Anyone can claim to see the future. Few put a date on it in public and let you check later. Every item below was published on Reality Shift(s) before the outcome was known — including the one that hasn't fully played out yet.
In "The AI Homecoming", Richard forecast that model compression (distillation techniques producing smaller, equally capable models) would drive AI processing back on-device and on-premises — reversing the assumption that AI meant sending data to distant cloud servers.
What happened: the edge AI market was on track to reach roughly $9.5B by the end of 2025; enterprises considering on-premises infrastructure on equal footing with the cloud for new AI applications rose from 37% to 45% in a single year. Ten months later, SPAN began installing residential GPU compute nodes — paired with a smart panel and 16kWh battery system — on the exterior of suburban homes, with a stated plan for 80,000 nodes nationwide by 2027. Early adopters of on-premises AI reported cutting operating costs by up to 70%. See the follow-up for the full picture.
In "En(vision)ing the Future", written when Worldcoin's iris-scanning Orb was framed as an AI/Metaverse curiosity, Richard forecast that a single, one-time scan wouldn't be sufficient — that a multimodal, continuous approach to identity verification across apps, sites, and physical spaces would be necessary.
What happened: in "The Orb Isn't Watching You. Your Agent Is." (Jun 2026), the follow-up confirms adoption scaled to roughly 18 million people verified across 160 countries, with World ID integrated into Docusign, Okta, Tinder, Zoom, and Vercel. But the deeper forecast — continuous, in-session verification, as opposed to a one-time entry check — remains unbuilt, especially now that AI agents, not humans, are increasingly the ones acting online. Naming what's still missing is part of the track record too.
In "The Real Revolution Isn't Robots", Richard argued that AI-orchestrated smart battery and energy systems — not flashy humanoid robots — would become the invisible backbone making AI infrastructure economically and physically viable, citing early cases of 30–40% peak cost reductions and facility-level annual savings near $250K.
What happened: in "The Signal Was in Your Walls" (May 2026), Richard confirmed SPAN's residential rollout as "the energy intelligence layer I described in August, deployed exactly as I outlined" — the same smart panel and battery system paired with the GPU nodes from the AI Homecoming forecast above, arriving together in a single commercial product.
In "The Mask Came Off", Richard argued VR and AR headsets keep failing for a biological and social reason, not a technical one — people reject things on their faces — and forecast that environmental and spatial computing would win where face-mounted computing couldn't.
What happened: in "The Web Just Adopted Electronic Protection" (Sep 2026), Richard pointed to Zuckoff — an app that fingerprints the Bluetooth signature of Meta's smart glasses and alerts you when a pair is nearby — as that rejection becoming an actual paid product category: roughly 5,000 users in its first month, a $9.99/year premium tier, and EU regulators already in contact with the developer. The Signal Report built on this pattern traces where else it's showing up.
In "The Coming PC and Smartphone War," Richard forecast that mobile devices — not desktops or standalone AI gadgets — would become the primary catalyst bringing AI processing power directly into people's hands.
What happened: foldable and AI-optimized phones (including Samsung's Galaxy Fold series) shipped with on-device model compression that made the forecast concrete, alongside industry data showing over 75% of major AI models released in 2024 achieved the compression needed to run locally.
A Reality Check applies this same forecasting discipline to your specific sector and roadmap.