Friday links: December 5, 2025

This week I'm re-launching my links digest. Every Friday, I'll share a handful of pieces that caught my eye, usually somewhere in the messy, fascinating overlap of technology, media, and society. Occasionally, I'll stray further afield.

This week: what might come after the AI hype cycle, how new community platforms are trying to rethink participation, why our urban spaces may be pushing kids online, and how the Trump administration is using immigration enforcement to pressure online communities.

These pieces share a familiar tension: the friction between communities we build for ourselves and systems imposed from above. It's a question that runs through all my work: how do we design technology that strengthens communities rather than extracts from them?

What Happens After the Hype? Lessons from Mobile Internet’s Long Road to Success

We’re beginning to see the light at the end of the AI hype cycle. That doesn’t mean that there aren’t uses for aspects of the technology, but it does mean that some of the hyperbole will diminish as investors and speculators move on to the next thing.

As Ecrue founder Shomila Malik points out here:

“The question isn’t whether the current AI investment cycle will face a reckoning. It’s what form that reckoning takes — and what comes after.”

The lessons she draws from the mobile industry’s hype and decline also parallel what happened during the dotcom crash, when a lot of companies went away but a lot of underlying useful infrastructure was left for the next generation of innovations. But a facet of how those two events were different is exactly how they imploded:

“The difference between a pop and a deflation often comes down to how adaptable the infrastructure is. 3G networks built for one vision of mobile internet ended up powering something completely different — but they still got used. The investment wasn’t wasted, just redirected. Time will tell if AI will be a deflation like mobile internet or a ear deafening explosion like the dot com crash.”

Either way, investment is way ahead of proven capabilities or even business models. Companies like OpenAI are losing money hand over fist. At some point, these endeavors have to touch oxygen, and either they’ll find their way to stunning profitability, or they’ll fizzle into acquisitions at best and leave some interesting ideas behind.

My bet? Ten years from now we’ll be looking at a series of smaller, more focused models that perform well-scoped tasks really well, and we’ll look back at the hype around generalized megamodels — and particularly AGI — with rolled eyes and a slight shudder when we remember the environmental and human impacts.

Introducing Roundabout

Really interesting to see New_ Public announce its first community product from its Local Lab:

“The main thing to know, maybe the most important thing, is that this is not just another social media app. Roundabout is a community space, built from the ground up with community leaders and neighbors.

[…] As a project incubated within New_ Public, a nonprofit, Roundabout will grow incrementally, sustained by a diverse and balanced set of revenue sources. With business incentives aligned towards utility and everyday value, instead of engagement and relentless scale, we’re designing Roundabout to be shielded from the cycle of enshittification. The ultimate goal is to build for social trust — every decision, every design, optimized to build bonds and increase belonging.”

There’s a lot to comment on here.

It’s amazing to see a social product co-designed with communities. For the safety and equity of all involved, this is how it should be done. I really hope New_ Public shows off more of its methodology in the future. I’d love to dive into the meta-conversation about what they’ve learned about this kind of co-design. The descriptions of participating communities — in Burlington, NC; Richmond, VA; Lincoln County, WI; North Chattanooga, TN; and Lancaster, PA — are already really promising.

The technical lead is Blaine Cook, who you might remember as Twitter’s first employee and first CTO. Since then he’s been a strong, sharp advocate for decentralized social.

On Mastodon, New_ Public mentioned that it’s building the platform in a way that’s compatible with AT Protocol, although it’s not the main focus for now.

Over on Bluesky, Blaine said they’re “building on atproto primitives but off-network because it's currently not possible to push private/scoped data around the wider atproto network.” He also made the important point that it’s not worth building for interop until you know what the user behaviors are actually going to be — so it’s too early to focus on decentralization.

That community co-design is key, and it makes sense that this is the first step. Communities are human; they can’t be defined by protocols. The protocols should describe real human behavior, not the other way around.

I’m excited to see how the platform develops, and how New_ Public seeds the ecosystem conversations around it. And: this is only one of its community initiatives. There’s more to come.

Where Do the Children Play?

This meaningful discussion touches on the role — and dangers — of online spaces in the lives of children, but has a lot more to say about how our lives and environments are designed overall.

“[…] Digital space is the only place left where children can grow up without us. For most of our evolutionary history, childhood wasn’t an adult affair. Independent worlds and peer cultures were the crux of development, as they still are among the BaYaka; kids spent their time together, largely beyond the prying eyes of grown-ups.

But in the West, the grown-ups have paved over the forests and creeks where children would have once hidden. They have exposed the secret places. So the children seek out a world of their own, as they have for millennia, if not longer. They find a proverbial forest to wander. They don’t know what we know: this forest has eyes and teeth.”

There are some really striking statistics here. 45% of American children aged 8-12 have not walked in a different aisle than their parents at a store; 61% have not made plans with friends without adults helping them. That’s so far away from my own childhood to be unrecognizable. It’s also wildly oppressive. Of course kids are looking for spaces where their helicopter parents aren’t constantly hovering overhead.

The biggest source of underlying fears from parents, as well as feelings of isolation from children themselves? A car-centric culture. Parents are worried about their children being hit by a car, which unfortunately isn’t unreasonable. Children, on the other hand, can’t drive, and often find themselves fully unable to visit friends or have their own lives without the participation of someone with a license.

So, yes, online spaces need to be safer for children — but our lived-in spaces also need to be more human. American infrastructure in particular is architected for control; it’s rare to live in a walkable, safe environment. Jane Jacobs would have had a lot to say about this analysis, and her own solutions still hold up. We need to stop dividing communities, break the influence of the car, and create spaces that allow humans to roam, live, eat, and play — permissionlessly.

Trump administration orders enhanced vetting for applicants of H-1B visa

The US has taken the extraordinary move of blocking H-1B visa recipients who work to keep online spaces safe from abuse:

“The cable, sent to all U.S. missions on December 2, orders U.S. consular officers to review resumes or LinkedIn profiles of H-1B applicants - and family members who would be traveling with them - to see if they have worked in areas that include activities such as misinformation, disinformation, content moderation, fact-checking, compliance and online safety, among others.

If you uncover evidence an applicant was responsible for, or complicit in, censorship or attempted censorship of protected expression in the United States, you should pursue a finding that the applicant is ineligible," under a specific article of the Immigration and Nationality Act, the cable said.”

The message here is very clear: the people who make online communities safe are not welcome in the United States. Trust and safety is a very wide field, which encompasses the policies, processes, and technologies online platforms use to protect users from harm, ensure a secure environment, and maintain user trust. Compliance ensures that safety rules are adhered to. None of these activities constitute censorship.

These capabilities are required to make online communities livable. Without them, online spaces become toxic very quickly. The ultimate effect of this policy — if spread beyond H-1B visa holders — would be to make US-run online spaces unusable, and drive users elsewhere. If you want the future of social to be defined in other countries, it’s a great path to take. Otherwise, it’s stunningly short-sighted.

My Life is a Lie: How a Broken Benchmark Quietly Broke America

Simplify Asset Management’s Chief Strategist and Portfolio Manager Michael W Green examined how the US poverty line is determined, and discovered that it has been wildly miscalculated for years. In fact:

“[…] if you measured income inadequacy today the way Orshansky measured it in 1963, the threshold for a family of four wouldn’t be $31,200.

It would be somewhere between $130,000 and $150,000.”

The big differentiator is childcare, which in the US averages out at $32,773 a year, but it’s not the only differentiator. Our costs are enormous, and a poverty line at $31,200 only really helps legislators avoid having to provide (and therefore pay for) real support.

A family of four that genuinely earns $32,773 will receive all kinds of state help. A family that earns $80,000 does not. As Green points out, the difference largely comes from costs that went away during the pandemic:

Childcare ($32k): Suspended. Kids were home.

Commuting ($15k): Suspended.

Work Lunches/Clothes ($5k): Suspended.

Of course, many incomes also went away, depending on the jobs that were keeping these families afloat. Knowledge workers were relatively sitting pretty, while people who worked in retail, etc, were in trouble. There’s a lot more help that we can provide everyone. But this is one reason why I cannot stand return to office mandates, particularly when peoples’ salaries are under the $140,000 threshold that Green identifies. (Hint: outside high earning categories like big tech, it’s almost all of them, and inside those categories there are plenty of people who are earning lower.)

In future posts, it sounds like Green is moving to show that 401(k)s and similar instruments are also a scam for most ordinary earners — something I tend to agree with (at least compared to more pro-social alternatives). If society is an operating system that allows people to live well, start businesses, be healthy, etc, it’s failing us on every level. I say it’s time for an upgrade.

"Disagree and Let's See"

This feels emotionally honest and an idea I can get behind, as an alternative to the popular “disagree and commit”:

““Disagree and let’s see” allows you to stay aligned with the team without forcing you to pretend you had conviction you didn’t have. It lets you walk into a room with your team and be honest:

“Here’s the path that was chosen. It wasn’t my first pick, but here’s the experiment we’re running, and here’s what we’re trying to learn.””

Committing to something you disagree with is an emotional contortion that is hard to do in practice. But the work of every team is a series of experiments at its heart, and by changing the onus from “let’s commit to this thing we don’t all agree with” to “let’s try it and see what happens”, we move from steamrollering dissent to mutually agreeing on an experimental hypothesis and testing it. You’re learning based on agreed criteria.

That’s much harder to argue with — and at the end, there’s no “I told you so” or winners and losers. There’s just a “here’s what we learned” and an implied set of next steps. Bliss.

It's worth saying that this was the original intention when Jeff Bezos coined the phrase in his 2016 letter to Amazon shareholders:

Third, use the phrase “disagree and commit.” This phrase will save a lot of time. If you have conviction on a particular direction even though there’s no consensus, it’s helpful to say, “Look, I know we disagree on this but will you gamble with me on it? Disagree and commit?” By the time you’re at this point, no one can know the answer for sure, and you’ll probably get a quick yes.

This isn’t one way. If you’re the boss, you should do this too. I disagree and commit all the time. We recently greenlit a particular Amazon Studios original. I told the team my view: debatable whether it would be interesting enough, complicated to produce, the business terms aren’t that good, and we have lots of other opportunities. They had a completely different opinion and wanted to go ahead. I wrote back right away with “I disagree and commit and hope it becomes the most watched thing we’ve ever made.” Consider how much slower this decision cycle would have been if the team had actually had to convince me rather than simply get my commitment.

The perception among many people is that it's morphed into a kind of corporate authoritarianism. So in some ways, "disagree and let's see" is just a course correction back to its original intent.

imperfect notes & my second subconscious

I’ve never been a successful notetaker. Winnie Lim enumerates the many reasons why not, which seem to be very close to her thinking too:

“Because of my personality I tend to solve for the whole before wanting to do something. For years I wanted to figure out how I could retrieve the notes in a meaningful manner before I committed to making them. If I cannot remember I had made the note, did the note really exist?”

The problem is that you end up trying to come up with a smart taxonomy of notes ahead of time — and that’s always bound to fail, at least for me. I’ve lost count of the number of times I’ve wiped my Obsidian vault clean because I didn’t like the structure or the maintenance of it all threatened to overtake any utility. Instead, as Winnie points out, the best thing to do is just write the note. It’s a bit like throwing the text into a big bucket, and that’s okay.

I hate to say it, but this might be a decent use case for some kind of personal LLM (ideally on-device so I’m not sharing my private notes with a third party I don’t trust). If you’re constantly just making notes without structure, being able to ask something about their content feels like it would have a lot of utility — again, at least for me. I’d love to be able to have my notes about a certain topic summarized when I need them. Or even have the summary proactively come up for me depending on my context.

Then again, maybe that doesn’t matter at all:

“My brain is constantly holding scattered bits of information so it is just better to offload them somewhere in one place. I think the main difference is I don’t see obsidian as my second brain, I see it as my second subconscious.”

I like that. Blogging is a little bit that for me, but blogging has an audience. There’s something useful in being the Harriet the Spy of your own life and putting words to things that otherwise might go unsaid. There’s poetry in it, too, which is very obvious from Winnie’s post.

I’ll give notetaking another try.