Embeddings are one of the most useful but unfortunately underdiscussed concepts in the artificial intelligence space relative to the modern generative AI gigahype. Embeddings are a set of hundreds of numbers which uniquely correspond to a given object that define its dimensionality, nowadays in a multiple of 128 such as 384D, 768D, or even 1536D. The larger the embeddings, the more “information” and distinctiveness each can contain, in theory.

These embeddings can be used as-is for traditional regression and classification problems with your favorite statistical modeling library, but what’s really useful about these embeddings is that if you can find the minimum mathematical distance between a given query embedding and another set of embeddings, you can then find which is the most similar: extremely useful for many real-world use cases such as search.

You wanna cut through the hype about AI? Here's the key takeaway: it boils down to a bunch of math nerds figuring out interesting relationships between numbers.

Which, of course, is useless to all of us non-math nerds... except for when you apply this information in the context of Pokémon.

Joking aside, I have a basic understanding of embeddings, but this article, with its basis in Pokémon lore, is the clearest explanation for how embeddings work in practice that I’ve seen.

Warning: there's still a lot of involved math happening here, but stay with it. You might learn a concept or two!

"Inside Out 2" has also showcased how vital the family audience is to the box office. This underserved crowd accounted for more than 70% of those in attendance during the film's domestic debut, according to data from EntTelligence.

While this audience came out in droves for Universal's "The Super Mario Bros. Movie," which generated more than $1.36 billion at the global box office, there was little for them to feast on until the recent releases of Sony's "The Garfield Movie" and Paramount's "IF."

We saw Inside Out 2 as a family the week it came out.

The anxiety attack portrayal in the movie got the tears rolling. I haven't felt so seen as it relates to mental health struggles, and I'm glad I have an example in the media I can show my kids as they get older and start dealing with stuff like this.

I don't understand why everyone keeps dogging on Pixar, saying they haven't released a good movie in years. Elemental, Turning Red, Soul, Onward, and Luca are all incredible movies.

The only turd since the pandemic is Lightyear. The reveal about Zurg's true identity made me literally yell "you've gotta be kidding me" out loud in a crowded theatre.

The article here does make a good point about the family audience being underserved essentially since the pandemic. We love taking the kids to our local Marcus theater, and there have been very few opportunities to do so with new movies.

The Garfield Movie was cute but also quite skippable. Better to find the 90s cartoon and binge that.

IF is not a kids movie; it's a movie geared towards aging parents who have lost touch with their inner child. (🙋‍♂️)

Puss in Boots: The Last Wish was dark as hell. I enjoyed it, but my daughter had nightmares for a week after seeing it.

So yeah, I'm grateful for Inside Out 2, and I'm looking forward to more family friendly movies coming to theaters here yet this summer like Despicable Me 4, Harold and the Purple Crayon, and Transformers One.

The Web we have was not born out of neglect. It has taken intentionality to become what it is. The Web we have today will not continue to be what it is and what we envision it to become if we do not involve ourselves.

Yes, it’s good to take a break when your burnt out and tired. Yes, it’s good to know when to stop or circle back when something isn’t working. Yes, it’s good to humbly trust others. These are all healthy, necessary things to do if we want to see the Web thrive, but do not remain extinguished, stalled, or sidelined.

The Web needs you and me.

In early 2023, an 18-year-old college student decided to make her first-ever shoegaze song. Her friend sent her a “beat,” a grungy shoegaze instrumental crafted by the producer grayskies, and she spent two hours recording herself singing over it into her phone, using her everyday Apple earbuds as a microphone. No guitars were strummed, and no reverb pedals were stepped on. The next day, she titled the song “Your Face” and uploaded a snippet of it on TikTok, posting under the artist name Wisp. The video gained 100k views overnight, so she made another. That one got 600k views. She made another. That one quickly racked up 1 million views. Soon after, “Your Face” was being streamed millions of times on Spotify, and before Wisp even released a second song, she had signed a deal with Interscope Records.

Fast-forward eight months later and “Your Face” has been streamed nearly 30 million times on Spotify, almost twice as much as My Bloody Valentine’s classic Loveless closer “Soon.” The official sound snippet has been used in 126k TikTok videos, almost as many as Mitski’s runaway TikTok goliath “Washing Machine Heart” (174k videos). In the real world, Wisp sold-out her first-ever show in less than a half hour, and then her second just as quickly.

Consider this article a bit of a “shot, chaser” to my previous post.

I’ve been really into shoegaze lately. This article does a fantastic job of highlighting how zoomers used TikTok to give the genre a renaissance.

It's a good reminder that social media isn’t innately awful. It warms my heart to see the children using these incredible technologies to unite under the banner of ethereal and somewhat depressing tunes.

Go check out Duster's album Stratosphere.

The whole point of the web is that we’re not supposed to be dependent on any one company or person or community to make it all work and the only reason why we trusted Google is because the analytics money flowed in our direction. Now that it doesn’t, the whole internet feels unstable. As if all these websites and publishers had set up shop perilously on the edge of an active volcano.

But that instability was always there.

The only social network I post on anymore is LinkedIn. I have close to 2,000 followers there.

Lately, I’ve noticed that the “engagement” on my posts is increasingly sparse. Earlier this year, I was routinely seeing thousands of views per post. These days, I’m only seeing hundreds, and when it comes to sharing links to my newsletter, I’m seeing only dozens.

Meanwhile, here on my rag tag blog, I know my thoughts end up reaching people who matter the most to me.

It’s certainly less than the 2,000 people who follow me on LinkedIn, and substantially less than the tens of thousands of people a week who “engage” with my “content”¹ there… but I don’t care.

By posting here, I’m taking the harder route of building an audience without the flashy shortcuts promised by platforms like LinkedIn and Google.

Whenever I try to take shortcuts and play SEO games, I end up doing things to my website which make it feel less authentic.

And these days, I find myself asking, “what exactly do I need to take a shortcut for?”

Robin also quotes this piece by Jeremy Keith where he discusses our need for human curation:

I want a web that empowers people to connect with other people they trust, without any intermediary gatekeepers.


The evangelists of large language models (who may coincidentally have invested heavily in the technology) like to proclaim that a slop-filled future is inevitable, as though we have no choice, as though we must simply accept enshittification as though it were a force of nature.

But we can always walk away.

It’s tough to walk away from the big tech companies, but I can assure you it is possible.

Facebook used to dominate my daily existence, but besides perhaps Marketplace for selling my junk, I do not miss any of Meta’s properties since I left several years back.

Google was my portal to my email, search, and maps for years. In the past few years, I have switched to primarily using Fastmail, Ecosia, and Apple Maps. Here in 2024, they all work well.²

I do my best to avoid ordering stuff off of Amazon, and I hardly stream anything on Netflix anymore.³

I haven’t made the move over to the Light Phone yet, and I find it hard to believe that I’ll give up my Apple Watch, Apple TV, or iPad/Macs… but I do find myself questioning the prolific presence of Apple in my life more often than I did, say, ten years ago.

As I continue to experiment with LLMs, I’ve noticed that the locally-run, open source models getting closer to the performance you see in closed source models like GPT-4o and Claude Sonnet 3.5 Sonnet. It’s only a matter of time that they’re good enough to do the tasks that I find myself turning to ChatGPT to complete today.

Enshittification isn’t inevitable. Like depression, it’s an indicator that something in your digital life needs to change.

This is a dope 8-bit-styled CSS framework that makes your site look like something straight out of an old school Nintendo game.

Someday, I’ll add some themes to this site, and this would be a fun theme to incorporate.

I saw this article referenced while reading Bill Mill’s recap of relaunching a website, which in and of itself is a delightful read for those of us who nerd out on large-scale system architectures.

I am almost certain I’ve read Dan’s piece on boring code before, but I wanted to share it here because it serves as a great reference for those of us who are sick of making bad tech stack decisions for bad reasons.

In particular, the ending here sums up my experience consulting with many different tech teams:

Polyglot programming is sold with the promise that letting developers choose their own tools with complete freedom will make them more effective at solving problems. This is a naive definition of the problems at best, and motivated reasoning at worst. The weight of day-to-day operational toil this creates crushes you to death.

Mindful choice of technology gives engineering minds real freedom: the freedom to contemplate bigger questions. Technology for its own sake is snake oil.

The teams which move the fastest are the ones who are aligned on a vision for what is being built.

Often, these teams hold a “strong opinions, loosely held” mentality where they decide what tools they’ll use, and they’ll use them until they no longer solve the problem at hand.

Put another way: in a business context, experimenting with your tooling is a huge organizational expense that rarely yields a worthwhile return on investment.

Your focus should be on what you are building rather than how you’re building it.

We’ve seen a lot of AI giants engage in questionably legal and arguably unethical practices in order to get the data they want. In order to prove the value of Perplexity to investors, Srinivas built a tool to scrape Twitter by pretending to be an academic researcher using API access for research. “I would call my [fake academic] projects just like Brin Rank and all these kinds of things,” Srinivas told Lex Fridman on the latter’s podcast. I assume “Brin Rank” is a reference to Google co-founder Sergey Brin; to my ear, Srinivas was bragging about how charming and clever his lie was.

I’m not the one who’s telling you the foundation of Perplexity is lying to dodge established principles that hold up the web. Its CEO is. That’s clarifying about the actual value proposition of “answer engines.” Perplexity cannot generate actual information on its own and relies instead on third parties whose policies it abuses. The “answer engine” was developed by people who feel free to lie whenever it is more convenient, and that preference is necessary for how Perplexity works.

So that’s Perplexity’s real innovation here: shattering the foundations of trust that built the internet. The question is if any of its users or investors care.

Well, I sure do care.

As everyone was celebrating and feeling good, I was barely functional. Truthfully, I had never felt closer to death in my life. I’ve done hard workouts before. I know what it’s like to push myself. I’ve been running for over a decade. But what I experienced after crossing that finish line was something else entirely.

And for what? To have a 07:25 pace instead of a 07:30 pace? Remove my two sprints from the race and I come in maybe 30 seconds later. What difference would it have made in my life? None. I don’t win some extra prize by coming in at 25:57 instead of 26:27. 

So why did I do it? Yes, I wanted to push myself. Yes, I wanted to beat my goal. But, ultimately, I did it because I was selfish.

I love a good running analogy.

I heard Derek Sivers make a similar point with biking a few years back. Pacing is an important aspect to a well-lived life.

I also enjoyed this Josh Brown quote he included in this article:

Make yourself useful to smart, successful people. That’s how you should spend the first ten years of your career.

Surround yourself with smart, successful people and then bet on them. That’s how you should spend the next ten years.

When you embrace the idea that, yes, you were lucky, the fear drops away. And then you become more open to the possibility that the universe will continue to guard your back.

Because here is a truth that only a few discover: when you look for signs that the Universe is ‘friendly’ you will find them everywhere.

It is far better to live in a ‘friendly’ universe than an ‘indifferent’ or ‘hostile’ one.

I’m used to ascribing neutrality as the universe’s default mode, but I didn’t consider the possibility that a neutral universe can be harnessed in whichever way you want.

As a developer, whenever I see my code works right, I often squint at it in disbelief, wondering what I did wrong, feeling like it’ll break the second I push it to production.

Maybe in those moments where my impostor syndrome is peaking, I should accept the pat on the back from the universe and give it some flowers.