Elliot Hopkins' Blog

Cleaning up the mess I made building things I love

Sun, 12 Apr 2026 00:00:00 GMT

I've added many unique features to my website over time, and that's by design. My website is my sandbox. It gives me a place to play with ideas I'm passionate about and put them somewhere real right away. This habit of quickly adding new, interesting things has allowed my site to evolve dramatically over time, despite no single change being very big on its own.

However, while this approach lets my website evolve in a way I'm happy with, it sometimes leaves me with tech debt that's more trouble than it's worth. Specifically, two features in particular had become hard to maintain. It was time for a little spring cleaning.

The first feature was how I displayed blog posts from the Atmosphere (a new social technology that powers things like Bluesky) alongside my regular blog posts on the homepage. When I first built the Atmosphere integration, it was simple and client-side. Post previews loaded directly in the browser via JavaScript, and they linked to the posts on their original pages. Later, I was getting more excited about AT Protocol (which powers the Atmosphere) and how easy it made building things off of it. Because of this, I got a bit ambitious and made posts from Leaflet (a blog platform on the Atmosphere) display as "native" content, styled exactly like the blog posts already living on my site.

It worked! And it worked pretty well. But there was a problem. I would have to update my site to support every type of content block Leaflet supports, with new content blocks being added frequently. Plus, I was now displaying posts from across the larger Atmosphere on my site — not just Leaflet. I didn't want to add this for every one of them, and giving Leaflet special treatment felt a bit odd.

So I went back to a client-side version of this feature, like I had before. It still keeps the part I like most — all my posts showing up in one place — but is simple and uncomplicated.

The second feature I had to cut down was the read-along feature, which had been causing me issues for a while. This feature highlighted text line by line as you listened to my article read in my own voice. The feature was inspired by how Firefox handles its text-to-speech highlighting in reader mode. I thought, "Why couldn't I do that on my website, but with my voice?" I still love the idea, but unlike text-to-speech, there's no great way to track what is being said in each section of an audio file without a perfect captions file, so I had to improvise to create a different solution. I threw together a weird timing system (based on auto-generated approximate caption files that then had to sync up with the actual text) that was "good enough." It was never good enough; it was a headache.

The feature was thousands of lines of code, took multiple complete rewrites, and kept breaking in new ways every few months. I would frequently notice a small bug while working on something completely different, tell myself I would just do a quick fix, and end up rewriting the whole system again. For a feature most people never used, it was taking up way too much of my time.

This spring cleaning wasn't as exciting or thrilling as building something new, but it made me feel like a weight had been lifted off of me. I now have the space to play around with new features on my website.

"Move fast and break things" is a motto that has some truth to it, but breaking things still comes at a cost. If something you made while moving fast is so fragile that it constantly breaks, it's not worth keeping around; it means you either always have something broken or you can no longer move fast trying to avoid breaking it again. Neither was worth it, so I'm glad I no longer have to worry.

My complicated relationship with the em dash

Tue, 09 Dec 2025 00:00:00 GMT

I have a complicated relationship with the em dash. It's a telltale sign of AI writing, and yet, it's one of my favorite literary devices.

If you don't know, the em dash (—) is that long dash that you might see in books, newspapers and frequently in ChatGPT and similar AI chatbots. You can use the em dash to create contrast between two statements, mark an interruption or aside, or for pure emphasis. In certain types of writing, it's commonplace, like in written fiction or journalism. But, at least before ChatGPT, it wasn't as common in most people's writing. Most people would just use a normal dash "-" in its place. Really, most people don't know the difference. The standard dash often has the same effect as the em dash, but it's just easier for people to type on a standard keyboard.

Now, I'll admit, I didn't properly know what the em dash was until ChatGPT came out. I had seen it in use, especially in news articles, but I didn't realize it was a fully separate character. I would use just a double dash "--" when I wanted to serve its purpose.

Once I found out what it was, though — after ChatGPT came out and people pointed out how much it loved the use of the symbol — I sort of fell in love. It's sort of elegant and fancy, even with its overuse in things like headlines. And it makes sense. I write a lot in a journalistic style for my student newspaper, including things like headlines. And as someone who likes a good aside in my personal writing, it serves a perfect purpose, especially when taking the role of parentheses for little asides. Using em dashes for an aside keeps the thought more grounded in the text, while parentheses separate the thought as a less important note or detail.

However, I want my writing to feel personal and human, and with em dashes increasingly becoming a sign of AI text, my use of them only serves to compete with that purpose. I want it to be clear that when I write something, it comes from my brain — not an AI. Not to say that I've never or will never use AI to help me put my thoughts into words or rewrite a sentence in a clearer way (I believe that AI is an incredible accessibility tool, including for things like this), but my writing should still overall feel human. It's important to me that when people read things I write, they feel they've made a connection with a human, not a bot.

Still, the em dash has found its way into my personal writing style. Should I avoid it to make my writing better serve its purpose and intention, even if it means letting go of a part of my personality in my writing?

I think, for me and in my writing, the answer might lie somewhere in between. In my last blog post, I used a lot of em dashes. I overused them for sure, which takes away some of their power — similar to how it's becoming overused in some news headlines, turning it into a cliché. I think I'm going to keep using the em dash, but just attempt to use it more rarely and tastefully. And using something tastefully isn't something ChatGPT can understand, at least for now.

How to know when to trust the news — from a student journalist

Tue, 02 Dec 2025 01:04:15 GMT

This is an external post Elliot wrote via the Atmosphere on a publication called from the desk of elliot. View it on the web here.

The post content should be in the RSS feed, but it might not look perfect.

Do you really know what a journalist does?

Before I joined my school's journalism class, I thought I knew what a journalist did. But I did not. Actually, I'm willing to say that most or at least a good number of young people don't properly understand the job of a journalist or the policies that they are expected to follow.

But after joining the journalism elective at my school and, in turn, becoming a student journalist for my school's newspaper, I learned the general standards, processes, and ethics professional journalists follow.

Not all people who call themselves "journalists" follow all of the standards, and journalistic policies and ethics are always a topic of discussion. But most good-faith journalists, especially from major newsrooms, agree on certain standards and policies to follow.

In today's news landscape, it can be hard to tell what to trust, and trust in journalism seems to be falling. I think that, in part, we have failed at communicating what the job of a journalist is, how they should act, and how you can know to trust them. And the trust shouldn't be blind; you should be skeptical of what you read — and trust should be earned. Critiquing how things are covered and asking the right questions is also important for keeping journalism trustworthy and transparent.

I want to do my part in educating people on when you can trust where you get your information and the role and ethical principles of journalism.

The Basics of Journalistic Ethics

SPJ’s Code of Ethics

Members of the Society of Professional Journalists believe that public enlightenment is the forerunner of justice and the foundation of democracy. Ethical journalism strives to ensure the free exchange of information that is accurate, fair and thorough. An ethical journalist acts with integrity.

www.spj.org

So let's start with the general ethics professional journalists should follow. Most journalists follow the Code of Ethics by the Society of Professional Journalists (SPJ). The SPJ is probably the most well-known organization where all sorts of different journalists come together to promote free speech and encourage ethical journalism. Sometimes journalists follow other policies besides the SPJ Code of Ethics, but these tend to mostly agree with or overlap the SPJ Code of Ethics.

Really, the main thing the everyday person should take away is that journalists should Seek Truth and Report It, Minimize Harm, Act Independently and Be Accountable and Transparent (these are the "pillars" of the SPJ Code of Ethics). If someone who calls themself a journalist doesn't seem to do this, you should question that.

Finding out if a source is trustworthy

If you're not super plugged into the world of journalism, it can be hard to determine which publications you should trust. To make things easier, I've created a list of some "green flags," "yellow flags," and "red flags" you can look out for when evaluating a source. This is not the end-all-be-all by any means, but they serve as a good reminder of what to keep in mind.

A "green flag" is a good thing, and it means a source is more likely to be reliable/trustworthy.

A "red flag" means a source is probably not reliable/trustworthy.

A "yellow flag" means to stop, pause, and look deeper into it.

✅ Green Flags:

Easy-to-find ethics policy
A good about or information page
Articles have clear bylines (author name(s) listed)
Authors have about pages
Quotes are clearly attributed to sources
Publication uses neutral language in news reporting
When people are fully or partially anonymous in an article, a reasonable descriptor is used and rationale for anonymity is given
Factual information has clear sources, often linked
Publication is a non-profit or similar
Ads and promotional content is clearly labeled and separate from the publication's reporting
Articles are edited by people other than the author before publishing

⚠️ Yellow Flags:

Can't find ethics policy
Articles don't have proper bylines or attribution
Many sources are unnamed or anonymous
Publication is not well-known
Publication is owned by a billionaire or similar
Website is littered with ads
Articles are NOT edited by people other than the author before publishing

⛔ Red Flags:

Opinion articles are not properly marked or separated from standard news content
Author inserts their own opinion into a factual news story (this is known as editorializing)
Publication uses words to dramatize an event or create fear, particularly in headlines
Headlines are "clickbaity"
Facts and claims are not easily verifiable
Articles are sponsored or have integrated advertisements
Conflicts of interest are not labeled properly

Hopefully you now have a better understanding of journalism and are better equipped to find out whether or not you can trust a publication.

If you have any questions, concerns, found anything I got wrong or you think I can improve or add, let me know!

glasswing butterflies

Fri, 03 Oct 2025 01:01:03 GMT

This is an external post Elliot wrote via the Atmosphere on a publication called elliot's random rabbitholes. View it on the web here.

The post content should be in the RSS feed, but it might not look perfect.

it's true. i've fallen down another rabbithole. a few nights ago, i discovered the glasswing butterfly, or the greta oto.

like i said in my post, i don't remember how i came across this, but that means that now i get to retrace my steps.

it seems like this all got started when i was testing out a few ai models with kagi assistant, especially after the new minor update to deepseek.

now, i was trying to think of some sort of interesting research question to ask the AI, and I must've somehow thought of butterflies. probably because i have this image of a teal butterfly on my website, that I think looks quite cool.

i wanted to put an image here, but i'm having issuessss lol, so if you want to see it just go to my website:

Elliot Hopkins

elliothopkins.com

so i asked my ai the following:

what is the prettiest blue butterfly species in the world? provide links to pictures or articles about them. i specifically am interested if there are any teal blue / green butterflies

i think i was specifically wondering if a teal butterfly, like the one on my website, was actually possible... spoiler alert, it probably isn't, at least in most lighting conditions :(

... but I did get some good names about other cool blue butterflies (like the blue morpho)

anyways, i think i must've started asking follow ups about the coolest butterflies in general, and well, that's where i heard about the glasswing butterfly

and well... glasswing butterflies look really cool, just see for yourself:

they are stunning.

i continued to look at the wikipedia and pages with info and facts about glasswing butterflies, and especially, i looked through a lot of random pictures of glasswings

i was very curious how they are able to have such beautiful transparent wings, and if you are to, i want to recommend this amazing video to you from this channel which i hadn't heard of before called "deep look" from PBS and KQED

here's the video:

ok i hope you enjoyed hearing about glasswing butterflies, stay tuned for more random rabbitholes in the future? maybe??

Little touches that make my website truly mine

Sun, 24 Aug 2025 00:00:00 GMT

I've always loved when people had super clean personal websites, so when I migrated my blog from Ghost to 11ty, mine started out the same way — minimal & elegant, without a whole lot else design‑wise. But as much as I love simplicity, I also wanted it to feel more me, especially as I started thinking of it as more than just a blog. That's why over time, I've added small touches here and there — colors, fonts, little details — that reflect me & my personality. And, to be honest, I might not ever stop.

The first personal touch on my site has been there from the start: the beautiful teal accent color. It used to be a darker shade of teal, but now it's a tad lighter for my personal preference. Right now, teal is my favorite color (though that tends to change every few years, lol), so it makes a lot of sense to me to have it as the accent color for my webpage.

I also love using fonts to show my personality in small ways. The main body font is one of my favorites: Inter. It's a simple sans serif, but I tweaked a few settings. Notice the capital "I" has proper stems and doesn't look like a lowercase L. And the lowercase "l" has a little curve at the end. If you're using a screen reader, reader mode, or hearing me speak this out loud on my website, then this sentence won't make sense — but just know I'm not crazy.

But the other font on my site — the handwritten one — is the most interesting. That's because it's actually my own handwriting (I tried to write neat, though). I did this because someone's handwriting is very personal to them, so it felt like a perfect touch to add to my webpage.

A newer personal touch on my site is audio. You can actually hear me read my own blog posts out loud. It's a personal website, and what's more personal than someone's voice? Of course, I had to take it a step further and sync the on‑page text with my voice. It was a pain in the ass to implement, and after a solid day, I finally found a solution that was "good enough," but I'm glad I put it in there. It's a fun feature that you don't see on everyone's personal website.

And even as I'm writing this post, I've added some more elements to my website to display my personality. I spent a solid couple of hours getting a really nice teal-blue textured gradient footer on the site, inspired by a bunch of fun gradients I keep seeing on Pinterest.

I also added a little blue butterfly icon, inspired by things I kept seeing on Pinterest as well.

I realized that I saved tons of fun design ideas on Pinterest (that's mostly what I use it for), but I rarely use them as inspiration for anything. They just sit there, even though they feel very "me." That's what led me to add the butterfly icon and the gradient footer to my site.

My website will likely never be "finished," at least not in a traditional sense. But honestly, that's the beauty of it. It's constantly evolving; it creates a place for me to mess around and grow over time, with it growing right there alongside me. And even though I update it often, each change is less about perfection, and more about making the site a truer reflection of me.

So long, Arc.

Wed, 19 Feb 2025 00:00:00 GMT

I've been a loyal Arc Browser user and fan for a while, but The Browser Company's recent announcement that they would step back from developing Arc and that they would now be focusing on developing a new web browser called "Dia" made me feel a bit betrayed. This came right after they had just made a bunch of big promises about a major update to Arc, codenamed "Arc 2.0." But now, they said, "Actually, we think Arc is a completed product," despite the fact that Arc still had so much to continue to work on, especially with their Windows version.

While I could have just kept using Arc until it eventually died off, that's just not my style. It was time to find a new browser. Arc was hard to leave, though — I had grown fond of its vertical tabs and its many other niceties. And, to put it simply, I now knew that something better was possible for a web browser, and I couldn't just go back. My old web browsers just didn't cut it anymore, so I began the search to find a new one.

After checking out the usual suspects, like Firefox and Brave, I gave the Orion browser another try. I had used Orion before, but only to test it out. Orion is like Safari, but better, and it's made by the team behind my favorite search engine, Kagi. It seemed like a good option, but while it was a good browser, honestly, a very good browser, it still felt like a web browser of old; it was very much a "pre-Arc" web browser. Sure, it had vertical tabs, profiles, and other features that made it better than other "pre-Arc" browsers, including a tight integration with Kagi, but it wasn't an Arc replacement by any means.

Somehow, I eventually stumbled upon the Zen Browser. At the time (which, again, was only a few months ago), Zen was still in alpha, and it definitely felt like it. Despite that, it was looking promising. It had a lot of the features I was looking for in a web browser, and it was built on Firefox, which is a plus in a browser landscape dominated by Chromium-based browsers. Even though it was at a state that I was willing to put up with, it was definitely not something that I would recommend to others, at least not yet.

In these past few months, though, the team behind the Zen Browser has been working overtime to make the browser feel even better to use with each update, and they've been adding new features, squashing bugs, and making the general user experience even better. The improvements they've made have made Zen look less like a Firefox fork and more like a refined version of the Arc Browser. All that is to say that Zen has evolved dramatically. Zen has exited its alpha stage and is now in beta. The browser now feels native to MacOS (and to other operating systems, I've heard) and is just overall very nice to use. It's kept all the things that I loved about the Arc Browser and refined them. Switching from Arc to Zen now truly feels like an upgrade.

The only thing I'm still missing from the Arc Browser is tab folders, but the devs are working on it. Last I checked, they were waiting for Firefox to release their version of tab groups so that they could piggyback off that implementation, rather than having to build their own implementation completely from scratch.

Another thing I like about Zen is that it's easier to move to Zen Browser from a more traditional browser like Chrome or Firefox than Arc ever was. Zen brings new ideas to the browser, but keeps enough of the features of a traditional browser, like the bookmarks bar, that makes it easier to dip your toe in than Arc, which made you abandon basically everything you knew about from your old browser.

The Zen Browser is a beautiful browser and an even better Arc replacement. It's open source, privacy-respecting, has a beautiful UI, and is available for all desktop operating systems: Windows, macOS, and Linux. It's improved so much in the past few months from where it was in its alpha stage that I can say it's a browser that I can now confidently recommend. I'm even willing to say that it's currently the best desktop web browser.

The Zen team is really doing great work, and they deserve so much support. If you're interested, you can download Zen on their website and also donate to the devs there if you'd like.

Why I Left Ghost - and the process of moving my blog

Wed, 09 Oct 2024 00:00:00 GMT

Ghost is a beautiful piece of software for blogs, newsletters, and publications alike. There's a reason why modern publications like 404 Media and Platformer choose to host their sites on Ghost, rather than something like WordPress. Actually, Ghost is a lot like a modern reimagining of what WordPress could be if it never stopped focusing on the core idea of hosting publications, rather than expanding to more generic websites as well.

That's why when originally choosing where to host my blog, I chose to do so using Ghost. It had all the features I needed and was the right amount of simple and easy-to-use while still being extremely functional. It was perfect for my use case at the time. The only downside for me was the cost. While $11/month wasn't a bad price for its features, since I wasn't posting very often, I was mostly just paying to keep it up.

But use cases change, and so does what you want software to do.

As time went on, I started to want to customize my blog a bit more. I couldn't make a custom theme for Ghost because the plan I was on didn't allow that, so unless I wanted to pay an extra $20 per month, I'd have to inject style code into the HTML of the site myself. It's basically a hackier and more limiting way of adding a little bit more customization to my site. I mostly just used it to remove a couple of buttons and play with the font.

I also wanted to start taking my blogging a little less seriously. Up until this point, I've called my blog posts "posts" or "articles" and my blog was marketed more as a weird one-person "publication" of sorts. But I found I didn't really care for this in the long term as it added unnecessary pressure for these posts to be super polished and informative and I just wanted to blog and rant, you know? Like I wouldn't have written "you know?" because my blog was a bit more formal.

As I started thinking about my blog in this way, Ghost just felt like overkill for me. Especially as I started thinking of other ways I could host my blog for free. And over the past couple of years, there have been some great new ways to host simple blogs super easily. There is even a pretty cool simple blogging platform called weblog.lol built into a service I already pay for (omg.lol)! These services were great, and they were almost what I was looking for, but they weren't quite perfect. And you know what they say...

If you want something done right, do it yourself.

What I ended up doing is custom-coding my blog myself. Well... with a little help from some friends. The base of my website, which includes the original code and design, was made by me. But I'm using Eleventy, a static site generator, to make it easy to update the whole website at once or write a new post without coding anything. For example, I've set it up to automatically generate the "Recent Posts" section to always show the most recent blog posts I've written. And I write my posts using basic Markdown, which 11ty automatically converts into a webpage for the post.

I want to be clear that Eleventy is not for everyone, as it requires a bit of technical knowledge to get set up. But if you want to make a simple little website that's easy to update and you already know some HTML and CSS, go for it.

Of course, 11ty doesn't host your website anywhere on the internet; it just spits out some HTML files for you to upload somewhere else. But once you have the files for a static website, it's super easy to find somewhere that'll host it for free since static pages aren't all that resource intensive. Heck, you could even host it on a Raspberry Pi, which I've tested out and works great as long as you're fine keeping your Raspberry Pi and its internet connection up and running (I didn't opt for this because it would just be a bit annoying to manage keeping my site up myself when someone else could host it for me for free and probably keep the website up better than I ever could, but it's absolutely a viable option).

I'm excited to post more on my blog - and potentially tweak the design a bit here and there too. But that's all from me for now. I hope to be posting more blog posts in the future about... well, whatever I want.

Why I'm Paying for a Search Engine

Tue, 23 Apr 2024 00:00:00 GMT

This is something I thought I would never do. Something that, before I did, would've seemed ridiculous to me. I started paying for a search engine. Recently, I've begun to use (or try out) a search engine named Kagi, which is $10/month. I know it seems a bit crazy, I mean, paying for a search engine? Why would anyone do that when you can just use Google or DuckDuckGo or Bing or even Yahoo for completely free? Free as in zero, nothing, no money spent (and not even necessarily paying in your data, either). Well, I decided to pay for Kagi anyway. Let me explain...

Before this, I had heard of Kagi from a couple of places, specifically through the Orion web browser for Mac. Orion is a cool little web browser for MacOS that works a lot like Safari but with more quality-of-life features. Since Orion was developed by Kagi, I knew they were a niche paid search engine. I think I may have tried a couple of searches with them, but nothing really caught my attention, especially not enough to pay for a search engine. So I went on with my life as normal until another product caught my eye.

It was called Perplexity, and it was an AI-powered search/answer engine. (The idea behind an "answer engine" is that, as opposed to standard search engines, "answer engines" answer your question right away, and don't just provide search results). I found this idea very interesting, especially since I tried a couple of searches and it seemed to work quite well. I started by using Perplexity for the occasional question I had, then began testing out the "Pro Search" mode for doing research, then finally slowly used Perplexity more and more alongside my search engine, until finally, I thought, why not just make it my search engine? After all, it also gave you a list of "sources" that worked similar enough to a traditional search engine for when you needed to do that sort of search.

As I used Perplexity more, I started to think more about its Pro plan, a $20/month subscription. $20 a month was certainly a lot, but this tool was providing me with a lot of value, and near-infinite "Perplexity Pro" queries seemed enticing. I eventually convinced myself to get it by comparing Perplexity Pro to other AI services. I told myself that Perplexity offered multiple Large Language Models (which you can think of as different versions of ChatGPT from different companies), while other AI products, like OpenAI's ChatGPT Plus, charge the same amount for one LLM in a wrapper that definitely weren't as good as Perplexity's. Now, all of a sudden, I was paying for a search engine, although an untraditional one. I told myself I was paying for an AI search engine, with the emphasis on that it was the AI that I was paying for, but over time it made me start to feel like paying for a search engine wasn't so crazy after all.

All was good for a while, and Perplexity and I were going steady with most of my web queries going through them. I liked just having the information there, plain and simple, while still being able to dig deeper if I wanted to. It felt like it was the best of both worlds (as Hannah Montana would say). But, over time, I found myself using Perplexity's "sources" section more for traditional web results, but then not being satisfied with the links it provided, and ending up having to go to a traditional search engine anyway. For lots of things, I found out I just wanted normal, clean, traditional search results, especially when it came to finding information and news on recent events, or to find a very specific website or webpage (which happened more often than you'd expect). I started exploring how to use search engines alongside Perplexity, like trying to figure out whether Perplexity or my traditional search engine should be the default option when typing in my browser's toolbar. I also had to figure out what sorts of queries were best for Perplexity vs the ones that were best for traditional search. But even still, I just couldn't give up Perplexity. Something just couldn't beat those AI overviews when I just needed a quick answer or a way to point me in the right direction of where to look.

So, thank you 404 Media, because a 404 Media article about Kagi search is what got me to look at it again, this time more willing to pay for a search engine (since I already had been paying for a version of one for a while). I started with a free trial or test account of just 100 searches, and what I found was a really nice search experience. The searches were quick, the results were good (with me finding that there were very few SEO articles in my results), and all the listicles were bundled together. (It's actually so great, why don't more search engines do that?) I could also have Kagi rank certain websites higher or lower in search results based on my preferences. But the real killer feature that had me rethinking where my money for search each month should go was Kagi's "quick answers". Kagi's quick answers are Perplexity-like summaries of results that only show up when you tell them to. You can do this in 3 ways: if your search ends with a question mark, if you press the "quick answer" button after you do a search, or if you press the "q" key on your keyboard after searching.

This was genius! While all these companies working on search and AI have been trying to figure out how to detect user intent of searches and whether or not to show AI summaries or just plain web results, this tiny niche web company Kagi cracked the code. Just let the user, or more accurately the person, tell you what they want. If I want AI results, I either type a question mark or press "q". If I don't, I don't.

Kagi is also cheaper than Perplexity at half the price, and I don't really need all of Perplexity's Large Language Models, so I've decided to give Kagi a shot more long term. I've canceled my Perplexity subscription this month and signed up for Kagi. So far, I'm feeling pretty good about it. But we'll see how it holds up...

How Letter Grades Cultivate a Toxic Environment for Students

Sat, 12 Aug 2023 00:00:00 GMT

Important note from Elliot: This is an opinion piece based on research, anecdotal evidence, personal experience, and a survey with an extremely small sample size that should not be seen as accurate. Still, this article brings up some important ideas and conversation starters around grades, the school system, and grade pressure culture. It is a piece of work I continue to be proud of a year after adapting this for my blog. Still, it should be noted that this is a deeply argumentative and opinionated post and that you should do your own research before coming to any conclusions. (note dated Aug. 29, 2024)

💡This article has a brief mention of self harm.

At first glance, everything is going well at my local middle school. Most students achieve highly academically and come from very well-off backgrounds with parents who are engaged in their student’s learning. But if you ask students that go to this school, they'll tell you a different story.

There’s a constant gloomy feeling that wasn't there before, and it’s increasingly becoming a worse environment to be in.

If you step into this middle school as a student, you'll notice the constant use of derogatory language and hateful speech by students right away. This feels out of place considering how progressive the area the school is in. You wouldn’t expect this behavior from such highly-achieving and well-off students. Grade level wise, this behavior started at the beginning of middle school; Before that, it was very uncommon. For quite a bit of time, I’ve been trying to find out what caused this large cultural shift in the environment entering middle school.

Originally, I thought homework could be part of the problem. But I could never quite get my finger on what was exactly wrong with it. I felt like there was some sort of underlying issue that I couldn’t quite grasp. So I tried to look at the bigger picture: the connection between my dislike of homework and the toxic environment of my school.

The grading system.

Finally, I realized that the grading system had been the underlying issue this whole time. Students were only so stressed because their grades were used as a measurement of themselves as students, and they were getting pressure from their parents to get straight A’s and do the very best. Incomplete homework gets a grade of zero, even if it's mostly done, and that grade reflects badly on students. No wonder the students are always talking about their grades or test scores and even cheating on schoolwork with each other so they can get a higher test score. I wondered whether the flaws of the grading policy was connected to this toxic behavior and student mental health

The standard grading system is straightforward and easy to understand. When you get a letter grade, you know exactly what it means. If you get an A, you did good. If you get an F, you failed. Easy and efficient. An article I found from THNK noted:

At first, grades seem like a textbook example of efficiency improvement, allowing the differences between students to be quantified and allowing teachers to understand more students in less amount of time.

This would be great, but using a letter to represent a student’s mastery of a subject just doesn’t work out. Letter grades are one-dimensional. As one teacher said,

“If at the end of the grading period when one of my students got a C in reading, I wasn’t sure if it was his reading fluency, his inability to understand cause and effect, or if he wasn’t constructing the sentences in an accurate way.” (Strobel)

The current grading system fails to work as a tool that's supposed to show where students are succeeding and where they are struggling.

💡 Most students who took my anonymous survey don't think grades fully measure their knowlege of each subject.

Data from ten local middle school students who took my anonymous survey.

The current grading system is also quite punishing to students, as it expects perfection and doesn't allow much room for mistakes. For example, if a student gets a low score on a test, perhaps because they’ve been going through a rough time, or because they’ve missed a couple of days when the teacher explained the concept, it permanently counts against them on their grade, creating unnecessary stress and causing the student to strive for perfection instead of focusing on what's important: learning.

Grades push students to strive for perfection instead of focusing on what really matters: learning.

This is hypocritical as we often hear about how students should learn from their mistakes, and how schools are the place to make mistakes. For example, my local school's handbook states:

“Let students make mistakes and fail. Some of life’s best lessons and “teachable moments” come from mistakes and failures.”

If this is the case and mistakes should be “teachable moments”, then punishing our students for mistakes in the form of low scores and zeros doesn't make sense.

On its own, the grading system isn't enough to create such a large toxic environment, because, really, it's just a tool. A flawed one sure, but still just a tool. And since everyone would use it as just a tool, it wouldn’t really matter, especially once people noticed how flawed it was.

But because of the emphasis we put on grades, it becomes a much larger issue.

When we mix in parents who want their children to succeed, as I hope all parents do, this is especially true. Often parents want their children to succeed in school, they do so by pressuring their children. This extra pressure causes all the aforementioned issues. Of course, when students are under such pressure, they act differently. Some students’ mental health gets worse, and others may lash out by being hateful, creating a toxic environment for everyone else as well. This is backed up by Madeline Levine, a practicing psychologist and children’s therapist, who noticed a pattern, specifically when a fifteen-year-old girl from a loving and financially comfortable family came to her with the word “empty” carved into her arm. Conversations with educators and clinicians across the country as well as diligent research confirmed her suspicions that competitive pressures in the school system made these kids so unhappy. In her book, The Price of Privilege, she references her own experiences as a therapist as well as other research. (Friedman; Levine)

The instant feedback provided by grades is co-dependent with parental pressure.

💡8 / 10 students who took my anonymous survey say they feel pressured to maintain good grades.

Data from ten local middle school students who took my anonymous survey.

So we need to find a way to break this this co-dependency between grades and parental pressure. Of course, we could try to stop this on the side of parental pressure, but this seems unattainable because the school can't control what parents are gonna do. To me, the only reasonable solution is to fix the problem with the grading system. If grades were not flawed but instead made a useful tool, there would be a lot less of a problem.

That presents the idea of what a world without traditional letter grades looks like. Of course, we would still need some other tool to measure a student’s strengths, weaknesses, accomplishments, and learning, but it would have to fix the flaws of the current system. After thinking about this for a while, I’ve made a theoretical system that, with some tweaks of course, would address most of the problems I'm aware current system has.

When creating this system, I wanted to:

celebrate student achievement and improvement
address a student's strengths and weaknesses
capture a student as a complete individual

Student progress would be measured in a very different way based on improvement, effort, and other such things. Students would still have to try their best, but the final result wouldn’t solely be determined by tests and quizzes. With this system, students wouldn’t be punished if they were struggling or if they made a mistake. Furthermore, student achievement wouldn’t only be measured by the result, but by the process of getting there.

A “score” sheet for each student would have different categories such as planning skills, follow-through, teamwork, understanding material, timeliness, etc., and would be scored on a scale similar to as follows:

Needs Work ➡️ Working on it ➡️ Improving ➡️ Good

These score sheets would only be available about four times a year, when the student and the teacher would review it together, highlighting what good work the student has done and what the student should focus on working on. I believe this would remove the pressure of standard letter grades as well as force people to not be so concerned about grades anymore, due to them not having constant access. Instead, this would finally allow grades to be used as the tool it should've been.

Quote from a local parent who took my anonymous survey.

While working towards a better grading system, we will have to work in baby steps. The easiest first step to take is to realize grades are supposed to be a tool and aren’t the most important thing in life. That mindset alone will help calm the toxic mentality of grading. The next thing I would propose is more lenient homework policies as well as policies that forgive zeros and low scores. If we continue making similar baby steps over time with the intention to try to reach a better grading system, we will eventually get there.

However, if you want to make a change right now, here's what you can do.

Students and parents - have a conversation about using grades as a tool rather than striving for perfection. Talk about the importance of mental health, and how grades alone do not determine everything about a student’s future. Discuss carving out times in the day for the student to pursue their chosen hobbies and activities.

Teachers - try to be more forgiving on your grading policies and frequently leave useful notes in the “notes” section of the grade book. Have ways for students to easily fix their mistakes and learn from them. If you are a less harsh grader but give more specific comments on improvement, it will allow your students more wiggle room to learn.

Everyone else - Advocate for better grading policies at your local school district by getting in touch with your school board.

By reimagining what grades look like, we open the door to a better learning experience for everyone.

Works Cited

“3 Reasons Grades Are Bad for Education.” THNK, 29 June 2021, www.thnk.org/blog/3-reasons-grades-bad-education/.

Friedman, Hilary Levey. “Children and Competitiveness.” Oxford Bibliographies, www.oxfordbibliographies.com/display/document/obo-9780199791231/obo-9780199791231-0061.xml. Accessed 4 June 2023.

“REDACTED's School Handbook” Google Docs, 2023, link redacted. Quotes from this source have been slightly edited for privacy reasons.

Levine, Madeline. “The Price of Privilege.” Madeline Levine, Ph.D., madelinelevine.com/books/the-price-of-privilege/. Accessed 4 June 2023.

Strobel, Kim. “Five Problems with the Traditional Grading System and Why Standards-Based Grading Is a Better Alternative.” Strobel Education , 30 Mar. 2023, strobeleducation.com/five-problems-with-the-traditional-grading-system/.

What the TikTok Algorithm shows about the future of AI

Mon, 23 Jan 2023 00:00:00 GMT

Recently my cousin was in town visiting my family for the first time in about 3 years. We quickly found things to make conversation about and we had a long catch-up about everyone's lives. During this catch-up session, I noticed that her, my sister, and I were all bonding over things that we have seen online, specifically on TikTok. So, naturally, I asked her if she had TikTok.

She didn't. She said that she used to have it but realized she was wasting too much time scrolling away on TikTok instead of getting her work done. I too have had the same experience. I felt like I was wasting away hours of my day mindlessly scrolling on TikTok. I also deleted the app to regain that time back.

This reminded me of a discarded podcast episode I made a while ago on the dangers of the TikTok algorithm. I took another look at the research I did for that podcast episode and once again found it very interesting. Because of that, I decided I should write a post mainly based on that research I did, but edited for an post format including more of my thoughts and opinions.

If you didn't already know, TikTok is a short-form video based social media app primarily used by teens. Its main appeal is its For You page, which serves up new content using an algorithm that attempts to give you content it thinks you'll like based on various statistics it collects. This is similar to other social media algorithms, except the main focus of the app, it's For You page, is based on this algorithm. In fact, the majority of time spent on TikTok is devoted to the For You page (powered by its algorithm), rather than the following tab or trending page like other social networks. This is what makes the TikTok Algorithm so significant, as it is providing the majority of the content that people view on the app. This algorithm is insanely good at finding out what will keep people on TikTok.

But to truly understand TikTok, you also have to understand its history.

The History of TikTok

In 2012, a Chinese entrepreneur in his late 20s founded a tech startup called ByteDance. He began the company out of a small apartment in Beijing and was inspired by the fact that people in China were using their phones to read the news instead of reading traditional newspapers. ByteDance created an AI-driven news platform, which was very successful, and the product still exists in China today. In 2016, ByteDance released a short-form video sharing app that allowed users to add popular songs in the background of videos. This was also marketed to brands and advertisers as a way to get in touch with the new generation, Gen Z. But since the app was based in China, it had to abide by Chinese regulations. So, for the app to go global, it needed to make some adjustments. After making the proper adjustments and splitting off the global app from the Chinese app, in 2017, the world was introduced to TikTok.

In its first year after launching globally, TikTok acquired Musically, which was another popular short-form video-sharing app, and merged them together. From there, TikTok's downloads skyrocketed mainly due to its appeal to young people. Over the following years, it overtook many other popular social media apps and made its way to the top of the charts.

The Appeal of TikTok

As stated before, the main appeal of TikTok is how insanely good the algorithm that runs its For You page is. To fully understand how good the TikTok algorithm was, I interviewed a local teen and TikTok User for more info.

According to this teen, they spend about 95% of their time in TikTok on the For You Page, powered by TikTok Algorithm. This Algorithm keeps users engaged with content on the platform.

What are some of your favorite things about TikTok?

Just like being able to look at it forever and just always do entertaining things.

The power of the TikTok Algorithm is further emphasized by the fact that, unlike other social media websites, it doesn't really matter if your friends are on the platform.

How much do you think it matters if your friends are on TikTok?

Not a lot. I think one of the great things [about TikTok] is that it's like an opportunity to make new friends. And like you got to learn new things that you enjoy, which leads to like making [mutual friends], which just leads to a lot more different friends that you maybe didn't have the chance to make before.

So the Algorithm has a lot of power over people. It keeps them engaged and addicted to more content. And you don't select the content you want, the algorithm does (although you do influence what it shows you, you aren't the one in control). So what's going on under the hood of the TikTok Algorithm?

How the TikTok Algorithm Works

We don't know exactly how the TikTok algorithm works, but we can get a pretty good look at it. In a YouTube video by the Wall Street Journal, they did some digging on the TikTok algorithm. They made hundreds of bots scroll on TikTok to find out how the TikTok algorithm works. Each bot was assigned ages and interests. None of the interests were ever entered into TikTok. The only way that TikTok knew any of the user's interests was by rewatching or pausing videos with related hashtags. At first, all the bot accounts got a variety of popular videos. But then as the algorithm saw what the bots were interested in, the view counts of those videos dropped more and more. TikTok's algorithm could learn these bots' interests in less than two hours, but some in less than even 40 minutes.

The For You pages of these bots started with general interests, but these interests got more and more extreme. For example, the bot with a general interest in politics wound up serving videos about election conspiracies and QAnon.

As you know, users spend 90 to 95% of their time on TikTok on the For You page. Just with simple factors such as how long you linger over a video and what you like, TikTok learns your hidden interests and emotions. And this can drive you down different rabbit holes of content on the platform.

Would you say that there are probably more toxic areas or more bubbles or more positive bubbles or like they kind of even each other out?

Thing is, I think there's less toxic bubbles, but they are bigger. [...] So like, while there is a lot more positive levels, there is just huge, like toxic bubbles.

Conclusion

The power the TikTok algorithm holds is wild. That paired with the insane accuracy of it makes for a scary combo. It shows us a glimpse of a potential future with AI if we are not careful. AIs like ChatGPT are progressing faster than ever and if we are not careful, these AIs will hold more and more power over us, which can only end in disaster. We should use the TikTok algorithm as a guide of what not to do, and a lesson on the dangers of AI if not managed correctly.

Although, as much as I covered the negative sides of TikTok and its algorithm in this post, it's important to note that it's not all bad. There are a lot of good things about TikTok and its algorithm. One example of this is that TikTok and its algorithm makes a place on the internet for people to reach a wide audience of people that they may have not reached before. It's another tool that can be used to spread new ideas and empower otherwise unheard voices.

The TikTok algorithm isn't the problem, it's us. We need more transparency with AI products and we need to be skeptical about most AIs for the time being.

We need to be careful.

Mentioned in this post:

The Rise of TikTok: From Chinese App to Global Sensation (Wall Street Journal)

How TikTok's Algorithm Figures You Out (Wall Street Journal)

No, Web3 Won't Change the Internet.

Sun, 22 Jan 2023 00:00:00 GMT

It's no secret cryptocurrency, NFTs, and other blockchain-related products have blown up recently. It's been all over the news, from the fall of FTX, to NFTs selling for millions of dollars only to crash in value. But one thing remains constant. Every-time a new blockchain based product comes out, it seems that we are hearing how revolutionary it is and how it will be the future of the Web as we know it and help shape a better "Web3.0".

What's Web3?

To understand what Web3 is, you first have to understand what Web1.0 and Web2.0 are. Web1.0 was the first iteration of the web. Web1.0 contained mostly static websites that you could view digitally. This is why Web1.0 is known as a "read only" model. Web2.0 was the next version of the Web, and the version of the Web used today. Web2.0 brought interactivity to websites, meaning that not only could you view a website, you could also comment on websites and post on social media. This is known as a "read-write" web.

Ethereum.org - What is Web3?:

"As more people came online [Web2], a handful of top companies began to control a disproportionate amount of the traffic and value generated on the web. Web 2.0 also birthed the advertising-driven revenue model. While users could create content, they didn't own it or benefit from its monetization."

Web3.0 is a concept for the next version of the web based on the idea that blockchains, cryptocurrencies, NFTs, and other technology will give the power of the web to the users. It's supposed to be a trustless system that will give more power to users and creators alike. A "read-write-own" model.

The underlying technology behind this is blockchain. Blockchain is effectively a giant spreadsheet with all transactions ever to have happened shared across everyone on the blockchain. This ensures no one can tamper with the history, since it would have to align with everyone else's copy of the blockchain data. This is a gross simplification of course, but long story short, it's incredibly secure.

The Problem with Web3

There are some problems with Web3 though. Beginning with that the idea of Web3 is very vague.

This "vision" for a better future web has no clear path of how to achieve the future in question.

Also, important aspects of Web3, such as cryptocurrency and especially NFTs, don't produce anything of material value. There is almost no good use for NFTs, as they don't really let you own an image or other file (you know, the revolutionary thing they are meant to do). With cryptocurrency, it's riddled with scams and pump-and-dump schemes, and is mostly advertised as an investment. But don't let the ads fool you, because most people will lose money on crypto, as it's basically a giant experiment in the "greater fool" theory of investment, which suggests that people can make money through buying overvalued assets by later selling those assets at a higher price to a "greater fool". I think this is getting close to scam territory, personally.

Also, another reason the idea of Web3 exists is because too much of the web is centralized around a handful of companies. While this is true, it's not like you have to bow down to these companies and do what they say in order to have a voice on the internet. Anyone can create a website for anyone else to use for fairly cheap nowadays. If you want a social media platform free from the influence of corporate giants, you can use Mastodon and other open social media platforms. Web3 doesn't need to fix the issue of centralization, as the technology of Web2.0 is just the right amount of centralized while maintaining the openness of the internet. If we really want to solve the issue of centralization on the web, we should get more people to use services like Mastodon and start their own websites and blogs. There's no need for a giant redesign of the whole world wide web.

Conclusion

In my personal opinion, the idea of "Web3" is mostly a marketing gimmick designed to get you to buy into Cryptocurrency and/or NFTs of some sort, so that whoever is advertising it to you can make a profit. With this said, Web3 isn't completely a bust. Blockchain, the technology powering most of Web3's ideas, is a very promising technology with lots of interesting use cases. Because of this, we should expect blockchain technology to show up in more and more places and industries. While the current technology built upon blockchain is not to be desired, the underlying technology is no doubt incredibly useful.

Elliot Hopkins' Blog

Cleaning up the mess I made building things I love

My complicated relationship with the em dash

How to know when to trust the news — from a student journalist

Do you really know what a journalist does?

The Basics of Journalistic Ethics

Finding out if a source is trustworthy

✅ Green Flags:

⚠️ Yellow Flags:

⛔ Red Flags:

glasswing butterflies

Little touches that make my website truly mine

So long, Arc.

Why I Left Ghost - and the process of moving my blog

If you want something done right, do it yourself.

Why I'm Paying for a Search Engine

How Letter Grades Cultivate a Toxic Environment for Students

There’s a constant gloomy feeling that wasn't there before, and it’s increasingly becoming a worse environment to be in.

The grading system.

Grades push students to strive for perfection instead of focusing on what really matters: learning.

But because of the emphasis we put on grades, it becomes a much larger issue.

The instant feedback provided by grades is co-dependent with parental pressure.

By reimagining what grades look like, we open the door to a better learning experience for everyone.

Works Cited

What the TikTok Algorithm shows about the future of AI

The History of TikTok

The Appeal of TikTok

What are some of your favorite things about TikTok?

How much do you think it matters if your friends are on TikTok?

How the TikTok Algorithm Works

Would you say that there are probably more toxic areas or more bubbles or more positive bubbles or like they kind of even each other out?

Conclusion

No, Web3 Won't Change the Internet.

What's Web3?

The Problem with Web3

Conclusion

Notes and Updates