Notices by Chris Trottier (atomicpoet@atomicpoet.org), page 2

AI companies said we’d reach artificial general intelligence (AGI) by now.

That clearly hasn’t happened and we’re far away.

RE: https://atomicpoet.org/objects/32c419aa-6010-42b9-9880-479c55fd69ed

USA is blatantly working to break-up Canada.

Not only is the Trump administration in talks with Alberta separatists, they’re actively working to destabilize my country.

https://www.thedailybeast.com/trump-teams-secret-meetings-with-group-plotting-to-break-up-canada-exposed/

David Rosen, co-founder of SEGA, just died. He was 95 years old.

Wait. He wasn’t Japanese?

That’s right. SEGA started as an American company. The reason SEGA is capitalized is because the original name was Service Games. And the “Service” in that name refers to the American military—its first customer base.

SEGA remained largely American until the 1980s, when David Rosen—along with Japanese business partners—bought the company from its parent, Gulf+Western, which also owned Paramount Pictures.

This initiated one of the most innovative and creative periods in video game history. SEGA produced classics like Space Harrier, OutRun, Sonic the Hedgehog, Streets of Rage, and Virtua Fighter.

For nearly two decades, SEGA was the primary rival to Nintendo, separating itself through speed and attitude.

Along with Atari, it was one of the companies that defined my childhood.

R.I.P., David Rosen. May you enjoy that great arcade in the sky.

https://www.theguardian.com/games/2026/jan/05/sega-co-founder-david-rosen-dies

I don’t care if generative AI exists.

If it collapses, gets regulated out, or proves useless, fine. Oh well.

What I don’t like are contradictions, gatekeeping, and techno-puritanism pretending to be ethics.

The “is AI art or not” debate collapses the moment you watch how IP holders behave. They’re not suing AI companies into oblivion. Instead, they’re licensing, investing, and monetizing.

When remixing happens without permission, it’s theft. When remixing happens inside a system they profit from, it’s innovation. Same act. Different cheque. That already tells you everything you need to know.

That said, when it comes to the “is AI art” debate, the philosophical escape hatch—the one everyone leans on—is intent. Humans have it. AI doesn’t. Sounds clean—supposedly. Except, when humans use AI, they have intent.

Which is why the supposed difference between AI and every other tool based on derivative works—such as sampling, collaging, quoting, etc.—doesn’t survive contact.

Intent doesn’t live in tools. Samplers don’t have it. Cameras don’t. DAWs don’t. LLMs don’t. People do.

If someone is prompting, re-prompting, rejecting outputs, selecting results, and deploying them for a purpose, intent exists. You don’t accidentally iterate. Complaining that the machine does all the work while also complaining people are prompting wrong is just holding two incompatible ideas at once. Pick one.

When that contradiction shows up, the argument mutates. Now it’s about quality. Or effort. Or struggle. Or depth. Or audience response. Or how it feels wrong.

None of that matters. Bad art is still intentional. Lazy choices are still choices. Audience response is downstream. Art doesn’t exist because people like it. It exists because someone made it. If struggle or popularity were required, most art history wouldn’t qualify.

None of this is new for me. This is the same position I’ve held for years. Composition over technique. Exploration over purity. Feeling over formula.

Tools don’t matter. Credentials don’t matter. Approved processes don’t matter. That’s why purity tests set off alarms.

Once you strip them away, the debate stops being about art and starts being about power. One rule for individuals. Another for systems backed by capital. Funny how that keeps happening.

RE: https://atomicpoet.org/objects/1d743389-00c1-48fc-932d-9a35cddb3442

The Gunk was my favourite game I played in 2025.

I played 435 games this year. This was the one I spent the most time with. I enjoyed every minute.

Right now it sells for $0.74 USD. That’s a 97% discount. The game is 100% DRM-free.

It is absolutely worth it.

https://www.gog.com/en/game/the_gunk

My unpopular coffee opinion:

Robusta is not worse than Arabica.

If you treat Robusta like Arabica, it tastes worse. If you treat it like Robusta, it can be excellent.

The key is controlled development. Robusta benefits from a steadier roast curve and a medium-dark to dark finish. The key, though, is that it needs proper heat to push it through second crack without baking it.

Where Robusta shines is in slow but measured extractions: espresso, phin drips, and cold brew. Vietnamese phins run long compared to pour-overs. Cold water suppresses bitterness and brings out the chocolate-heavy profile that good Robusta is known for.

Handled with the right roast and the right method, Robusta can be remarkable. Good Robusta has hints of cocoa, nuts, and slight spice. But you won’t get there if you brew it like Arabica.

@jrcruciani Quantum computing is still way too early in its tech hype cycle. A lot of people are already banging the drum on it, but I haven’t seen a clear path for how it changes people’s lives in real, practical ways yet. It’s fascinating science, but not something I spend much time modeling because we’re nowhere near meaningful deployment.

Cisco is on the verge of doing something genuinely absurd.

It’s inches away from reclaiming its March 27, 2000 intraday all-time high of 82.00. This is the price level everyone swore would remain frozen in amber forever. Yet here we are.

Twenty-five years later, the chart is knocking on the same door it slammed into during the dot-bomb. If it breaks through, that’s a quarter century of history closing like a tab no one thought would ever get paid.

And the significance is massive.

Cisco was no mere casualty of the dot-com bubble. It was the mascot. The crowned king. At the peak, it briefly became the most valuable company on Earth.

Then gravity kicked in, the bottom fell out, and Cisco lost more than 80% of its value in record time. For years, analysts pointed to Cisco as the archetype of speculative mania. It became shorthand for “brilliant idea, terrible valuation.”

Now, right in the middle of a brand-new hype cycle—the AI gold rush—Cisco is quietly about to remind us all of that ancient trauma.

However, the conditions today are nothing like 2000.

Cisco is no longer the poster boy for speculation. Its forward P/E is roughly 18.72, its PEG is 1.82, and it’s been paying reliable dividends since 2011. You don’t get a 2.16% yield from a company powered by hype. You get that from a slow, heavy, cash-generating machine that investors treat like infrastructure—because that’s what it builds.

Which means we can’t say Cisco’s return to its dot-com high is necessarily a bubble signal. Perhaps it’s fundamentals finally catching up to a price investors placed on it a generation ago.

Thing is, the dot-bomb valuation wasn’t ridiculous because Cisco’s future never materialized. It did materialize. The internet became the foundation of the entire global economy. Everything runs on networks now. Cisco’s hardware is the plumbing of modern civilization.

What investors got wrong wasn’t the future. It was the speed.

The market priced in 25 years of growth and adoption as if the whole thing would happen in 5. The vision was right. The timeline was fantasy.

And that’s why this moment is historic. Is Cisco the canary in the coal mine for yet another AI bubble? Or is it proof that sometimes the world really does move in the direction everyone predicted—but not necessarily on schedule?

If Cisco finally clears 82.00, I don’t think it will be a warning. It will be closure. A full cycle completed.

And the final proof that the dot-com era wasn’t wrong about the internet. It was wrong about how long it would take to turn prophecy into cash flow.

People ask why I work in finance.

Most of the time, it’s boring. Truly boring. Markets drift. People get in their feels. Nothing happens except sentiment swinging from euphoria to despair and back again.

But when it isn’t boring? It’s the greatest show on Earth.

We just spent two straight weeks screaming about an AI bubble. Every headline was doom. Every pundit was prophesying collapse. You’d think Western civilization was getting priced for liquidation.

Meanwhile, my position never changed: whether or not we’re in a bubble, sentiment didn’t match the fundamentals. On a macro level, the Fear & Greed Index was sitting at extreme fear. Volatility was spiking. Tech sold off hard. But at the company level? The numbers weren’t cracking. Not at Microsoft. Not at Google. And certainly not at Nvidia.

For saying that, I got called an apologist more than once. As if noticing a disconnect between emotion and data is somehow ideological. It isn’t. It’s my job. I don’t declare boom or gloom. I look for when the narrative and the numbers stop lining up.

And sure enough—here we are. Nvidia beats expectations again. EPS $1.30 vs $1.26. Revenue $57B vs $55.2B expected. Data center revenue alone at $51B versus the $49B consensus. Then they guide Q4 revenue to $65B when Wall Street was bracing for $62B.

That’s not a miss. That’s a statement.

I don’t know where the market goes next. Nobody does. But I can tell you this much: the world didn’t end. Western civilization continues. The apocalypse has been postponed.

The Metaverse doesn’t need Meta. It needs the Fediverse.

Meta has spent more than US$70B on VR/AR since 2020. Reality Labs lost US$17.7B in 2024 alone, and its cumulative losses are now over US$70B according to earnings reports. That investment only makes sense if Quest becomes a mass-market computing platform, not a niche headset. Meta has openly said that VR and AR are the next major platform after the smartphone. But Quest never hit the ubiquity that turns a device into generational tech.

The iPhone became essential because people used it constantly and its app ecosystem became woven into daily life. VR hasn’t made that leap. Most people don’t live in a headset. There’s no universally indispensable “iPhone moment” app. And even if Meta built the most polished XR experience imaginable, polish has never guaranteed inevitability. If it did, Google+ would have conquered social networking.

And while Quest hardware has improved, sales have fallen year-over-year and Reality Labs revenue remains under 1% of Meta’s total revenue. Those numbers don’t point to a platform about to reshape the world. They point to a very expensive stall and an industry precariously balanced on one company’s willingness to keep losing billions.

That’s where the real danger begins. Corporate-owned social platforms do not fail gracefully. MySpace, AIM, Google+ all vanished when the company walked away. XR raises the stakes even higher. A closed social network loses posts. A closed Metaverse loses entire worlds. The claim that “VR will stagnate without Meta” isn’t a defense of Meta. It’s an indictment of an architecture where one corporate balance sheet can freeze an entire medium.

This is why resilience matters. If you are building VR worlds, tools, communities, experiences, you need to know your work won’t evaporate because a CFO adjusted the five-year plan. You need identities that persist. Worlds that persist. A social layer that outlives corporate strategy. Convenience absolutely matters, but it isn’t real convenience if the universe you built can be unplugged overnight.

Open systems provide exactly that kind of continuity. Email never died. RSS never died. XMPP never died. These open protocols survive because no single corporation controls them—and the Fediverse operates on the same principle. That matters for XR, because any social layer meant to outlive a company’s budget cycle needs infrastructure that doesn’t disappear when the funding does.

But continuity is only half the story. The other half is scale. People often talk about the Fediverse as if it’s “Mastodon and its ~10M users,” when Mastodon is just one cluster of thousands of ActivityPub-speaking servers. And the real footprint extends far beyond it. WordPress—which powers roughly 43% of the entire web—is Fediverse-capable. Ghost is Fediverse-capable. And Threads, now with more than 400 million monthly active users, is actively integrating ActivityPub. The potential addressable market for a federated XR ecosystem isn’t 10M. It’s already in the billions—and waiting for the right applications to tap into it.

And the missing XR layer is already here: ActivityPub. Years ago, Immers Space experimented with bridging WebXR to the Fediverse. It didn’t scale then—not because it was impossible, but because everything around it was early. WebXR was immature. XR adoption was small. The Fediverse was tiny compared to what it is now. The idea was right. The ecosystem wasn’t ready.

Now the timing finally lines up. Valve’s upcoming Steam Frame is PC-first, Linux-powered, OpenXR-native, and built by a company that treats openness as a strength rather than a liability. Steam Frame isn’t a guaranteed savior. What it does create is space. Space for XR that isn’t welded to a single company’s garden. Space for developers, modders, and open-source communities to build worlds that don’t vanish when one corporation changes direction.

I see the same pattern running GoFedi, a managed Fediverse hosting service. Communities thrive when their digital lives aren’t tethered to one company’s survival. Apply that logic to XR and you get a Metaverse that behaves like the internet itself: resilient, interoperable, and impossible to kill with a budget meeting. Not instead of good hardware, but alongside it—as the continuity layer that keeps progress from evaporating the moment a CEO decides it’s time to pivot.

The future of shared worlds shouldn’t hinge on whether one corporation can absorb another US$15B loss. It should belong to the people running their own servers, building their own spaces, and refusing to let entire universes disappear because a company changed its mind.

Valve isn’t just the biggest force in PC gaming, and they’re not just the newest console manufacturer swaggering into the arena.

They’re morphing into something far bolder: the Apple of Linux.

If you’re not a gamer, that might sound unhinged. Maybe even a little deranged. But if you’re already deep in the Steam ecosystem—if your library scrolls so far it needs its own municipal transit system—you know this isn’t wild at all. It’s practically destiny.

Let’s rewind. When Steve Jobs returned to Apple, he didn’t reinvent the wheel. He just drew a big cross on a whiteboard and said: four products. iMac, Power Mac, iBook, PowerBook. Four neat squares. Four clean market segments. And everything Apple built slotted neatly into that grid.

Apple didn’t suddenly leap to 30% marketshare. They barely scraped 3%. Didn’t matter. Because the money wasn’t really in the hardware. It was in the ecosystem.

Buy a Mac and suddenly you’re buying OS upgrades, iLife apps, office software, music tools, the whole glittering Cupertino starter kit. That stack of software made the hardware profitable, and that hardware made the software inevitable. The loop fed itself.

Now fast-forward to Valve. Look at what they’ve assembled.

Four core hardware pillars:

• Steam Controller
• Steam Deck
• Steam Machine
• Steam Frame

Four segments. Four use cases. Four doors into the same house.

Already have a PC? You grab the Steam Controller.

Want your library in your backpack? Steam Deck.

Want it in the living room? Steam Machine.

Want it strapped to your face? Steam Frame.

And the moment you buy any one of these, something interesting happens: the rest of the ecosystem starts making sense. Buy a game on Steam and it works everywhere. Your save files carry across devices. You can stream titles between them. The more hardware you add, the smoother it all feels, and the more the ecosystem pulls you deeper in.

But here’s the part I really want you to notice: I didn’t say Valve wants to be the Apple of gaming. No. They want to be the Apple of Linux.

And that’s where this gets concrete. Their hardware ships with Linux that isn’t locked down or lobotomized. It has a real desktop environment hiding under a slick UI.

Which means Valve can evolve SteamOS in ways Apple never aimed to with macOS. Apple built a general-purpose OS that occasionally supported games. However, Valve built a gaming OS that can naturally branch outward into media, creative tools, and productivity. “Gaming-adjacent” doesn’t require a conceptual pivot. It’s the next logical step.

What might that look like?

• A native media center built directly into SteamOS—think Plex or Jellyfin, but officially blessed and seamlessly integrated.
• First-party creative tools that take advantage of Proton and GPU acceleration—video editors, music tools, asset creators.
• A productivity layer—file syncing, cloud storage, collaborative apps—that piggybacks on your Steam identity.
• A SteamOS app store that isn’t just for games. Apps, utilities, editors, streaming clients, the works.

They’ve already dipped into this with Big Picture Mode’s media features, Steam Link, Steam Input configurators, desktop mode on Steam Deck, and Proton opening the gates for thousands of non-gaming applications. Nothing stops them from extending that further.

That’s why Valve—private, secretive, and small enough to fit inside an Amazon lunchroom—is still one of the most valuable forces in the entire industry. Not because they sell hardware like Apple, but because they’re building an ecosystem like Apple. Except this one runs on Linux.

If you’re a PC gamer, none of this is news. But if you’re outside the gaming bubble and this future arrives exactly how I’ve described, just know: it didn’t come out of nowhere. You just weren’t looking in Valve’s direction.