context where engagement is a metric the platform optimizes for.

What is my solution to this? To ban this speech? To ban twitter, as a platform that is optimized for anger and hatred? My solution is to close the tab and to read a book.

The most basic decision is not what to say, but what to listen to, since what we say is in some sense a response to that, and we will always respond in kind.

If we read a lot of angry tweets, we’ll write a lot of angry tweets. If we read a lot of books, we’ll write a lot books? It’s not that easy. Of course writing an angry tweet is easier. I don’t plan to write books anyway, but I’d like to write longer essays, I’d like to talk about ideas rather then news.

If you wish to make an apple pie from scratch, you must first invent the universe.

— Carl Sagan

The existence proof argument for general intelligence, the idea that artificial general intelligence must be possible because human general intelligence exists, would not even be conclusive if by artificial general intelligence we meant a brain grown in a vat — the existence of a thing does not prove that it could have come into existence in any different way — but it’s absolutely absurd if by artificial general intelligence we mean a computer running certain software. The existence of anything does not prove that a computer running certain software could perfectly simulate that thing. This is not even true for many things far simpler then a human brain, such as a double pendulum.

If we took the existence proof argument seriously anyway, we would have no reason to be more concerned about artificial intelligence then we are about human intelligence, we would have no reason to be more concerned about an artificial intelligence explosion then we are about a human intelligence explosion, about a human reaching, with the perfect notes app and the perfect nootropics mix, a level of intelligence that then allows him to become even more intelligent ever faster, turning the entire universe into brain mass in the process.

Concerns about general artificial intelligence are based on the idea that it would be fundamentally different from human intelligence. But the existence of one thing does not prove the possibility of a fundamentally different thing.

Producers today do not publish DAW source files.

Why? It’s no longer necessary. Classical composers had no other choice, sheet music was the only way to reproduce music. Today, sound can be reproduced directly.

Today it is common for people to convince themselves that they are blocking ads to protect privacy or because money is the root of all evil — if you earn through ads, and not by selling luxury devices.

Of course, ads are way better for privacy then paywalls and even donations — you don't have to give someone your credit card number if they monetize their website with ads. Paywalls as a business model are bad for privacy even before you're asked to pay, because the first step to getting you to pay is getting you to make an account. Ads are the reason so many services exist which do not throw a "sign in" popup in your face. For an ad-funded service, that would risk you closing the tab, which would directly lose them money. For a paywalled service, if you don't want to login, you're a lost case anyway.

Of course, money is the root of all evil — regardless of how you earn it. If you want to monetize a website, you have to get into people's way somehow. They want to do whatever they want to do, and you show them an ad, ask for a donation, or hide whatever feature or content they want behind a paywall (or a loginwall as the first step to get them to pay). All these things can be done in an honest way, and yet the temptation for dark patterns is always there. But what if you're not running a website at all? What if you're selling luxury devices? Is that not a business model that is morally pure?

Well, compare iMessage to WhatsApp. Everyone with every device can use WhatsApp as an equal. iMessage is only accessible on Apple devices.

In the anti-ad worldview every morally questionable action by Google and Facebook is blamed on their ad-based business model. But have these companies done anything as cartoonishly evil as building a social network that excludes people who can't afford an expensive device?

I'm sure the founders of Spotify pirated music themselves, but they did not rationalize and moralize their behaviour, so they were able to think of a better alternative.

But in the weeks after Fukushima, there were three state elections, in which both these parties lost votes. The FDP fell below the 5% barrier and out of parliament in two states and their chairman announced that this had been a vote about the future of nuclear power.

Maybe if germany's election calendar had been different, if these elections had not happened in the shadow of Fukushima, if that memory had faded away as people started taking climate change more seriously, germany would still be using nuclear power.

Instead of thinking of dark mode and light mode, we should think of dark mode on or off.

If dark mode is on, every background is dark. If dark mode is off, backgrounds can be light or dark.

Colors mean things, and unless we're in dark mode we shouldn't restrict our palettes of colors and meanings.

Light means professionalism. Dark means fun. That's why an office app uses light backgounds while a game uses dark backgrounds.

Dark means privacy and anonymity. That's why an incognito window uses dark backgrounds.

Dark means hacking. That's why a terminal uses a dark background.

Light emphasizes. That's why the taskbar should have a dark background to deemphasize it, to allow to focus on the app you're currently using, rather then the operating system, and whatever other apps you have open or pinned to your taskbar.

You might think that giving the user the choice whether to trust the web app with his data would be great for privacy, but Apple disagrees:

Web apps are coming to Mac. With macOS Sonoma, you can add a website — any website — to your Dock. Just go to File > Add to Dock, adjust the name and icon if desired, and the web app icon appears in your Dock.

When a user adds a website to their Dock, Safari will copy the website's cookies to the web app. That way, if someone is logged into their account in Safari, they will remain logged in within the web app. This will only work if the authentication state is stored within cookies. Safari does not copy over any other kind of local storage. After a user adds a web app to the Dock, no other website data is shared, which is great for privacy.

There are, broadly speaking, two ways for a web app to store data on the client:

• Cookies, which get sent with every HTTP request.

• IndexedDB and localStorage, which do not.

These map to two kinds of web apps.

• Client-server web app, where everything is stored on the server and only a session cookie is stored on the client.

• Offline-first web app, where everything is stored on the client, and the server is only necessary for sync.

In the name of privacy, Apple copies over cookies but not IndexedDB and localStorage, supporting client-server web apps but not offline first web apps.

Because any directory can contain subdirectories, the file system can be arbitrarily deep. This organization of nested directories and files is called a “hierarchical” file system. Again, though the advantages are obvious in hindsight, hierarchical file systems were not widely available before Multics and then Unix. For example, some file systems limited the depth of nesting; CTSS limited it to two levels.

— Brian Kernighan, UNIX: A History and a Memoir

If the idea of a filesystem limiting the depths of nesting sounds insane, imagine a filesystem where every level of nesting is conceptually different,  where there's one program that you use to manage some levels of nesting and another program for other levels of nesting, where you can't move things from one level to another.

That is the state of window management today:

• Different levels of nesting: workspaces, windows, tabs.

• One program (the window manager) that you use to manage the first two levels (workspaces and windows), another program (the browser) that you use to manage the next level (tabs).

• You can't turn a tab into a window, or a window into a workspace.

Data might have connotations of numbers or tables, but the economic methods that the economics imperialist applies to subjects that belong to others are also data. In a functional language function is also a data type.

Zawinski's law states that "every program attempts to expand until it can read mail. Those programs which cannot so expand are replaced by ones which can." Conquer or be conquered.

A unix command line tool adds another flag. A search engine company develops a browser, then an OS, then another OS. (A web search engine develops an image search engine, maps, news, books, flights, finance.)

Data can be used in limitless ways, and that leads to the limitless expansion of everything from the unix command line tool to the search engine.

Unix command line tools are supposed to "do one thing and do it well." There's no similarly clear and confused demand of companies, but there are ideas such as "fair competition".

When the `ls` command adds a sorting option, is that unfair competition with the `sort` command? When google shows a google flights onebox is that unfair competition with skyscanner?

That is/was tab-to-search. I'm not sure I should speak in the past tense, since it still works for some websites, and you can still manually activate it for every website, but it doesn't work by default for every website that supports opensearch anymore.

Does this matter? Can't you just use google for everything? For example instead of searching HN with tab-to-search can't you just add "hackernews" to your google query? That might work for HN (though it lacks certain options to sort and filter results, is not live and not as convenient) but it doesn't work for your notes, emails, messages …

What kind of UI could make it easier to use other search engines?

• tab-to-search: just as you can press enter to go to the suggested website in the addressbar, you can press tab to search it. (this feature already existed in chrome, but now has to be manually actived for every single website.)

• search with different search engines at once. this could mean opening different tabs with the results from different search engines, opening one tab with different columns for different search engines or aggregating the results.

How could a browser that implemented such features be funded, which companies would have an incentive to support it?

• big companies with search functions that wouldn't work as the default search engine, for example amazon, twitter and reddit.

• companies that provide search engines to other websites, such as algolia and typesense.

Any website that has a search function would have some incentive to promote such a browser.

When people used "RT" and copy-and-paste to retweet people, twitter build the retweet feature.

When people linked to images, twitter started showing images directly in the app, and later even made it possible to upload images.

When people posted screenshots of tweets, twitter should have asked why and how people were doing that—mostly to express outrage, often to sic their followers on the OP, but also just to "steal" tweets, so that they, rather then the OP, get the retweets. And learning that, they should have stopped supporting images. Instead, they build the "quote RT" to make that easier.

If up- and downvotes were shown separately, it would encourage unpopular views and it would help people distinguish between comments that express unpopular views (such comments would always have a few upvotes) and comments that are badly formulated, which people wouldn't even upvote if they agree, because they'd be embarrassed by them.