This is a letter from the CEO of ETAPX about the future we are building — a connected ecosystem where artificial intelligence amplifies human creativity instead of replacing it. Across Whistlr, GLSRM, and Ocsidian, we are making one bet: that the next era of software belongs to companies obsessed with craft, built in-house, and answerable first to the people who use them.
I want to write to you the way I would talk to you if we were sitting across a table — not in the language of a press release, and not in the language of a pitch deck. I have read a lot of those letters over the years. Most of them say almost nothing in a great many words. I am going to try to do the opposite. I want to tell you exactly why ETAPX exists, what we believe, what we are actually building, and what we are asking of you if you choose to come along.
ETAPX is a technology company. But that description undersells what we are trying to do. We are building three products that share one spine: Whistlr, the social platform; GLSRM, the front page of AI; and Ocsidian, the agentic game creation platform. They look different on the surface. One is where you talk to people. One is where you keep up with a field that reshapes itself every week. One is where you build worlds. Underneath, they are the same company making the same argument over and over again: that the tools you use every day can be better, faster, more honest, and more respectful of you than the ones the industry handed you.
Why ETAPX Exists
Every company has an origin story, and most of them are tidied up after the fact. Ours is simple and a little stubborn. We started ETAPX because we were tired of using software that treated us like a resource to be harvested. Tired of feeds engineered to keep us anxious. Tired of "creator tools" that took more than they gave. Tired of AI products that were thin wrappers over someone else's model, shipped fast and abandoned faster. Tired, frankly, of the sense that the people building the tools had stopped caring whether the tools were good.
The conviction underneath ETAPX is that the relationship between a person and the software they rely on should be a fair one. You give a product your attention, your data, your trust, and your time. In return, it should give you leverage — real leverage to connect, to learn, to create, to earn. When that exchange is fair, technology feels like a gift. When it is rigged, it feels like a slow extraction. Most of the software economy of the last fifteen years was built on the rigged version. We think the next fifteen years will be built on the fair one, and we intend to be one of the companies that proves it can win.
So ETAPX exists to build that fairer relationship across the three domains where we think it matters most right now: how people connect, how people understand the most important technological shift of our lifetimes, and how people make things. That is not a coincidence of three random products. It is a deliberate bet on three frontiers that are all being remade at the same time by the same force.
"We are not building features. We are building a relationship with the people who use what we make. Every decision either strengthens that relationship or strips it for parts. We have decided, permanently, to strengthen it."
— CEO of ETAPX
The Force Reshaping Everything
That force is artificial intelligence, and I want to be precise about how we think about it, because the conversation around AI has become so loud that it is hard to hear anything true inside it.
There are two unserious positions on AI, and both are everywhere. The first is that AI is a magic trick that will soon do everything, so human skill no longer matters and we should all step aside. The second is that AI is a fad, a plagiarism machine, a bubble, and that serious people should ignore it. Both positions are comfortable, and both are wrong, and we have built ETAPX on the harder ground in between.
Here is what we actually believe. AI is the most powerful amplifier of human intent that has ever been invented. It is not a replacement for taste, judgment, or craft — it is a multiplier on top of them. Point it at a person who has nothing to say, and you get faster noise. Point it at a person with vision, taste, and something they care about, and you get work that would have taken a studio a year, done in an afternoon, without losing the fingerprint of the human who made it. The interesting question of this decade is not "what can the model do?" It is "what can a person with real intent do when the model is finally good enough to keep up with them?"
That is the question every ETAPX product is built to answer. Whistlr asks it about connection. GLSRM asks it about understanding. Ocsidian asks it about creation. The answer in all three cases is the same: the human stays in the driver's seat, the AI does the work that used to stand between intent and result, and the floor under what an individual person can accomplish rises by an order of magnitude.
Why Now
People ask why we are doing this now, as if timing a company is like timing a market. It is not. You do not get to choose the year you are born into, and neither does a company. You get the moment you get, and the only real question is whether you understand it.
The moment we are in is rare. Three things have become true at the same time, and they may not stay true together for long. First, the underlying AI models crossed a threshold from "interesting demo" to "genuinely useful collaborator" — they can now hold context, reason across steps, and act with enough reliability to build real workflows on top of them. Second, the cost and latency of using them dropped far enough that you can put them inside everyday products instead of behind a paywall for specialists. Third, an entire generation of people has grown up expecting software to be fast, beautiful, and personal, and they have run out of patience for the bloated, ad-choked, dark-pattern products they inherited.
When the capability, the economics, and the audience all arrive at once, you get a window. Windows like this open maybe once a decade. The companies that define an era are almost always the ones that recognized the window and committed fully while everyone else was still debating whether it was real. We recognized it. We committed. That is the entire reason "now" matters.
- The models grew up: AI moved from novelty to dependable collaborator, reliable enough to build durable products on, not just demos.
- The economics flipped: Intelligence got cheap and fast enough to live inside everyday tools rather than behind specialist paywalls.
- The audience changed: A generation raised on instant, beautiful software has lost patience with extractive, slow, ad-heavy incumbents.
- The incumbents got comfortable: The biggest players are defending business models, not reinventing them — which is exactly when a focused newcomer can move.
The Bet: AI Plus Human Craft
The single sentence that explains ETAPX better than any other is this: we are betting everything on the combination of artificial intelligence with human creativity and craft. Not AI alone. Not craft alone. The combination, treated as inseparable.
It is worth slowing down on why we phrase it that way, because the word "craft" is doing real work. Craft is the part of building that cannot be automated and should not be: the thousand small decisions about how a thing should feel, the taste to know when something is almost right but not yet right, the refusal to ship something mediocre because it is technically finished. Craft is what separates a product you tolerate from one you love. The industry largely abandoned craft over the last decade because craft is slow and expensive and does not show up cleanly in a growth metric. We think that was a historic mistake, and we think AI is what finally makes craft affordable again at scale.
Here is the logic. AI handles the volume — the rote, the repetitive, the parts of building that used to consume the hours that should have gone into judgment. That frees the human to spend their time on the part only a human can do: deciding what is good. When you take a small team of people who genuinely care about quality and give each of them the throughput of a department, you do not get a worse product faster. You get a better product, made by people who finally have time to make it better. That is the whole thesis. AI buys back the time; craft spends it well.
"AI does not lower the bar for what we ship. It raises it. When the machine handles the volume, the only thing left to compete on is taste — and taste is exactly what we refuse to outsource."
— CEO of ETAPX
Built In-House, On Purpose
One of the least glamorous and most important decisions we made early was to build everything ourselves. ETAPX is a build-in-house company in an era that worships the opposite — an era where the default advice is to assemble a product out of other people's services, wrap them in a thin interface, and rush to market.
We understand the appeal. Building things yourself is slower at the start. It is harder. It requires people who can actually engineer, not just integrate. But the moment you depend on someone else's stack for the core of your product, you have handed them control over your quality, your roadmap, your costs, and ultimately your relationship with your users. When their service degrades, your product degrades. When they change their pricing, your business shakes. When they decide your category competes with theirs, you discover that you were a feature, not a company.
You can see this conviction most clearly in Ocsidian, where we built our own engine from the ground up. The BlackFrost Engine is a proprietary runtime and editor core — a full 3D scene composition pipeline with physically based rendering, environment lighting, shadows, and post-processing like screen-space reflections, ambient occlusion, temporal anti-aliasing, and volumetric light scattering. It runs integrated Havok physics natively, both in the editor and in exported projects. It has a TypeScript-first scripting system with hot reload, a complete asset pipeline with compressed texture caching, a timeline-based animation and cinematic editor, navmesh generation for pathfinding, and visual node editors for materials, particles, geometry, render graphs, and flow logic. That is not something you assemble out of a weekend's worth of integrations. It is a serious piece of engineering, built because we refuse to put the creative heart of a product in someone else's hands.
Building in-house is a tax we pay up front and a moat we collect on forever. It is the reason we can promise quality and actually keep the promise. You cannot guarantee craft on top of a foundation you do not control.
One Account, One Ecosystem
The three products are connected by design, and the connection starts with something deceptively simple: one account follows you across the entire ecosystem. Sign in once, and your identity, your preferences, and your place in the community travel with you from the social platform to the AI front page to the game creation tools. You are not three separate users to us. You are one person, and we treat your presence as continuous.
This matters more than it sounds. The old model fragments you. Every product wants its own login, its own profile, its own walled copy of you, because fragmentation is how platforms lock you in. We went the other way on purpose. A single connected account is a statement that we would rather earn your loyalty than trap it. It also unlocks real value: what you discover on GLSRM can spark a conversation on Whistlr; the community you find on Whistlr can follow your work in Ocsidian; the things you build in Ocsidian have a natural place to be shared and discussed. The ecosystem is not three products bolted together. It is one continuous relationship expressed in three rooms.
- Whistlr: The social layer — where people connect, talk, and build community without being farmed for engagement.
- GLSRM: The front page of AI — where the industry's news, models, releases, benchmarks, and research are tracked on time, all the time.
- Ocsidian: The agentic game creation platform — where an idea becomes a playable world with speed, taste, and full creative ownership.
GLSRM: The Front Page of AI
Let me take you through each product properly, starting with GLSRM, because it is the clearest window into how fast our world is moving.
The problem GLSRM solves is simple to state and brutal to live with: the AI field now moves faster than any human can track. New models ship weekly. Benchmarks shift. Labs leapfrog each other. Research that would have been a landmark a few years ago now scrolls past in a day. If your job, your business, or your curiosity depends on understanding artificial intelligence, you are drinking from a firehose, and most of what reaches you is either marketing or noise. GLSRM exists to fix that. It is an intelligence platform for the AI industry — a single, continuously updated, editorially clean surface that tracks the field as it actually moves.
What GLSRM Is Made Of
GLSRM is organized the way a serious person would organize this information if they had unlimited time, which is exactly the time they do not have.
- The Wire: A live newsfeed across the categories that define the industry — News, Models, Releases, Research, and Agents. Curated headlines, refreshed around the clock, built to be the first thing you check and the fastest way to stay current.
- Data Pulse: The industry at a glance, in numbers. Hundreds of models and dozens of labs tracked daily — top intelligence, Elo standings, best coding and agentic performance, the fastest and lowest-latency models, the biggest context windows, the sharpest prices, and the newest releases, all in one panel.
- Releases: A cinematic walkthrough of the frontier. Every major release placed in context — capabilities, positioning, and the data behind the hype — presented as an immersive panorama rather than a dry changelog.
- Benchmarks: Standings you can act on. Intelligence, coding, agentic performance, speed, and cost, tracked over time rather than frozen at launch.
- Watch: A curated video wall of launch events, deep dives, interviews, and breakdowns, organized for the way people actually catch up.
- Circuits: The community layer — threaded discussions in focused SubCircuits, from frontier-lab watchers to vibe-coding practitioners, with profiles, messaging, and notifications built in.
The thing I am proudest of in GLSRM is its restraint. It would have been easy to build another hype machine, another feed that rewards whatever is loudest. We built the opposite: an editorially clean surface that respects your intelligence and your time. GLSRM treats benchmarks as things to be tracked over time, not screenshots to be cherry-picked at launch. It places releases in context instead of amplifying the spin. That editorial discipline is craft, applied to information. It is the difference between a product that makes you feel informed and one that actually informs you.
Why We Built GLSRM the Way We Did
Every design decision in GLSRM traces back to one observation: in a field moving this fast, the scarce resource is not information — it is trust in the information. There is no shortage of takes about AI. There is a desperate shortage of a place that tells you what is real, what matters, and how it fits together, without an agenda. So we made choices that a purely growth-driven product would never make.
We separated the categories — News, Models, Releases, Research, Agents — instead of blending everything into one undifferentiated stream, because the person tracking research has a different job than the person tracking releases, and a good tool respects the difference. We built Data Pulse to show the industry in hard numbers because opinions are cheap and standings are not; when you can see top intelligence, Elo rankings, coding and agentic performance, latency, context windows, and pricing in one panel, you stop arguing about vibes and start reasoning about evidence. We made Releases cinematic rather than a changelog because a frontier model launch is a genuine event in the history of the field, and treating it like a list item would be a failure of proportion. And we wrapped a real community layer — Circuits, with focused SubCircuits, profiles, and messaging — around all of it, because understanding a fast field is a team sport, and the conversation is half the value.
None of these are obvious or free decisions. Each one made the product harder to build and slower to ship. We made them anyway, because GLSRM is not trying to be the biggest feed about AI. It is trying to be the one you trust. That is a smaller goal that turns out to be much harder.
Ocsidian: From Idea to Playable World
Ocsidian is the boldest expression of the ETAPX thesis, because game creation is the hardest creative discipline there is, and it is exactly where AI plus human craft has the most to give.
Building a game has always required you to be ten people at once: a designer, a 3D artist, a programmer, an animator, a physics engineer, a sound designer, a level builder, a technical director, a QA tester, and a producer. The tooling was sprawling, the learning curve was cruel, and the result was that the vast majority of people who had a great idea for a world never got to build it. The gap between imagination and execution was simply too wide. Ocsidian is built to close that gap.
A Vertically Integrated Creation Platform
Ocsidian is vertically integrated, which is a precise way of saying it does not ask you to assemble your toolchain out of a dozen disconnected programs. It brings together a native desktop editor, a real-time 3D engine, AI-powered agentic workflows, project scaffolding, cloud account management, and a modern web presence into a single cohesive product. The platform handles the complexity — scene composition, physics, rendering, scripting, asset management, live preview, and deployment — so the creator can focus on the only thing that actually matters: the experience they want to build.
Ocsidian Os, the desktop editor, ships as a native application on macOS and Windows. It opens to a branded home screen with your recent projects and account sign-in. You can scaffold a new project from real templates — Next.js, Electron, Vanilla TypeScript, Solid.js, or Nuxt.js starters — each pre-configured with the engine packages, Havok physics, and editor tooling so you are building, not configuring. One click launches a live preview in play mode against a local dev server. Authentication persists across windows and restarts, with a secure handoff that connects the web sign-in flow straight to the desktop editor. There is even self-healing project repair that quietly patches import paths, engine links, and physics binaries so your project keeps working instead of breaking under you.
What "Agentic" Actually Means Here
The word "agentic" gets thrown around carelessly, so let me be concrete about what it means inside Ocsidian. It means AI that does not just suggest — it acts inside your project, under your direction, to move you from an idea toward a playable world. It is the embodiment of the whole ETAPX bet: the agent handles the laborious construction, and you keep full creative control. The tagline we hold ourselves to is "from idea to playable world — with speed, taste, and full creative control." Every word in that line is a constraint we accept. Speed, because the gap between imagination and result should be small. Taste, because the output has to be good, not just generated. Full creative control, because the work is yours — Ocsidian is built for solo creators, small studios, and teams who want to build without giving up ownership of what they make.
"The dream is not a machine that makes games for you. The dream is a machine that finally lets you make the game you always saw in your head — and gets out of the way the moment you take the wheel."
— CEO of ETAPX
The Engineering Behind the Promise
I want to dwell on the engineering for a moment, because the promise of "speed, taste, and full creative control" is only as good as the machinery underneath it, and a lot of products in this space make that promise on top of foundations that cannot keep it.
The BlackFrost Engine is what lets Ocsidian keep its word. A real-time scene graph and rendering pipeline with PBR materials, environment lighting, shadows, image-based lighting, and a deep post-processing stack means the worlds you build actually look like worlds, not prototypes. Integrated Havok physics running natively in both the editor and exported projects means things behave the way your instincts expect — objects fall, collide, and respond without you wiring up a separate physics layer. A TypeScript-first scripting system with decorators, hot reload, and editor-aware metadata means programmers feel at home and changes appear instantly instead of forcing a rebuild. The asset pipeline accepts 3D models, textures, HDR environments, materials, sounds, particles, sprites, and GUI layouts by drag and drop, with compressed texture caching and KTX2 transcoding so performance does not collapse as your project grows. Timeline-based animation with curve and keyframe editing, cinematic sequencing, and render-to-video export means you can direct, not just assemble. Navmesh generation with Recast integration gives your characters the ability to navigate. And the visual node editors — for materials, particles, geometry, render graphs, and flow logic — give creators who do not write code a real path to authorship rather than a toy.
I list all of this not to brag about a feature matrix, but to make a point about the relationship between depth and freedom. The reason a creator can move fast and stay in control inside Ocsidian is precisely that we did the deep, slow, unglamorous engineering underneath. The agentic workflows can be helpful only because there is a serious engine for them to act inside. Convenience without depth is a demo. Depth plus convenience is a tool you can build a career on. We are building the second one.
Whistlr: Connection Without Extraction
Whistlr is the social platform, and it is where the human side of our thesis is most personal. Social media is the technology that touched the most lives and, in many ways, betrayed them the most thoroughly. It promised connection and delivered comparison. It promised community and delivered metrics. It promised to bring us closer and instead got very good at keeping us scrolling, anxious, and alone together.
Whistlr is our attempt to build the version that keeps the promise. The design principle is connection without extraction: a place to be with the people you actually care about, where the product's incentives are aligned with your well-being instead of opposed to it. Inside the ecosystem, Whistlr is also the community fabric — the same account that connects you across GLSRM and Ocsidian connects you to the people you talk to and build with. The community layer is not an afterthought bolted onto the products; it is the connective tissue that makes an ecosystem feel like a place rather than a collection of tools.
The Old Way Versus What We're Building
It helps to name the thing we are reacting against, because our whole approach is a deliberate inversion of it. The old way of building consumer technology had a recognizable shape, and we have set out to do the opposite of it at every turn.
- The old way optimized for attention; we optimize for outcomes. The previous era measured success in time-on-app and impressions. We measure it in whether you connected, understood, or created something real. If you close the app satisfied, that is a win, not a loss.
- The old way assembled; we build. Where the industry wraps other people's services, we build our own foundations — our own engine, our own surfaces — so we own our quality.
- The old way fragmented you; we keep you whole. Instead of a separate trapped profile in every product, one account follows you across the ecosystem as a single continuous person.
- The old way treated craft as overhead; we treat it as the point. Quality was the first thing cut under deadline pressure. For us it is the non-negotiable, made affordable again by AI handling the volume.
- The old way feared AI or worshipped it; we put it to work. We neither pretend AI changes nothing nor pretend it changes everything. We aim it carefully at the work that stands between human intent and human result.
What This Means For You
All of this philosophy is worth nothing if it does not change your actual experience, so let me be direct about what it means depending on who you are.
If You're a User
It means software that is on your side. A social platform that wants you to feel good and connected rather than anxious and addicted. An AI front page that informs you instead of overwhelming you. One account that respects you as a single person rather than slicing you into profiles. We are asking for your trust, and we understand that trust is earned in specifics, not slogans. The specific we offer is this: every product decision we make is weighed first against whether it serves you, and we have structured the company so that nobody's incentive is to betray that.
If You're a Creator
It means leverage. The whole point of combining AI with craft is to raise the ceiling on what one talented person can make. With Ocsidian, the idea you could never afford to build becomes buildable. With GLSRM, the field you need to track becomes trackable. With Whistlr, the community you need to reach becomes reachable. And critically, you keep ownership of your work — Ocsidian is built around the principle that you do not surrender creative control to use powerful tools. We win when you win, and we have tried to wire that into the products rather than just say it.
If You're a Builder
It means an invitation. If you are an engineer, a designer, a researcher, or anyone who has ever been frustrated by getting handed a watered-down version of what you knew you could make — ETAPX is built for the kind of person who would rather build the hard, in-house, craft-driven thing than the easy, assembled, disposable one. We are betting that there are a lot of you, and that you are tired of being told that quality does not scale. We are building proof that it does.
The Hard Parts We're Not Hiding
I would not respect you if I only told you the good parts. Building this way is hard, and there are real risks, and you deserve to hear them from me directly.
Building in-house is slower than assembling, and slowness is dangerous in a fast market. We have to be excellent at engineering, not just integration, which means we have to attract and keep genuinely great people. The AI field moves so fast that any specific capability we describe today may be table stakes by the time you read this — which is exactly why we anchor on craft and judgment, the things that do not expire, rather than on any single model or feature. And building three connected products instead of one focused product is a discipline test: the ecosystem is our strength, but only if each product is genuinely excellent on its own. A weak link does not just fail by itself; it weakens the whole.
We are clear-eyed about all of this. We chose the harder path deliberately, because we think the harder path is the only one that leads to something durable. Anyone can ship the easy version. We are not interested in the easy version.
Frequently Asked Questions
What is ETAPX?
ETAPX is a user-first, craft-driven technology company that builds three connected products in one ecosystem: Whistlr, the social platform; GLSRM, the front page of AI; and Ocsidian, the agentic game creation platform. Everything is built in-house, and one account connects a user across all three.
How does ETAPX use AI?
We treat AI as an amplifier of human creativity and craft, not a replacement for it. In every product, AI handles the volume and the rote work that used to stand between a person's intent and their result, while the human stays in control of judgment, taste, and final decisions. The bet is that AI plus human craft beats either one alone.
What are Whistlr, GLSRM, and Ocsidian?
Whistlr is a social platform built for genuine connection without extractive engagement tactics. GLSRM is an intelligence platform that tracks the AI industry — news, models, releases, benchmarks, research, and agents — on a single, editorially clean, continuously updated surface. Ocsidian is a vertically integrated, agentic game creation platform, powered by the in-house BlackFrost Engine, that takes a creator from idea to playable world while keeping full creative ownership.
Why does ETAPX build everything in-house?
Because owning your foundation is the only way to guarantee your quality. When you depend on someone else's stack for the core of your product, you hand them control over your roadmap, your costs, and your relationship with your users. Building in-house — like the proprietary BlackFrost Engine inside Ocsidian — is a cost up front and a durable advantage forever.
What does "one account, one ecosystem" mean?
It means a single account follows you across Whistlr, GLSRM, and Ocsidian. Your identity, preferences, and place in the community travel with you instead of being fragmented into separate, trapped profiles. It unlocks real value across products and reflects our choice to earn loyalty rather than lock it in.
Is ETAPX a social media company, an AI company, or a gaming company?
It is all three, on purpose — and the connection is the point. Social, AI intelligence, and game creation are three frontiers being remade at once by the same force, and ETAPX is built to be excellent across all three while keeping them connected by a single account and a single philosophy of craft.
What does ETAPX mean by "craft"?
Craft is the part of building that cannot be automated and should not be — the taste to know when something is good, the thousand small decisions about how a thing should feel, and the refusal to ship mediocrity just because it is technically finished. We believe AI finally makes craft affordable at scale, because it buys back the time that craft requires.
How can I get involved with ETAPX?
If you are a user, use the products and tell us where they fall short — every decision is weighed against whether it serves you. If you are a creator, build with them and keep ownership of what you make. If you are a builder who would rather make the hard, craft-driven thing than the disposable one, ETAPX is built for you.
The Future We're Building
I want to close with the part I think about most, which is not this year or next year but the decade ahead.
I believe we are living through one of those rare hinge moments when the floor under human capability rises permanently. Not for a few specialists, but for everyone. The person with a story but no studio will tell it. The person with an idea for a world but no team will build it. The person trying to understand a field that reinvents itself weekly will keep up. The communities that the old internet pulled apart will have somewhere honest to reassemble. This is not a fantasy. The pieces exist now, today, and the only question is whether the companies that assemble them will do it with respect for the people they serve, or with the same extractive habits that defined the last era.
ETAPX is our answer to that question. We are building the version where the technology is on your side. Where AI is an amplifier of your creativity, not a replacement for it. Where craft is the point and not the overhead. Where the relationship between you and the software you depend on is, finally, a fair one. Whistlr, GLSRM, and Ocsidian are the first three rooms in a house we intend to keep building for a very long time.
To the users who give us their trust, the creators who give us their work, and the builders who might give us their talent: thank you for reading this far. We are early, we are ambitious, and we are not going to get everything right. But we know exactly who we are building for, and exactly why, and we are not going to compromise on either. The future is going to be built by someone. We intend to make sure it is built well.
"We are building the future where technology is on your side. That is the whole company, in one sentence. Everything else is just the work of keeping that promise."
— CEO of ETAPX
