WriteArm.ai

Just got published in the The AI Journal The piece is about something I've been obsessed with for ten years: why are we making AI regenerate the same intelligence over and over again? Every time you ask an LLM a structured question, it rebuilds the answer from scratch. Every single time. That's like hiring someone who forgets everything overnight and has to relearn their job each morning. We built something different. Matrix-OS treats intelligence like an artefact. You structure it once. Then you just... execute it. The results are kind of absurd: -99% less computational waste. -Answers in milliseconds instead of seconds. -Same result every time (no hallucinations, no drift). -And it remembers what happened yesterday. It's not magic. It's just... structural. If you're spending serious money on AI that keeps recomputing the same stuff, or you need answers you can actually audit and trust, take a look. https://lnkd.in/edsaBFW3

Last month a big shift happened in the world of AI. One that carries the trace of a far greater shift to come. University of Cambridge’s Polymathic AI collaboration announced two new AI models, called Walrus and AION-1, both of which are trained on physics rather than words. Instead of learning patterns in language, they learn patterns in physical systems. They can then apply the knowledge gained to completely different fields. For example, Walrus can tackle systems ranging from exploding stars to the movement of bacteria. As WriteArm.ai's Martin Lucas sets out in his latest article, ‘Cambridge is training models on the rules of reality, not the patterns of language. That is not a technical tweak. That is a philosophical break. Because it quietly admits something the industry has avoided for years: The future of intelligence is not more text. The future is structured reality.’ One of the Cambridge scientists, Miles Cramer said of the breakthrough, “I continue to be awed by the fact that a multi-disciplinary physics foundation model works at all, let alone at this level.”  But, as the article reveals, a new and even more awe-inspiring form of AI is on the way; one which Martin has spent over a decade developing. It’s called Synthetic Deterministic Cognitive Intelligence (SDCI). At their heart Walrus and AION-1 are, like LLMs, probabilistic. SDCI, by contrast, is deterministic. It’s not AI trained on physics. Rather it creates a physics for AI. As Martin explains, it’s not “a better model” rather: ’It is an infrastructure layer that replaces probabilistic generation with deterministic cognition.’ In summary, Polymathic has changed what AI learns. SDCI changes what AI is. Read the article here: https://lnkd.in/eyT6SD8T

Why is Nike’s messaging so effective? The answer might surprise you. And what’s equally surprising is that it can be mapped scientifically. WriteArm.ai’s Martin Lucas is the man behind the science. He sets out his analysis of the company’s success in his latest Medium article: ‘When people reach for Nike, they’re rarely asking: “Is this the best product?” They’re asking something quieter: Am I still moving? Am I still trying? Am I still in this? Nike works because the answer is almost never “the best”. Rather, it’s:  ‘“I’m still in motion.” “I haven’t given up.” “I’m not done yet.”’   Nike doesn’t attract one type of person, but rather one state of mind: ‘Nike never asks: “Are you good enough?” It asks: “Are you willing to keep going?” That single question travels across ages, bodies, ambitions, and ability levels. And that’s why Nike scales without fragmenting.’ Martin has mapped what’s going on in the attached image. It's kinda simple. And kinda not. Want to find out more? Then read the full article here:  https://lnkd.in/eCat-5Pn Go on, just do it! #copywriting #aicopywriting #nike

The genAI industry is looking shaky. Why? ‘It has been a statistical mimicry machine scaled to planetary size. The industry knew this. The public did not.’ The words of WriteArm.ai's CIO Martin Lucas in his latest Medium article. Google’s CEO Sundar Pichai this week issued two warnings. One was to the world about the consequences of the AI bubble bursting. The other was to users, urging them not to blindly trust AI outputs. The two warnings are umbilically tied, because it’s the steadily eroding trust that will help burst the bubble. That’s a huge problem for the industry, because, as Martin explains, current AI models have hallucination baked in. They are all, in essence, probabilistic, meaning that every step of the generation process is the most statistically plausible one, rather than the most truthful or appropriate one. ‘It works brilliantly until you need truth, logic, or governance. At that point it collapses.’ How can this trust gap be solved? By applying a new form of AI. 👉 One that thinks rather than guesses. 👉 One that has an in-built truth protocol. 👉 One that’s coming soon. To find out more, read Martin’s article in full here:  https://lnkd.in/e5UM7XMq #genai #AIcopywriting 

Er, you don't say! https://lnkd.in/eAxQh_fh

Most marketing fails to connect on a subconscious level. But a new form of AI based on decision science, invented by our CIO Martin Lucas, is set to change all that. Find out more in this episode of The Digital Executive (one of Apple’s top 10 tech podcasts). Listen on Apple: https://lnkd.in/eGEfea_m Spotify: https://lnkd.in/e9-QHDSC  YouTube: https://lnkd.in/ewkD7UmP