Newsletter Mix #03

Ån

Where do we begin?

As of a few weeks ago, Feral Research LLC became a real entity. As of a few days ago, we actually signed a handful of clients and some neat projects, not enough to really celebrate yet, but we're alive until the end of the summer.

I'm looking to focus on human-centric and internet-reconstruction technology and strategy work and consulting. I'm not sure where this is going yet, maybe nowhere, and this isn't even an announcement so much as a very light line in the sand, but between you me and the handful of people on this mailing list (and per the last newsletter) I am not happy with the current state of things. And: 1. Have been here longer than the current batch of people running our industry and 2. I intend to outlive the worst of them.

I don't know how this goes, but we've got work to do.

Earlier this afternoon a small community tech chat that I'm a part of (less than 40 people) started discussing two things simultaneously: 1. The consequences of opening up the group or merging it with a larger one and 2. The viability of Stoat, a Discord-alternative (formerly Revolt).

The first topic was an excellent chat about what we all got out of the group, why interacting at different scales felt harder or easier for each of us, what we valued out of being able to communicate with peers vs. feeling "on stage" and so on. The second topic came up because a friend of mine who I share some Discord chats with (less than 100 people each) was interested in what it would take to spin up our own servers in light of Discord forcing facial-recognition based age-verification onto the platform.

So Uhh... what is going on with age verification?

I do have a lengthy analysis I want to share at some point on what exactly is going on with age-verification and trust-and-safety in this moment. It's a strange-soup of conservative moral politics, legal and financial wrangling, lowest-common-denominator, lazy T+S efforts primarily focused on liability-risk-reduction and misogynistic and anti-queer sentiment tied up with a "save the children" bow. And boy is it a doozy, just in time for the midterms. If you happen to be reading this in MA, please educate yourself about 359. Because it's somewhat urgent, here's the letter I sent to my state reps.

Suffice to say if you're even tangentially aware, either as a worker or a user of modern social and UGC platforms: shit is NUTS right now. Ask yourselves how the exact same MAGA cult can spawn Q-Anon, Pizzagate and simultaneously maintain Epstein buddy Donny "grab them by the pussy" Trump as the guardian of Christian family values and you're barely starting to scratch the surface.

Ultimately it's just easier to outsource In loco parentis to a SaaS platform, even a poorly written AI-based one, because that's far cheaper and less complex than doing the right thing by your users.

So that topic deserves some detailed unpacking, especially if we'd like to avoid repeating providing cover for Epstein-scale human trafficking in the coming years. I'm not tackling that today, but if you want a primer I recommend starting with the Financial Times 2024 podcast Hot Money, Porn, Power and Profit. In particular, pay attention to the story of OnlyFans and how its existence today is largely a reflection of who handles the payment processing. Then ask if you'd prefer your kids be raised by yourself or the dude that owns Visa.

For today, and for very different reasons, the conclusion is that Stoat isn't going to replace Discord, even for our small-scale use case. The reason has nothing to do with the UX or the UI or most of the features it offers, it won't work specifically because secure, multi-mux, high-definition, high-framerate video and audio sharing that is resistant to bad internet weather and also doesn't leak your IP is a hard problem.

It tastes the same (if you close your eyes)

If you ask most people if X is a good replacement for Y, they will provide you with a point-by-point UX comparison, and/or they will wax poetic about protocol choice and how it is useful for supporting crypto-trading and free speech without limits while providing protection for whistleblowers under state surveillance. If you configure it right. Sometimes. Kinda hard to use though, it just needs stickers.

It turns out that while free speech and protection for dissidents is a valid and valuable use case and while interface is what most people understand software to be, neither describe the reality of the platforms we use or why we use them.

Very few of us are threat-modeling operating out-of-view from state surveillance and almost none of us will switch platforms even if you completely replicate the UI/UX of an existing platform, because neither of those things are actually important to us, not really.

This is mostly what I want to talk about right now: it seems self evident to me that scale-for-scale's sake has destroyed the internet and, arguably, the fabric of civil society. It seems useful to maybe flip the question and ask what problems scaling does solve for us, because that might illuminate where the work is that remains to be done. We need this to get a grip on the problems we should tackle if we're ever going to have a shot at making a better world for ourselves.

Furries (and everyone else) should probably stop using telegram. It is not safe to use.

But, they won't.

Most people won't stop using Facebook either, and apparently there are still more than a few people still hanging out on the dead corpse of Twitter? In the spirt of harm reduction then, we should figure out why.

More specifically, we should ask what is the core problem that is being solved by a given platform better than any other? (hint: In Telegram's case it's neither security nor security theatre, it's user-to-user privacy, which is something quite different).

What problem does scale solve?

An easier way to put this: why exactly does Discord exist?

Ask most users (those who are not cult leaders, influencers, politicians or corporations), and they'll tell you that Discord is a platform which provides a "third-space." It's a place to hang out and share moments with friends. Primarily gaming, but also anything you can put on a screen and talk about. That turns out to be a whole lot of things, and the user privacy and group-centric model means you can do it specifically with the people who also care about those things while removing or limiting the people who will cause you grief. It feels like a cozy, private internet for you and your like-minded crew. This, ultimately, is what we've all always wanted: a safe place to be validated and get-social with a concentration of people we care about.

I've never met a user who was not a politician, cult-leader or advertiser and who wanted a fully-public internet. Massive growth and outsize influence is lucrative, yes, but it's scary, unwieldy and corrosive and not where most people want to live.

When most of these early systems were built, sharing your status "online" meant a smaller pool of humans saw it. We designed for that scale, then handed it to a different generation of growth-hacking managers who exploded these systems well past their safe and useful boundaries.

I'm grateful for this nearly 20 year old article about how we ought to handle user data as carefully as nuclear waste, although then (and now) the headline reduction is wrong. I wouldn't read that as a call for not handling user data but rather a call for handling it with care (a responsibility we have completely and utterly abdicated as we find ourselves in our pre-superfund era.)

But even if things are undisputedly a dumpster-fire at the moment, understanding how we got here and the tendency we have to prefer civil discourse over slop and noise draws a clear circle around the platforms and communities which are thriving today versus those that dominated the last 15 years.

Discord, Telegram, SecondLife, Grindr, Nevermet, Fetlife, Google Drive, Facebook Marketplace (but not Facebook itself) are doing alright. The "big commons + real names" platforms like Meta, Facebook or Twitter, Truth Social? Not so much.

We clearly crave going to ground with our own like-minded friends and our real time and money is spent there. This is the crux of the "Dead Internet" theory and I suspect it's far more than theoretical. The myth of a coherent worldwide public commons is cooked, dead, buried. And frankly? Good riddance. All those efforts ever did was convince people they had a grip on reality that was never real to begin with.

The Dark Forest

These days there are entire movements and organizations designed around the principle of smaller communities in plain sight (see: Kickstarter founder Yancy Stricklers Dark Forest theory of the internet and DFOS project). This is ultimately an overlay network: the same idea that underpins VPNs, the dark web and BitTorrent. A network on top of (or within) a network. And it mirrors the invisible ties we're familiar with from life offline: no technology requires me to "friend" someone to visit their home, and my neighborhood is traced with my own personal mental map of relationships. The house where I met an ex across from the apartment where I lived near the place a friend once threw a party near where I got that bad news that one time.

Perhaps more interestingly (because for various reasons it now exists on Discord but not independently), I'm deeply interested in New Design Congress' Undersco.re soverign project, and I'm watching the European Union's internet sovereignty efforts with deep interest.

Because in order to accomplish this sense of cozy, to sustain an overlay network, the infrastructure needs to work and it needs to work well. Technical sniping aside, Discord works well. Its killer feature, fundamentally, is high-definition, high-throughput video and audio streaming that is reasonably secure and tolerant of large variations in network quality and weather. In short the thing people want Discord for (a place to just hang out whenever they feel like it) is made possible by an enormous amount of effort to "keep the lights on" 24/7, internationally, without most users even noticing.

And for most users, it does that for free.

Discord exists because it uses economy of scale to solve a single hard problem: secure and reliable video and audio streaming. It then gives that away so that users take it for granted. That's it: that's the user lock.

No platform, service or project will ever replace Discord until it can handle video in the same way. Everything else is a compromise that feels unnecessary because this centralized platform exists where you can just have it, without cost to you.

Scale provides "difficulty-arbitration," but time makes "difficult" into "easy"

The tech industry largely makes its insane-scale mark by arbitraging difficulty. The history of the success of most of our communication platforms draws a straight line from a Hard Problem to centralization by solving something difficult and then giving it away "for free."

Thus YouTube solved sharing video on the internet. Justin.tv solved streaming, Whatsapp solved status. Newer entries in this mode solve VR and AR world distribution and file sharing. Google solve email and word processing.

Making these hard problems easy for users builds platform buy-in, but the same industry can also increase their profits by making these hard problems easier, and thus there's a distinct benefit to making these difficult problems shared, open-source or both. Maintaining infrastructure is difficult. Writing operating systems is hard. But what if you could just nerd-snipe the planet and get that work done for free? Over time, the hard problems are no longer so hard anymore. Which means if you pay attention, the platforms tend to outlive their usefulness, running mostly on momentum.

So vanilla HTML now handles most of what Flash used to. Edge caching MP4 files is sufficient for most cases of "streaming." LetsEncrypt gives away certificates that used to cost $900/year. Cloudflare gives away devops tools. Spinning up a server on the open internet is close to free for most people and most use cases. Gitlab holds all the code. The internet is three websites in a trenchcoat.

I don't mean to devalue any of this - in many cases the fact that so much utility is available free of cost is a function of folks who understand there is value in quality infrastructure. What I want to highlight is that understanding most of this is free even without "platform" helps us zero in on the original "hard problems" we need to solve to build viable alternatives.

The most common mistake self-soverign alternatives to centralized services make is to replicate the "platform" aspects of an existing service before, or instead, of solving the hard problems first.

"Friending" is not difficult. "User profiles" are not difficult. "Dark mode" and "stickers" are lovely, but not killer features that will lure folks away from well-lit well-maintained servers to your $5/mo droplet that is likely to be down in a week.

I don't want to say that doing these things well is trivial, but you don't solve secure distributed video and audio sharing with a nice UI for friends lists.

If you are concerned about Walmart having destroyed your local economy, the solution is not to build a differently-branded big-box store. The solution is to refocus your efforts on solving infrastructural problems well and re-imagining the economy along community lines. Make the "free" centralized infrastructure less attractive (especially if it's the ONLY option) and I firmly believe that smart people will accrue themselves around that energy. Tying that work to platform-scale thinking, however, is counterproductive.

Does your platform for 100 people need the same tooling professional YouTubers and AI content-slop farms need? Do you need to gather and keep the same user information as Facebook so you can sell it to advertisers? Do you need to lock down your users content with DRM so you can implement the RIAAs takedown system and negotiate with Nintendo? Does every message need to be cryptographically secure (actually yeah probably) in a way that you can only ever access with a yubi-key on a third-thursday from your jailbroken android runing kali (probably not).

Build useful infrastructure instead. Bulletproof, community-oriented basic services and operating systems. Solve hard problems elegantly. Solve at human scale, design well, and aim for making things last forever.

After that, if you really want to make some liquid-glass inspired knockoff buttons for your Figma portfolio, have at it.