Accessibility – Alice got Chained http://10.139.236.50:8023 and now she's breaking out of them. Thu, 19 Mar 2026 10:33:11 +0000 en-NZ hourly 1 https://wordpress.org/?v=6.9.4 We should have a Human Right to execute Unsigned Code http://10.139.236.50:8023/2026/03/16/we-should-have-a-human-right-to-execute-unsigned-code/ Sun, 15 Mar 2026 17:24:56 +0000 http://10.139.236.50:8023/?p=29 There is a unique branch of mathematics known as boolean algebra.

A boolean is either a 1 or a 0, true or false.

Boolean functions include AND, OR, and NOT.

Some have combined these functions into others, such as NAND, NOR, XOR, and XNOR.

It’s actually quite an interesting idea, but wasn’t thought about much…

…until humans invented devices that can function as logic gates for these functions.

That’s right. They combined electrical components to make logic gates. These logic gates can be combined to make a machine that can do stuff.

General-purpose computers are just fancy calculators

This eventually led to the integrated circuits of today.

The reason I bring this up is that there is a push by technology companies to make it so we can’t run software on our own general purpose computing devices, unless that software is approved by them. This is gatekeeping the functionality of a machine that would otherwise do it’s job.

The fact that embedded devices that are designed to do a specific task and nothing else (e.g. a game console) already exist, doesn’t change the fact of the matter that these are a different class of devices.

First, you got the single-purpose device class. Devices that do one thing, and do it well.

Then you got general-purpose devices. Devices that are specifically designed to do anything that a user programs it to do.

The problem is that these tech companies (and even some governments) want to prevent general purpose devices from being programmed in an ‘unauthorised’ manner.

My question is this:

Authorised by who?

Who owns the device?

If I had to choose between an iPhone (cost over 2 grand), and a game console (cost 800 bucks), neither of which allow for unsigned code execution, I’d get the cheaper option that actually has some good games on it.

If I pay for it, knowing it’s a general purpose device, I should be able to install whatever the fuck I want to install, run whatever I want to run, and destroy it if I ever want to.

If my choice is between a crap device that’s locked down, and a more expensive higher-end locked down device that can’t do any more than the crap one, I’d get the cheaper one.

This is compounded by another fact: Modern web browsers are designed to run arbitrary code on the website they’re browsing using Javascript and Web Assembly. Websites can do some rather interesting stuff with this, including running an entire x86 virtual machine inside the web browser. In addition, many modern websites will cease to function without the ability to run unverified code.

I mean hell, someone recompiled Super Mario 64 to web assembly using Emscripten, so it can be played inside a web browser. If only someone could be bothered to do this with GIMP, I’m sure it would mane many people’s lives easier if someone did, especially for iOS/Android users.

Locking down a general-purpose computing device is not feasible if one can run an entire operating system inside it’s web browser now, is it?

Then there’s the other problem. Stupid governments thinking they can dictate how operating systems work, locking them down to ‘protect the children’ by requiring age verification to use them:

  • How are you going to enforce this against older devices?
  • How are you going to prevent people from compiling their own operating systems, like what you can do with Linux distros?
  • How are you going to prevent people from using older operating systems, or smuggling alternative ones on SD cards?
  • How is this going to effect system users that aren’t representive of an individual person, such as ‘Admin’, ‘root’, or ‘nobody’?
  • What are you going to do about virtualisation and emulation?
  • What are you going to do about offshore developers that don’t give a shit about your bullshit?
  • How are you going to prevent people from building their own computes?

And let’s not get started with that ‘censoring the internet to protect the kids’ bullshit. This can easily result in access to life-saving resources being restricted from access to those who desperately need them. I can’t stress this enough, as someone who is in such marginalized circumstances (intersectionally, of course: trans and disabled).

Locking down devices will also make it much harder for people to learn how to develop software.

Another reason that companies use as an excuse to lock down their platforms is ‘security’, as if sandboxing and virtualization techniques didn’t solve that problem already.

Google cited fake banking apps as a reason to lock down Android, despite the fact that progressive web apps exist, essentially allowing anyone to fake a banking app? Someone should tell them about that.

Yes indeed, you can make a functional website that can seamlessly be installed as a full featured app, with many browser APIs providing many necessary app features too.

Web browsers have APIs for notifications, battey status, clipboard, file system, geolocation, device vibration, device sensors, MIDI devices, WebGL and WebGPU for 3d rendering and AI workloads, content security / DRM (if desired), WebRTC for real time audio/video communication, picture-in-picture, screen capture, and service workers (allows for offline web apps), just to name a few. These are more than enough for many apps, like social media apps, communication apps, streaming apps, navigation apps, text/image editors, and games. As such, for many apps, there is simply no real reason to publish to an app store anymore. Just maintain a generic web app, and one doesn’t need to compile for multiple platforms, or wait for any sort of attestation.

There are some experimental upcoming APIs for USB, Bluetooth, NFC, HID devices, idle detection, VR/AR/XR, payment, audio session, credential management, etc, that should allow for even more advanced web apps. A good list is available here.

The only thing this kind of device lock down will effect, are the apps that actually need to be natively run on the device, with platform-specific APIs. This includes accessibility utilities, automation apps, device assistants / smart home apps (at least until a WebAPI is made for this), media players, many games/emulators, VPNs, certain security apps, and other system-specific apps, like button remappers and widget apps.

Another thing this will effect, is offline app installation and offline app development. If there isn’t a workaround (e.g. Android’s ADB installation), this will prevent apps from being installed in mission-critical circumstances where the internet may not be available. E.g. deploying a Bluetooth communication app during an emergency, like a natural disaster, or a war zone.

Advice for politicians, and operating system developers:

If you cowards are concerned about children:

  • Don’t touch the operating system.
  • Only enforce age verification onto applicable app stores that provide 18+ content for purchase, like Steam or Google Play. All others should be considered unenforceable, due to alternative app sources, and the lack of verification that comes with that.
  • Leave the rest of the responsibility to the parents and caregivers. Any failure of parents’ to screen for inappropriate content should be considered negligence at best (if unintentional), and abuse at worst (if intentional). There are plenty of parental control tools that may be deployed if needed for compliance (i.e. for prevention, or for resolving a parental negligence situation).

If you cowards are concerned about malware:

  • Secure the system, and provide safe guards, like malware scanners, and additional user confirmation or a disclaimer for unverified app installation.
  • DO NOT PREVENT THE INSTALLATION OF UNVERIFIED APPS. Provide a sandbox for them if necessary.
  • Consider the possibility of progressive web apps masquerading themselves as app clones, before even thinking about malicious apps that masquerade as other apps, since web apps can bypass web filters as trivially as using another domain/subdomain/subdirectory. And again, provide realtime anti-malware scanning capability for detecting known malicious apps, and provide additional user confirmation or a disclaimer for unverified app installation.

Advice for centralised social media platforms, and porno sites:

  • Bring your own age verification, or pay someone else to do it. Especially for social media sites, since they already have enough information to verify age of users in most cases. As for porno sites, this can be as simple as verifying age via a credit/debit card processor (they can verify age at the bank’s KYC level).
  • Allow the option for a program, app store, or browser plugin to verify a users age if desired, but don’t use it as the only option, as while such a solution may be convenient, it will not be available in all circumstances.
  • Do not blame the operating systems for your lack of oversight. Own it, and clean up your own mess.

Read the advice above, and don’t make any of our lives worse. Amateur computing will live on in the same way as amateur radio, just without the need for a radio license. Hardware will always be available, and hobbyists will always get their fix, especially as STEM education fields remain increasingly important as time goes on.

If you bastards kill the internet, we will make our own. This can be as simple as running Yggdrasil (or a future variant of Yggdrasil?) on all computers, as it can bridge multiple transports, including leased lines, WiFi mesh, local networks, enterprise networks, metropolitan area networks, satellite links, licensed/unlicensed radio links, and the many surviving parts of the Internet. I2P can be run on top of it if needed.

]]>
I used TikTok for a year. Here’s what I found out about it. http://10.139.236.50:8023/2026/03/10/i-used-tiktok-for-a-year-heres-what-i-found-out-about-it/ Tue, 10 Mar 2026 07:51:40 +0000 http://10.139.236.50:8023/?p=12 Alright, I’m gonna write a testimony of my experience using TikTok a few years ago. This predated the proliferation of AI generated slop.

The app was surprisingly well designed… Too well designed…

It has a doom-scrolling casino-like infinite scroll design. You never know which video will show up next. Each swipe feels like you’re hitting the lever of a slot machine. Some may not think much of this, but this can put one in danger of the recommendation algorithm taking them to a place they wouldn’t expect to end up.

The videos themselves ranged from memes, to tutorials, to trolls and obvious rage bait. One of the more interesting features were the ones that allow for making reactions to other content, allowing for quite the interesting reactions and memes.

Occasionally, an ad will show up. These were surprisingly non-intrusive, and could be easily swiped passed. This seemed to promote a unique advertising META that encouraged advertisements to actually blend in with the other content on the platform, because as soon as people swipe and see a blatant McDonald’s ad, they would immediately swipe past it. This advertising system is something I actually have to give credit to, as it discourages the usual un-creative slop that usually plagues advertising.

As usual, I didn’t engage much with the comment sections of videos. I knew those were nothing but trouble since I saw the ones on YouTube. It didn’t take long to find everything from mundane insults to entire flame wars in the comments.

I also didn’t watch live streams on TikTok. It didn’t feel natural to me.

The more I used TikTok, the more I could feel my brain start to malfunction, with stress and anxiety levels peaking far easier than before I started using it. I only used TikTok sparingly. It wasn’t an every day thing. I only used it every few weeks, but even then, I could still feel the effects.

After around a year of usage, I deleted the account, and the app entirely.

Alright, let’s boil it own to the problems that I saw:

Accessibility Issues

Undocumented features

My hands are very finicky with touch screens. Sometimes a tap or a swipe may not register as a single tap or swipe. This lead to a problem: The app had a button I can press to like a video. I see it, I recognise the shape, I know what it’s for. Simple as that. But TikTok is an entirely different beast. It kept occasionally liking the video when I tapped the screen. I thought this was a bug, until I looked in deeper.

It turns out that TikTok has an undocumented ‘double-tap to like’ feature, that can not be disabled, and this is a problem. It felt like I wasn’t controlling the app. It felt like it was deciding to do stuff on it’s own. This is an example of railroading a feature that can effect users in a negative way.

Inconsistent design

When one swipes up or down, they expect to go to the next pr previous video, but TikTok railroaded a feature that allows for image slideshows to be posted. This wouldn’t be a problem if it wasn’t for one thing: the autoscroll. After a couple of seconds, it would automatically switch to the next slide. Not problematic yet? Well, did I mention that when you try to swipe vertically to change the video while the horizontal slide is changing, that it would cancel out the vertical scrolling? The horizontal slide change would take about a second to complete, and happened every 2 seconds. This vertical scroll-cancellation happened repeatedly, and pissed me off beyond explanation. I couldn’t filter out those slideshow posts, as they couldn’t be disabled. It’s almost as if I wasn’t in control of my own device again.

Obfuscated UI element names

While making an accessibility frontend using Tasker and AutoInput to overcome the above accessibility issues, I struck another issue: All the UI elements were obfuscated in a way that wasn’t consistent between app updates. This inhibits any accessibility software from recognising UI elements correctly and consistently, and makes it nearly impossible to make an app-specific accessibility overlay for TikTok, especially when UI element positions and sizes depend a lot upon the device itself, and the device’s settings. I’m pretty sure this is against Android’s accessibility guidelines for UX design, but I’m not a paperwork wizard, so what could I possibly know?

All I have to say is: GOOD LUCK USING THIS APP IF YOU’RE BLIND, LOL!

The usual social media crap

Doom scrolling infinite content machine

That platform was fill of garbage content, misinformation, and outright lies. This kind of stuff always shows up on user-generated content platforms like TikTok, but this time, it was exacerbated by the fact that on TikTok, you don’t browse for content (they do have a search feature, but it’s garbage). Instead, the algorithm feeds you what it ‘thinks’ you might like.

This can lead to a rabbit-hole effect, where you can start from knitting tutorials, and end up in some alt-right Qanon conspiracy niche that believes that a certain ethnic minority is drinking the blood from aborted fetuses or some shit, and the worst part is: the more these videos show up, the more watch time the algorithm sees, and thus, the more of this garbage is recommended by the same algorithm.

Image describing the problem with content recommendation algorithms going into loops, and how bad algorithms can come to stupid conclusions.

I think there’s a mathematical fallacy demonstrated in the image above…

The problem with these recommendation algorithms is that we don’t know how they work. We can, however, assume they keep total platform retention in mind for this. since they want to keep users on the app for as long as possible.

What keeps people on the platform longest? Why, just make it so time consuming to find content one actually wants to watch, and prioritise recommending shit that outrages people enough to start a flame war in the comments about it, send push notifications for each reply to these comments, and there’s the interaction and retention loop for ya.

This is an incredible method of distributing not just funny memes, but also rage bait, propaganda, and now with TikTok Shop existing, dick pill salesmen and other snake oil merchants.

Exploitable Ads

While I commend TikTok’s respectable effort to making a more sustainable advertising system than most other platforms do, this doesn’t change the simple fact that TikTok is rife with bad content. This is especially concerning when combined with the in-app advertising system, which makes the ads blend in with the actual content stream, making it hard to discern an ad from genuine content without carefully checking the content description each time you swipe, meaning any possible disclosure is therefore an afterthought in the user’s mind.

Online advertisers have a long lasting habit of poor quality control, and an eagerness to accept foreign political propaganda into the ads. Some astroturfing organisations and hate groups have used this method to spread their slop, but that’s an article for another day.

Push Notifications Galore

Imagine waiting for an important text message to come through, only to be flooded with random notifications that drown out the important stuff? Well, that’s the problem with platforms being too eager to escape the app they belong inside.

I’m a simple person when it comes to notifications. I enable phone calls, text messages, emails, and important web server notifications. Everything else can get bent.

Want to recommend content for me to watch? Wait for me to open the app and check. Online content doesn’t usually have that narrow of an expiry date. It can wait.

Want to notify me about comments and replies? Wait for me to check the app. It can wait.

I cracked down on this push notification stuff a long time ago. I realise that their only existential purpose is to hook you back into the app, and encourage an addiction cycle in the user. It’s sick, it’s predatory, it’s just bad for the attention span of the user (as if TikTok’s content format wasn’t already bad enough for that).

Chinese spyware concerns

One of the main concerns of the time is that TikTok’s business operations were based in China, and you know what China is like. The CCP want to spy on everyone.

Many people are relieved to hear that the US have obtained TikTok, but I’m afraid the US is no better than China, with all their bad privacy laws, data brokers and all that fun stuff.

I would go as far as to say that almost every centralised social media platform has this issue in common, as all the users’ data is stored in one place, easily accessible by law enforcement, foreign spies, or those participating in industrial espionage and social engineering attacks.

Don’t expect any of these platforms to be keeping your personal information safe, especially when many of them have a profit motive to sell users’ personal information.

TikTok Clones

The only thing worse that TikTok is a TikTok clone.

All of a sudden, every other platform started blatently copying TikTok, often with even worse results, which is funny. It’s usually the Chinese who make knockoff products of US goods, not the other way around!

YouTube made YouTube Shorts, which is somehow even worse than TikTok.

META/Facebook have cloned TikTok, with things like Instagram Reels.

Alternatives to TikTok

and other platforms like it

Decentralised and federated social media platforms like Mastodon, Peertube, and others exist, that can be self hosted on one’s own server if they wish. Loops is probably the closest to TikTok that these get (being an obvious clone), with Peertube being a close second (Peertube is a good YouTube clone). Most of these federated solutions are open source software, and can even cross-federate with each other in many situations.

While it is possible to gather metadata about everyone’s posts from a compromised instance, it doesn’t compromise the private personal information of users from other servers, as only their public profiles and public posts are available for public scraping.

Since these servers are operated by individuals and/or isolated groups, and the software is free and open source, there is no profit motive for predatory design over the whole social media network, or the software itself. Either find a server willing to accept you, or host your own.

These servers can even work and federate over darknets, like I2P and Tor. (I2P is probably the better one for this specific purpose, has much more flexibility, and better peer-to-peer capacity) The only requirement is that the network support connections over TCP either directly, or via a proxy, so other alternative non-anonymous networks like CJDNS and Yggdrasil, will also work with this.

Server operators can also blacklist servers they don’t want to federate with, or disable federation entirely. This helps administrators weed out bad eggs when it comes to poorly moderated servers.

I think the federated approach may be the future of social media, especially with things like age verification being enforced onto the larger centralised platforms. It also prevents one country’s stupid laws or one company’s stupid decisions from compromising the entire network.

While this might not be a perfect replacement for traditional social media, and may have their own problems, it’s certainly better than the alternative.

Conclusion

TikTok is garbage… plain, stinking, garbage… but it’s scarily well designed garbage, I’ll have to give it that.

A good analogy is to describe it like one of the original crack houses of social media: Many try to copy it, but most resort to lacing their drugs with fentanyl.

It’s just bad, and my mental health took a huge impact after this temporary use of TikTok.

Just stay away from it.

]]>