Ideas Weave Every Narrative with AI.
We’re obsessed with AI benchmarks, but the real battle isn’t about intelligence—it’s about trust. From hikers stranded by Gemini’s confident advice to OpenAI’s rogue agents, the AI race is shifting. The winners won’t have the smartest models; they’ll have the best accountability infrastructure, local execution, and safety nets.
Anubis spent a year shipping WebAssembly to make proof-of-work bot defense smoother for humans. But every client-side efficiency also optimizes the solver for AI scrapers. We aren’t building a wall against bots; we’re paving a highway for them, while taxing real users to burn CPU cycles just to prove they have a pulse.
The impending ‘shortage of everything’ isn’t a production failure, it’s a leverage failure. Once AI and robotics make human labor obsolete, the masses lose their sole bargaining chip. The economy will restructure to serve only capital owners, creating artificial scarcity in a world of abundance unless we decouple wealth from work.
You’ve watched the demos—an AI recreating Minecraft or turning a photo into a 3D model. But when you plug it into your workflow, it fails. Public AI benchmarks have become marketing artifacts, not scientific measurements. The real value isn’t in the model anymore; it’s in the harness layer.
We are told AI is just linear algebra. You can trace the gears, follow the gradient descent, and see the exact math turning. Yet, the emergent output feels indistinguishable from magic. Understanding the math doesn’t demystify AI; it just proves we are playing with dimensions we don’t actually comprehend.
Every SaaS company is desperately slapping an ‘AI Copilot’ onto their dashboard, but it’s panic masquerading as innovation. The real shift isn’t making existing tools smarter; AI is dissolving the tools entirely. The app, the UI, and the abstraction layers are evaporating. If your career depends on tool mastery, your expertise is melting away. The new scarce power isn’t execution—it’s accountability.
AI isn’t killing product manager jobs; it’s exposing how many were just mechanical translators all along. As AI automates execution, the ultimate career moat isn’t tech-savviness—it’s the messy, un-automatable human realities of physical context, office politics, and taking the blame when things break at 2 AM.
AI didn’t just make coding easier; it destroyed its value. When a designer can use ChatGPT to write 95% of an app’s code, technical execution is no longer a moat. The future of software isn’t about features—it’s about emotional resonance, aesthetics, and taste. Just ask the Pomodoro timer app making $18k a month off a hand-drawn cat.
The tech world is furious that America’s largest school districts are hitting pause on AI, calling it a barrier to progress. But the real battle isn’t safety versus innovation—it’s about who gets to run massive, unregulated experiments on children. Schools are finally demanding proof over hype.
We’ve all felt the dread of learning a new skill, knowing AI can do it faster. But when it comes to programming, AI’s fluency in strict languages like OCaml isn’t a reason to give up—it’s a signal. Learning difficult, compositional languages trains the exact judgment and auditing capabilities that probability engines lack. The future belongs to those who can spot the elegantly flawed code AI generates.