News

The Next Heinrich Problem Is Not More Demos
The Next Heinrich Problem Is Not More Demos EMPHOS Group, June 27, 2026, 7 min read A demo can be useful. A proof can be useful. A benchmark can be useful. Heinrich has produced many of them, and they matter because they show that individual capability patterns are possible. But a proof is not a product. The next Heinrich problem is not more demos. The next Heinrich problem is moving the proven ideas into the ordinary path a user actually touches: ask a question, activate COEUS, form support, let Mercury... Read more...
Learning Is Not the Same Thing as Mutating the Mind
Learning Is Not the Same Thing as Mutating the Mind EMPHOS Group, June 27, 2026, 7 min read There is a version of AI learning that sounds impressive and should make everyone nervous. The system answers. The user reacts. The system changes itself. The next answer is different. The loop is fast, autonomous, and opaque. The marketing calls it continuous improvement. The user is asked to trust that the machine knows which experiences should become truth, which corrections are malicious, which memories are outdated, and which changes should be rolled... Read more...
The Most Important Answer Heinrich Gives Is I Do Not Know
The Most Important Answer Heinrich Gives Is I Do Not Know EMPHOS Group, June 27, 2026, 6 min read The easiest way to make an AI product look impressive is to make it answer everything. The screen fills. The paragraphs sound coherent. The user sees confidence. The demo moves quickly. Nobody stops to ask whether the system had support for the answer, because the answer already arrived with the shape of authority. That is exactly the behavior Heinrich is designed to avoid. The most important answer Heinrich gives is not... Read more...
The Intelligence Has to Fit on the Machine
The Intelligence Has to Fit on the Machine EMPHOS Group, June 27, 2026, 6 min read There is a simple question every AI company eventually has to answer: where does the intelligence live? For most of the industry, the answer is somewhere else. A data center. A rented model endpoint. A cluster of accelerators behind an API. The product on your computer is an interface. The intelligence is remote, metered, and controlled by someone else's infrastructure. Heinrich is built in the other direction. The intelligence has to fit on the... Read more...
Heinrich Is Not One System. It Is Three Systems Working Together
Heinrich Is Not One System. It Is Three Systems Working Together EMPHOS Group, June 27, 2026, 6 min read Most artificial intelligence products are presented as a single surface. You type into a box. Text comes back. The marketing calls that box an assistant, a copilot, an agent, or a model. The user is not supposed to ask where the answer came from, what part of the system reasoned, what part spoke, what part acted, or what part checked whether the answer was supported. Heinrich is not built that way.... Read more...
Heinrich Cannot Hallucinate. And That Is the Whole Point
Heinrich Cannot Hallucinate. And That Is the Whole Point EMPHOS Group, May 30, 2026, 7 min read Every large language model deployed today shares a single structural property. It cannot tell you what it knows. It can only tell you what it produces. When the output is correct, that is a coincidence between the model's statistical patterns and reality. When the output is wrong, the model has no way of recognising that. From the outside, the two cases look identical. They have to be verified by something else. Usually a... Read more...
The Man the System Is Named For
The Man the System Is Named For EMPHOS Group, May 30, 2026, 6 min read Heinrich Rudolf Hertz lived from 1857 to 1894. He was 36 when he died. In the eight years between his doctoral thesis and his death, he produced one of the most consequential experimental results in the history of physics. And he did it with apparatus he built by hand, in a university laboratory, while almost nobody outside the small circle of mathematical physicists understood what he was looking for. He proved that electromagnetic waves exist.... Read more...
The Environmental Footprint of an Intelligence That Does Not Need a Data Center
The Environmental Footprint of an Intelligence That Does Not Need a Data Center EMPHOS Group, May 30, 2026, 7 min read Microsoft is building nuclear reactors. Amazon has bought a power plant. Google's emissions are up 48 percent over four years. The frontier AI companies are not being secretive about what is coming. They are openly investing in dedicated power generation because their architectures require it. The question is not whether the current trajectory is environmentally sustainable. The question is whether it is architecturally necessary. It is not. This is... Read more...
Heinrich and the Large Language Model. A Head to Head
Heinrich and the Large Language Model. A Head to Head EMPHOS Group, May 30, 2026, 7 min read Heinrich and a large language model can both answer the question "what is a mammal?" Both return a coherent English sentence. To a person looking at the screen, the two outputs are interchangeable. The physical work that produced them is not. An LLM produced its answer by running a forward pass through several hundred billion learned weights, predicting the most probable continuation of your question token by token, then stopping when it... Read more...
Heinrich as Software. What You Are Actually Running
Heinrich as Software. What You Are Actually Running EMPHOS Group, May 30, 2026, 6 min read Most artificial intelligence products today are a thin interface in front of a model that lives somewhere else. The model is enormous, the model is remote, and the model is rented. You are not running anything on your machine. You are sending text to a data center and reading the reply. Heinrich is the opposite shape. Heinrich is software. It has a directory on disk. It has files you can list. It has a... Read more...
VOXIS Is Trying to Build Speech Without Weights
The dominant assumption in AI voice is simple: better speech requires larger models. VOXIS starts from the opposite premise. If voice can be measured as acoustic structure and reconstructed from that structure directly, the future of speech may be lighter, faster, and more controllable than today's black boxes. Read more...
PRISM Is What Child-Safe AI Should Look Like
Most learning software assumes the child must adapt to the system. PRISM is built on the opposite assumption. It adjusts pacing, explanation style, and mode to the learner in front of it while treating safety as an architectural rule rather than a policy statement. Read more...
The Living Sphere Is Not Decoration. It Is the Interface.
Most AI products still present themselves as branded text boxes. The Living Sphere was built from a different belief: if an assistant is meant to feel present, the interface has to communicate state, attention, and responsiveness before a single sentence appears on screen. Read more...
ETTS Built Haven's Voice as Infrastructure, Not Decoration
A product voice should sound like it belongs to the product. ETTS exists to make that possible. It is not a demo generator. It is a corpus, training, verification, and export pipeline built so Haven can speak locally in a voice that is actually its own. Read more...
CAMS Code Is the AI Code Editor That Never Calls Home
Most AI coding tools treat your repository like something to upload. CAMS Code was built around the opposite assumption: the machine you are on is the boundary. The result is a local-first code editor that streams files into a live workspace, works offline, and does not need a monthly subscription to justify its existence. Read more...
HAVEN MD: Clinical Intelligence in the Palm of Your Hand. Zero Hallucination.
Every clinical AI tool on the market requires a cloud connection. Every LLM deployed in medicine can fabricate a drug name, invent a contraindication, or hallucinate a dosage. In medicine, that is not a product defect. It is a patient death. HAVEN MD does not guess. It cannot guess. That is not a design choice — it is how the architecture works. Read more...
We Solved the Device Shipping Problem for AI. Here Is How.
A HAVEN device needs to carry its intelligence layer in a file small enough to ship, update overnight, and reconstruct instantly on an ARM processor drawing 15 milliwatts. The Katz Protocol answered the question of whether that was possible — on real data, at real scale, with a pass condition of exactly zero error. It passed. Read more...
EMPHOS Group Has Applied to Be One of Canada's 50 Most Investible Ventures
The Foresight 50 is Canada's benchmark for the ventures most worth backing. EMPHOS Group Corporation — an AI company building intelligence that uses 11,000 times less energy per query than the industry standard — applied because the case for what we are building belongs in front of the people who fund Canada's technological future. Read more...
HAVEN Lens: The World, Explained. Privately. Honestly.
You walk into a building in Tokyo. Every sign is in Japanese. You do not reach for your phone. You look. The translation appears — instantly, offline, without asking. HAVEN Lens is not better smart glasses. It is the first AR intelligence device that runs entirely on-device, shows only what it knows, and has never heard of a cloud server. Read more...
HAVEN Watch: Health Intelligence That Cannot Lie to You
Every smartwatch on the market gives you data. HAVEN Watch gives you knowledge. The difference is Heinrich AI — a knowledge field that traces health relationships, reports what it cannot confirm, and runs entirely on your wrist with no server, no subscription, and no ability to fabricate an answer. Read more...
This Is What EMPHOS Is Building — And Why Every Product Is the Same Mission
EMPHOS Group Corporation builds software and hardware designed for people first. That is not a tagline. It is the filter every product is run through before a single line of code gets written. Here is the stack as it stands today — and why the products that seem most different are actually the most connected. Read more...
Nine Days: The Sprint That Built a Patent-Pending AI Company
On April 10, Heinrich AI was an idea. On April 18, it was a patent-pending technology inside a legally incorporated British Columbia corporation with a founding team, a live website, and 23 million knowledge concepts in its field. Here is what nine days of building actually looks like. Read more...
The Idea That Changed Everything — How Heinrich AI Was Born
Most ideas for new AI systems begin with a theory. Heinrich began with a question that had no satisfying answer in any existing system: what if AI stored knowledge the way physics stores information — not as text predictions, but as addresses in a field? The question led somewhere nobody had been before. Read more...
Haven Was the First Proof That the EMPHOS Philosophy Could Become a Product
 Before Haven existed, EMPHOS was a research operation with a philosophy and a protocol stack. Haven was the moment the philosophy became something you could actually use — an AI assistant built not to compete for your attention, but to earn it Read more...
Before Any Product Existed, EMPHOS Spent a Year Doing the Science
Most software companies build first and research later. EMPHOS did it the other way around. Before a single product shipped, before the company had a legal name, the work that would define every product in the stack was already done — measured, documented, and reproducible. Read more...
EMPHOS Group Corporation Is Officially Incorporated
EMPHOS Group Corporation is officially incorporated in British Columbia, Canada. Founded by inventor and CEO Victor Brodeur alongside Directors Kelly Meeds and Nadine Collison, EMPHOS is building a suite of AI and voice software products engineered around empathy, mindfulness, and presence — including Heinrich AI, Haven, VOXIS, ETTS, and HAVEN Ear. Read more...
The Environmental Cost of AI Is Not an Accident — And We Are Fixing It
The AI industry will consume more electricity this year than the entire country of the Netherlands. That number will double within three years. The response from the industry has been renewable energy commitments and more efficient chips. Neither addresses the cause. We are addressing the cause. Read more...
What the AI Industry Gets Wrong About Energy
GPT-4o uses 0.34 watt-hours per query. Heinrich uses 0.00003. That is not a 10% improvement. That is not even a 10x improvement. That is 11,000 times less energy per query — and the gap does not close at scale. It widens. Here is why. Read more...
We Are Building a Hearing Aid Powered by Wave Physics
Every attempt to put real AI in a hearing aid has failed for the same reason: the chips run hot. A conventional AI inference chip draws 1 to 5 watts continuously inside an ear canal. That is a burn hazard. Heinrich draws 3 milliwatts. That is not an optimization. That is a different physics. Read more...
Heinrich Can Now Hold a Conversation — Without a Language Model
Today Heinrich answered a question in plain English. Not because a language model predicted the words. Because the frequency field contained the knowledge, the HSR pipeline composed it honestly, and the result was a sentence that is guaranteed to be true — or guaranteed to say it isn't. Read more...
Heinrich Now Has 1.75 Million Nodes — And Still Uses 0.2% CPU
Three days ago Heinrich had 128 concepts. Today it has 1.75 million. The CPU usage is still 0.2%. The RAM is still 78 megabytes. This is not optimization. This is the architecture doing exactly what it was designed to do. Read more...
Heinrich Answered Its First Real Questions Today | EMPHOS Group
Today Heinrich AI had its first real conversations. Real questions, typed into a chat interface, answered by a knowledge field that has been learning continuously since this morning. The results were better in some ways, more honest in others — and in one case showed something no other AI system does by default. Heinrich said it didn't know. And it was right. Read more...
Why Local-First AI Wins | EMPHOS Group
The default assumption in AI software today is that intelligence lives in the cloud. Local is framed as the compromise. We think that framing is wrong — and the gap between local and cloud is closing faster than the incumbents want to acknowledge. Read more...
What We Learned Building Heinrich | EMPHOS Group
Heinrich AI was conceived on April 10, 2026. Two days later it was running software with a verified proof of concept and 532 passing tests. That timeline happened because the architecture was right. Read more...
Haven: What It Means to Own Your AI | EMPHOS Group
Every major AI assistant today comes with the same fine print: you do not own it. You rent it. Monthly. Indefinitely. Haven is built on a different premise. You buy it once. You own it permanently. Read more...
The Problem With How AI Remembers | EMPHOS Group
Every major artificial intelligence system available today is built on the same foundational architecture. Knowledge is compressed into billions of numerical weights. This produces remarkable results — but it has three structural problems that no amount of engineering can fully solve. Because they are built into the foundation itself. Read more...
Introducing EMPHOS HQ — Built the Same Way as the Software
Most software companies show you what they shipped last quarter. EMPHOS is showing you where it's going. This is the concept for the EMPHOS Group headquarters — a three-level Earthship-inspired building in Fraser Valley, BC, powered by solar, cooled by the hillside it sits in, and built on the same principle as every EMPHOS product: local-first, owned not rented, designed to last. Read more...
Why EMPHOS Exists | EMPHOS Group
We started with a simple observation: the tools people use every day are either technically shallow, visually forgettable, or exhausting to operate. The AI revolution promised to fix this. In many cases, it made things louder without making them better. EMPHOS is our answer. Read more...