During this episode of Search Engine, PJ and Casey kept coming back to the idea that Google thinks browsing the web is a chore, hence why they want to let AI do it for you. They, and much of the internet, seems to think that google is wrong here and, to their credit, Google is wrong about most things these days. But, in 2024, browsing the web certainly is a chore. It sucks!

Unless you spend the time to carefully curate an RSS feed or a list of bookmarks to regularly check up on^[1], browsing the web is quite terrible. The infamous personal essays that preclude what could have been simple recipes, the almost useless "Can Pikachu be Shiny in Pokemon GO?" articles, those horrible product spec sheet comparison sites, and immeasurable amounts of other detritus have mucked up the internet such that it takes a great force of will to get any value out of it.

Let me be abundantly clear: the answer to this problem is definitely not AI. Google is definitely wrong about that part. However, at least part of the solution has to be taking the profit out of “What time is the Super Bowl?" posts and putting into good content.

Google incentivized the SEO chum to take over the web and got rich doing it, but publishers played the same game. Much like tech companies during the pandemic, they were either so naive as to not see the end state of their actions or, more likely, simply didn't care about anything beyond next quarter's financials. Together, the likes of Conde Nast, Google, and countless others destroyed everything great that any of them created.

And now, Google is poised to eat its own tail. The media industry is currently in shambles because Google finally realized that people don't like SEO slop, and the companies that got fat on making it no longer have any idea how to make money without their favorite crutch. Journalists are being laid off in droves while Google tells us to put glue on pizza. At this rate, there will soon be no content left to shovel into the gaping maw of the LLMs.

Maybe, as Casey suggests in the episode, the Fediverse is the answer. Maybe email newsletters will save us. But, it took more than Google to ruin the internet.

The Humane AI Pin is bad. The writing was always on the wall for this one. Any product whose core promise builds on the current state of generative AI is going to be bad. And it's going to stay that way until we can trust the robots to not lie to us.

Not to mention, you have to clip the freakin' thing to your shirt.

The problem with voice assistants in general is that there is no way to know for sure if they did what you asked without checking for yourself. Once you go through the effort to double check, you could have simply done the task on your own. While you can generally predict the way that Siri or Google Assistant will fail (usually by simply not completing the task), AI throws a massive wrench into that problem since it will fail in strange new ways and act upon false information.

John Gruber on Daring Fireball:

I don’t know where Humane goes from here but this might be impossible to recover from reputationally. It seems borderline criminal that they shipped it in this state.

Humane massively overpromised and underdelivered here, and this whole thing had a Kickstarter-scam vibe from the very beginning. We learned over the last decade or so that startups with big ambitions are generally not to be expected to deliver. Media in general is far too bullish on AI, but tech media should have known better than to hype this launch to the levels they did.

It's not clear to me that Humane's ideal AI Pin is a genuinely useful thing, and the pin in its current state is clearly a shell of that potential. Last year, when the pin was revealed, we didn't know the extent of the train wreck that this launch would be, but the warning signs were there.

David Pierce, writing for The Verge:

Aboard is just one of a new class of AI companies, the ones that won’t try to build Yet Another Large Language Model but will instead try to build new things to do with those models and new ways to interact with them. The Aboard founders say they ultimately plan to connect to lots of models as those models become, in some cases, more specialized and, in others, more commoditized. In Aboard’s case, they want to use AI not as an answer machine but as something like a software generator. “We still want you to go to the web,” Ford says. “We want to guide you a bit and maybe kickstart you, but we’re software people — and we think the ability to get going really quickly is really, really interesting.” The Aboard founders want AI to do the work about the work, so you can just get to work.

Since it was featured in the Installer newsletter last year, Aboard has quickly become one of my favorite apps.^[1] Now, I'm deeply worried about its future.

In short, the Aboard app consists of multiple boards that can each house cards. Cards can then be further organized using stacks (folders) and tags. It excels when using the browser extension to clip web pages into cards, where it pulls relevant information such as an image, price, brand, or genre.

For me, this has manifested in the app being a truly excellent way to keep track of interesting products. I use only one board, but I have tags for things such as accessories for my mountain bike, a wish list for my ever growing collection of bags, and a list of clothing items I might want to buy. My board is like a big Amazon wish list that can contain anything on the internet. It works perfectly for someone who, like me, accumulates things they want much faster than disposable income.^[2]

Recently, Aboard announced it's march into becoming an AI-based software.^[3] In my experience, this type of announcement usually signals the death of the support for my simple and specialized use case. This AI foray means that you can now ask ChatGPT to create a board for you — for example, Best Picture nominees. You could then sort that board into stacks such as "To Watch" and "Finished" and create tags to rate them. Of course, the AI often makes an incorrect or incomplete board.

This method flies in the face of exactly what I liked so much about Aboard in the first place: curation. My Aboard database is carefully curated collection of objects that I find interesting or compelling, and I'm realizing that I'm quite attached to it. I'm somewhat conflicted about that attachment since it's deeply rooted in consumerism, but that discussion is outside of the scope of this post for now.

If it were only about AI features that are useless to me, then I could live with that as long as my use case remains intact. But this whole thing smells like a larger pivot — they are starting to monetize, and creators Paul Ford and Rich Ziade are explicit about focusing on the workplace market over personal users. The free tier, which used to to be quite generous, now offers just 50 cards per board, with a maximum of three boards. I believe that good software should cost money, but $12 a month is quite steep for AI hype.

This monetization strategy feels bad because it doesn't charge for the expensive part. Data is cheap, and I don't think that users having too many cards and boards is what's going to hurt Aboard's bottom line. The API fees for accessing LLMs are the costly part, but instead of monetizing that alone, Aboard is forcing users to upgrade in order to bypass arbitrary caps on storage. Presumably, this is because they know that the AI isn't very useful yet but they need to start showing cash flow somehow.

Again, good software should cost money, but paying for additional features is a lot easier to swallow than paying to bypass an artificial limit.

Aboard is a cool idea and I hope they find a niche. But I don't think that I'm alone in growing tired of seeing AI features crop up in every app I use. For now, I'm going to test out Raindrop as a promising alternative and, sadly, bid adieu to Aboard.^[4]

But not for the reason you might think.

Matt Birchler: Browse for me gets a big 🤨 from me (birchtree. Me)

Since receiving my iPhone 15 Pro, by Action Button has pretty much been only one thing: open the Google app. I like to keep the dozens of disposable Google searches I do every day separate from the Safari tabs I keep open to annoy myself into getting things done. Having that functionality conveniently tucked away in the action button has been surprisingly useful.

After downloading Arc Search, I immediately swapped my Action Button to open it. Now, to be clear, I don't make regular use of the Browse for Me button. In fact, I've only used it once or twice. The reason I use Arc Search on my action button is because it has cleared the very low bar of being better than the Google app^[1].

It opens the keyboard automatically, keeps old tabs easily accessible, looks nice, and breaks webpages far less often than Google's app. I've ended up with a common workflow of searching in Google and then having to open the link in Safari because the Google app can't display the page properly, or some functionality breaks. Arc Search, so far, has necessitated this workflow far less.

In the beginning, I experimented with using the Action Button to open Safari or use a shortcut to automatically search Google, but those solutions were always slow and clunky, and I always ended up with a bunch of useless Safari tabs open. It's unfortunate since Safari is by far the best way to browse the web on iOS. Arc Search has solidifed itself as second best in my book.

The Browser Company continues to inject some intrigue into a market that's been stale for more than 15 years, and I have to applaud them for that. However, I'm not the first to point out that the Browse for Me button is probably an existential crisis for the internet as we know it. Given the current state of AI, it's currently little more than a fun novelty to mess with, but I expect it to advance rapidly.

This leaves Arc Search is a weird spot. For me, it's a more pleasant way to do Google searches. But that isn't going to bring many others to the app, and it certainly won't make The Browser Company any money. Its hard to figure out what the endgame is here. The Browser Company makes a product for browsing the web (Arc), and then made another product that lets an AI do all the browsing for you. I'm interested to see how they reconcile this space.

All that said, just over the horizon is a wild new frontier of browser apps, after Apple opens up iOS to non-WebKit browsers. A new frontier that those of us who do not live in the EU will sadly only get to salivate over.

I tend toward being somewhat of a completionist. Not in the realm of getting all the achievements or reading every footnote, but in the sense of, once I start something, I want to see it to its Natural End. And, until recently, that Natural End did not include the idea of abandonment.

For years, I'd force myself to muddle through books I ended up hating, games I found boring, or podcasts that didn't click, constantly looking for the part where it gets good. And never finding it.

That is, until recently. David Pierce's excellent Installer newsletter at The Verge introduced me to a lovely app called Sequel. It's a relatively simple tracker for video games, books, movies, and TV series. A content consumption organizer, if you will. And I do love me some organization. I subscribed to the premium tier shortly after downloading the app, and I've been using it frequently for the past few months.

The app allows me to collect the things which I wish to consume in categories and lists, and I can assign a status to each one. In the example of video games: Wishlist, Backlog, Playing, Played, and Abandoned (We'll get back to that last one). It also keeps track of release dates where relevant and, perhaps my favorite feature, integrates with howlongtobeat to show an estimated time of how long it might take to finish that particular game.

Let's take a closer look at the Abandoned status. At first, I thought of it as a curiosity, something that's meant to be used infrequently when I get distracted by something else, or life gets in the way. But, slowly, I started taking it as a license to abandon finishing things. Keeping with the video games example, when I started to lose interest in Starfield, I decided rather quickly to move it to the Abandoned category and move on to something I'd enjoy more.

As I started to do this more and more, I found myself being much more interested with the content that I was following through on, since I can now know that I'm really into it. I've started to realize that free time is far too precious to waste it on things I'm not fully committed to. And, the beauty of this organization system is, I can catalog all the things I've abandoned and revisit them at some point in the future. Or perhaps never. It's the cataloging of these items that sets me free to forget about them.

The final piece to fall into place was a line in Citizen Sleeper that resonated with me. During one conversation, Castor says to the player, "It is the nature of life to leave things unfinished." That line coalesced all these thoughts for me and solidified that license to abandon. Mandating that I finish anything I start goes against the nature of life itself.

Consider this your license to abandon that game that didn't as good as you thought it would be, that book that isn't clicking, or that TV series that doesn't "get good" until season four. Stick with the things that are working for you, and abandon the things that aren't.