Day 22 of Green, Efficient AI is live: The Silicon in the Rack. A cloud-native platform team approved a hardware refresh. Newer accelerators. Better performance per watt. Clean operational case. The sustainability lead asked one question before signing off,…
Day 20 closed the arc on the system around the model with a single frame: once capacity is standing, the bill runs on wall-clock, and utilization becomes the number that matters. That frame covered operational emissions — the electricity, cooling, and compute…
Day 21 of Green, Efficient AI is live — The Footprint Already Spent. A financial services company published its AI carbon report. Every inference call metered. Every kilowatt-hour tracked. Grid intensity factored in by region. The number looked clean. An…
49 new efficient skills for your AI agent. The kind that spot waste — before your cloud bill or carbon footprint does. The latest edition of Technology Bytes is live — and it comes with something you can actually use, starting today. Last article, I wrote…
Organizations around the world are investing billions in Agentic AI. New foundation models are released almost every month, intelligent agents are becoming more capable, and the pace of innovation has never been faster. So why are so many organizations still…
In the last edition of Technology Bytes, I wrote about something our industry has been overlooking — one of the most important AI skills of all. Efficiency. Not efficiency as an afterthought. Efficiency as a skill that every AI agent should learn from the…
Organizations around the world are investing billions in Agentic AI. New foundation models are released almost every month, intelligent agents are becoming more capable, and the pace of innovation has never been faster. So why are so many organizations still…
Artificial intelligence has become remarkably capable. Today's AI agents can search the web, write software, analyze documents, reason through complex problems, interact with enterprise systems, and even collaborate with other agents to accomplish…
AI Skill = Capability × Efficiency Today's AI agents are measured by what they can do. Can they search? Can they reason? Can they write code? Can they use tools? Can they collaborate with other agents? But as enterprises move from deploying a handful of…
Day 19 covered the compute the team triggers itself — evals on merge, LLM-as-judge sweeps, notebook experiments — and walked the Meter, Gate, Sample, Substitute ladder that moves it from an unbudgeted reflex into a governed line on the bill. That closes the…
Day 20 of Green, Efficient AI is live — The Capacity That Waits. A retail team ran per-call optimizations across their assistant for months. The bill barely moved. When they pulled utilization data across the full stack — model endpoints, vector database,…
Speaking at Google I/O on Lean Agentic AI: as organizations scale from a handful of agents to thousands, systems must be efficient by design — across cost, energy, and carbon. Slides and demo available on GitHub.
It was a pleasure speaking at the Google I/O event on Lean Agentic AI. As organizations move from deploying a handful of AI agents to thousands, efficiency can no longer be an afterthought. We need to design agentic systems that are efficient by design—across…
Tomorrow, I'll be speaking at Google I/O Connect, Bengaluru on a topic I've become deeply passionate about: Lean Agentic AI. We're entering an era where organizations will deploy not one or two agents—but thousands. The question is no longer: can we build…
Artificial intelligence has never been more capable. Every few months, a new generation of foundation models arrives with better reasoning, larger context windows, improved tool use, lower latency, and reduced costs. From a technology perspective, the pace of…
🚨 AI is advancing faster than your organization can transform. That may be the single biggest reason why so many enterprises are struggling to demonstrate ROI from Agentic AI. Every month, we see smarter models, better reasoning, lower costs, and more capable…
Day 18 covered the data that AI systems leave behind — vector stores, indexes, embeddings that outlive the features that created them. That was the persistence side of the invisible bill. There is a compute side to it as well, and it fires before a feature…
Day 19 of Green, Efficient AI is live — The Test Tail. A team spent six months bringing production AI costs down. The playbook worked — production inference became dramatically cheaper. The monthly bill barely moved. The gap was hiding in plain sight. Every…
Day 17 examined what every call carries on the way in — the system prompt, the tool catalog, the examples, the payload that pays a bill on every request whether it needs to or not. Today's issue steps back from the call itself. Every AI feature leaves…
Day 18 of Green, Efficient AI is live — The Forever Data. Every AI system carries two footprints. The work it performs, and the data it leaves behind. The second one rarely gets named. A B2B platform running a retrieval assistant found that three routine…
AI doesn't need another breakthrough model. It needs a return to engineering fundamentals. Somewhere between foundation models and agentic AI, we confused capability with architecture. Because a model can reason doesn't mean every task requires reasoning.…
AI doesn't need another breakthrough model. It needs a return to engineering fundamentals. Somewhere between foundation models and agentic AI, we started confusing capability with architecture. Bigger models, more agents, longer context windows, and…
Day 16 traced the compute that safety adds on the outside of every call. Day 17 opens the same call up and looks at what is actually inside it — the request the model reads, not the message the user sent. The gap between the two is usually enormous. A user…
Day 17 of Green, Efficient AI is live — The Invisible Payload. A user asks a support agent to reset their password — seven words, roughly ten tokens. The trace shows the model actually receives thousands of tokens. A system prompt written months ago. Every…
Day 15 covered the routing discipline — the choice of which model answers a request, and the four-rung ladder of Filter, Classify, Route, Call that decides whether a call should ever reach a heavyweight model at all. That settles one piece of the system…