Right after college, I moved to San Francisco, a city where I knew one person. I had a lonely time at first, and in particular I struggled to stay connected to the friends I no longer shared a campus with. I wasn’t very good at calling people on the phone and my email correspondences were sporadic at best. But what I did have was my Reader Crew, a group of friends who were all devoted to Google Reader.

Some of you have already lit up at the mention of Google Reader—it’s got a devoted following of mourners. Reader was a short-lived aggregator of RSS feeds (RSS stands for “Really Simple Syndication”). Sites can publish RSS feeds which allow you to access that sites content in another program, called a reader, where you can scroll, sort, and search. These readers pull together any feeds you curate, keeping them updated and tracked. RSS feeds tend to be the posts and articles from a site—scroll down to the bottom of this page and you can see ours—but most RSS readers can also handle newsletters, Tumblrs, and even specific Google searches can rendered in RSS.

Google’s Reader was special because it had some very light social aspects: you were able to follow other people, who could share things from their own feeds into your feed, with or without a small bit of commentary. You could comment on or “like” these shares, but that was about it. There was no big public feed of everyone’s stuff, there was no push to discover other users, and there was no way to make content for Reader. It was just curation and light commentary, if you wanted it.

Molly White wrote a great piece recently for her newsletter that describes RSS aggregation as “curating your own newspaper,” and this was my Reader Crew’s experience. My feed felt like a magazine I was editing, with a small group of friends popping in to guest edit every now and then. It was small, pleasant, and slow.

We were pretty bereft when Google killed Reader, as were many other devotees. It’s hard to replace. Reader was similar to social media, where you can also curate what you’re reading, but without the massive public news feeds and the jockeying for attention. Reader’s more intimate size also felt a bit like a group chat, maybe, but less chaotic and ever-present.

Thankfully, we discovered The Old Reader, which aims to recreate the Google-axed experience and does it admirably well. If you miss Reader, it’s worth a look. But if you’re just starting out with RSS, don’t stress too much about which program to use. There are a lot of free and cheap options that others have aggregated—like Molly White’s from above. Really the question comes down to interface: what is pleasant for you to use and look at? But it’s easy to import and export your list of feeds, so you can always bop around if you want.

I really recommend giving RSS a try, especially if you’re tired of endless feeds that feel like constant, multidirectional fire hoses. I love RSS primarily because you can curate who and what you want to hear from. The pacing is self directed too, and never overwhelming. It feels like riding a bike: fast enough to get somewhere, but slow that the ride is enjoyable. And like reading, you control the frame rate, and can stop, slow down, or go back in your feed if you need to. Which is unlike the stationary bike of social media, where some red-pilled millionaire engineer is cranking a dial to make you peddle faster. Plus, you can get to the end of your RSS feeds, unlike a social scroll which is endless by design.

This scale and pacing issue seems to be part of why RSS never caught on with Silicon Valley business types. It’s a tech that was never flashy or engaging enough. David Pierce wrote an interesting deep dive for The Verge called “Who killed Google Reader?” that reveals how executives never got what was so special about Reader, and had it out for the product from the start. They saw it as “a humble feed aggregator built on boring technology” and “in meeting after meeting, they’d ask why Reader wasn’t just a tab in the Gmail app.”

It’s another reason to love RSS: seems like the tech lords hate it.

When you’re a businessman making dollar-sign eyes at things like Twitter and Facebook, you’re certainly going to be less horny about a product that is slower and less addicting. RSS lacks a stickiness that keeps you compelled to go back. It’s much closer to a tool, allowing you to create something unique and private, that is only as useful or enjoyable as you make it.

We need a new word other than “feed” to describe RSS. A “feed” is for molten metals extruded along an assembly line, or for bullets pumped into a machine gun. Maybe we should start calling the aggregation of RSS feeds “fields” or “pastures”: contained spaces where you can plant and harvest as you like, with no one butting in unless they’re invited.

Not too long ago, the internet was a place you visited. The family desktop sat in its designated closet or back office. In schools, there were rooms filled with computers blinking in tandem, waiting for your class to arrive and hop online. You had to purposefully arrive at the internet, and when done, you left it behind until next time. Now the internet pervades our everyday lives. We have eliminated the doorway, the conscientious effort, needed to access the internet. Always on, always watching, the internet is no longer a place to arrive at and explore but instead a panopticon-like enivronment that we are trapped within. Where the early internet once required intention, place, and presence, today it saturates daily life in ways that erode our capacity for rootedness, attention, and freedom; to recover a healthier digital culture, we must reimagine the internet not as an omnipresent miasma of distraction and surveillance but as a place we choose to enter—and leave—on human terms.

In the 1990s and 2000s, the internet was a deeply physical thing, a location. One “arrived” at the internet with purpose and intention. It was an embodied experience. You clicked and typed, and in the early days often coded via html, your own way through the web. The sense of exploring a physical world was embodied in the way that even website interfaces encouraged breaks in attention, you came to the actual bottom of a page of text rather than scrolling through infinitely, and then you had to make the decision to click forwards, backwards, or leave the page altogether, not unlike flipping through a book.

There was no algorithm, no feed. Instead, on popular web hosting sites like GeoCities, there were “neighborhoods,” wherein the platform and the personal websites it hosted were divided into virtual groups like “Hollywood” for pop culture sites or “Area51” for alien and science fiction pages, allowing users to find websites other users had created based on their interests. There was a sense of locality because of this. Even the language of the old web—homepage, computer room, website—connotes a deep sense of location and place. Once done in the neighborhoods, and done online, disconnection was natural, inevitable, and even restful. You purposefully logged out, shut down the computer, and left its designated spot. It did not follow you and was instead waiting until your next visit.

This experience of the internet, of “Web 1.0,” is sadly gone. It went slowly at first, Facebook and MySpace became new booming platforms, Yahoo! acquired GeoCities and did away with the concept of “neighborhoods” before shutting the platform down entirely, and, in perhaps the biggest development of all, the “infinite scroll” began structuring the internet in 2006. Now the internet is no longer a neighborhood to be visited and explored and left behind. Today the internet has expanded to our pockets and hands, watches and glasses, refrigerators and doorbells and speakers.

The internet is now a panopticon, wherein users are both watched and watching with no true exit. We have lost the doorway to the computer room and, with it, the distinction between private and public boundaries. Under surveillance capitalism, algorithms are always tracking, nudging, and shaping the behavior of users to scroll and interact as much as possible. The freedom and exploration of the internet is no more. This has resulted in fractured attention, anxiety, and sadly, a diminished sense of place and belonging even with the connectivity the internet could offer.

Where once the internet had clear thresholds of arrival and departure, today it offers no such rhythms. There is no doorway to pass through, no sense of beginning or end. Instead, we find ourselves caught in a stream without banks, pulled along by an attention economy that frays our ability to focus, to think deeply, or even to be fully present with those around us. The internet dissolves the difference between “here” and “elsewhere,” collapsing place into a single, endless everywhere. And in this condition of perpetual connection, what looks like freedom often masks a subtler dependence—our choices shaped, nudged, and constrained by forces we neither see nor control.

A healthier digital culture will require the reintroduction of boundaries and thresholds, a reclaiming of the doorway that once framed our entry into the online world. We can begin with simple practices: confining devices to designated rooms, choosing intentional moments to log in, keeping sabbaths from screens. Such acts remind us that the internet is a tool to be entered on human terms, not a condition of existence. But recovery will demand more than discipline; it will require the counterweight of embodied community and locality, the kinds of rooted ties that resist being flattened into the everywhere of the web. And finally, it will call upon our cultural imagination to picture the internet not as an omnipresent infrastructure humming endlessly in the background of our lives but as a neighborhood—a place we may visit, explore, and leave, on terms that honor human attention, freedom, and rootedness.

Not too long ago, the computer room stood as a threshold, a doorway into another world that waited patiently for our return. In remembering that doorway, we recover more than nostalgia; we recover a vision of what it means to return the internet to its proper bounds. The work before us is not to abandon digital life altogether but to give it limits, to render it once again something we visit rather than something that consumes us. If we can imagine and practice such boundaries, the internet may yet be a place to explore rather than a tower of surveillance, a neighborhood among neighborhoods rather than the endless everywhere. And in reclaiming that doorway, we may also reclaim our attention, our rootedness, and the freedom to dwell more fully in the places that are ours.

Image via Flickr.

Bots have crawled the web for a long time, but the past couple years has been something different as companies release their AI crawlers to scrape as much as possible. Cloudflare broke it down by type of bot and source.

Not all crawlers are the same. Bots, automated scripts that perform tasks across the Internet, come in many forms: those considered non-threatening or “good” (such as API clients, search indexing bots like Googlebot, or health checkers) and those considered malicious or “bad” (like those used for credential stuffing, spam, or scraping content without permission). In fact, around 30% of global web traffic today, according to Cloudflare Radar data, comes from bots, and even exceeds human Internet traffic in some locations.

A new category, AI crawlers, has emerged in recent years. These bots collect data from across the web to train AI models, improving tools and experiences, but also raising issues around content rights, unauthorized use, and infrastructure overload. We aimed to confirm the growth of both search and AI crawlers, examine specific AI crawlers, and understand broader crawler usage.

Every now and then I glance at traffic sources, and AI bots seem increasingly common. I wonder if or when bot traffic outnumbers human visits.

Tags: bot, Cloudflare, traffic

Leading Internet companies and publishers—including Reddit, Yahoo, Quora, Medium, The Daily Beast, Fastly, and more—think there may finally be a solution to end AI crawlers hammering websites to scrape content without permission or compensation.

Announced Wednesday morning, the "Really Simply Licensing" (RSL) standard evolves robots.txt instructions by adding an automated licensing layer that's designed to block bots that don't fairly compensate creators for content.

Free for any publisher to use starting today, the RSL standard is an open, decentralized protocol that makes clear to AI crawlers and agents the terms for licensing, usage, and compensation of any content used to train AI, a press release noted.

Read full article

Comments

I Replaced Animal Crossing's Dialogue with a Live LLM by Hacking GameCube Memory

Brilliant retro-gaming project by Josh Fonseca, who figured out how to run 2002 Game Cube Animal Crossing in the Dolphin Emulator such that dialog with the characters was instead generated by an LLM.

The key trick was running Python code that scanned the Game Cube memory every 10th of a second looking for instances of dialogue, then updated the memory in-place to inject new dialog.

The source code is in vuciv/animal-crossing-llm-mod on GitHub. I dumped it (via gitingest, ~40,000 tokens) into Claude Opus 4.1 and asked the following:

This interacts with Animal Crossing on the Game Cube. It uses an LLM to replace dialog in the game, but since an LLM takes a few seconds to run how does it spot when it should run a prompt and then pause the game while the prompt is running?

Claude pointed me to the watch_dialogue() function which implements the polling loop.

When it catches the dialogue screen opening it writes out this message instead:

loading_text = ".<Pause [0A]>.<Pause [0A]>.<Pause [0A]><Press A><Clear Text>"

Those <Pause [0A]> tokens cause the came to pause for a few moments before giving the user the option to <Press A> to continue. This gives time for the LLM prompt to execute and return new text which can then be written to the correct memory area for display.

Hacker News commenters spotted some fun prompts in the source code, including this prompt to set the scene:

You are a resident of a town run by Tom Nook. You are beginning to realize your mortgage is exploitative and the economy is unfair. Discuss this with the player and other villagers when appropriate.

And this sequence of prompts that slowly raise the agitation of the villagers about their economic situation over time.

The system actually uses two separate prompts - one to generate responses from characters and another which takes those responses and decorates them with Animal Crossing specific control codes to add pauses, character animations and other neat effects.

Via Hacker News

Tags: python, ai, prompt-engineering, generative-ai, llms, anthropic, claude, claude-4

As you know, AI-assisted coding is incredible. Unleashing creativity! So many people who can’t code are suddenly productive geniuses! Ten times more productive! It was so productive for each of them personally, honest! No, they don’t have any numbers on that. AI will cause a flourishing of new software! Apparently.

Mike Judge is a software developer. He’s coded for decades. In particular, he’s done a ton of small mobile apps. He also started using AI coding early. He was a fan of it!

Then Mike saw the METR study that showed developers thought AI code completion made them 20% faster, but actually made them 19% slower. He liked that METR put numbers on the question.

So Mike asked: if AI lets any bozo make an app now … where is it all? Where’s the shovelware? Where are the widgets? Where are the crappy little apps that one guy makes at home?

If AI coding is such rocket fuel for development, we should be seeing a flood of little apps being released. And we aren’t. [Substack]

So, numbers time! First, Mike measured his own productivity — he felt like AI made him 25% faster. But he kept track for a few weeks, and it slowed him down 21%.

But what about everyone else? There should be so much software coming out. Mike collected all the numbers he could.

Apple Store app releases have been flat for the past few years. Google Play app releases were flat, then they went down. Steam release rates have gone up steadily since 2023, but there’s no visible surge, from AI or anything else. New GitHub repositories have been flat since 2023.

There’s no surge. There’s no burst from AI or anything else. AI coding is not having any visible effect.

People are getting fired from their jobs for not using AI tools that don’t help. But the whole claim of efficiency from AI coding looks fake.

You’d almost think all the tech layoffs were because the real economy where people live is screwed right now, businesses overhired for the past few years, and no-one’s being adventurous in 2025. AI is not the reason for the layoffs, it’s the excuse.

If AI coding was such rocket fuel, those laid-off developers are the precise people who’d get bored and shove out some quick little apps. But the evidence is that they aren’t.

The standard slop coder excuses don’t work here at all. “It works for me” is great, very happy for you. But those graphs show it’s not working for most people, or we’d see it.

Prompting better isn’t the answer or it’d be working already.

Some claim AI coding is super effective, but not in public spaces. You can’t prove it’s false! Sure, but you’re not proving it’s true, and you’re not explaining the gap.

Saying AI coding supremacy will definitely happen in six months doesn’t say why it’s not happening now. The propaganda sure says it’s happening right now — and it just isn’t.

Some slop coders say it’s just unfair to expect a burst of new apps — but they always say any actual measurement is unfair. They prefer vibes.

This is while the AI code propagandists are trumpeting ridiculous gains — one study has 17% of developers (that they asked) claim they’re literally ten times as productive! And I mean, come on now. [Qodo]

Mike advises:

if someone — whether it’s your CEO, your tech lead, or some Reddit dork — claims they’re now a 10xer because of AI, that’s almost assuredly untrue, demand they show receipts or shut the f— up.

• Video version