The real power is shifting. While everyone else is busy arguing over which LLM has the highest benchmark this week, the industry’s "silent" giants are moving in a different direction.
Here is what’s actually happening right now:
1. The Agentic Era is here
Eric Schmidt recently highlighted a critical shift: the next multi-billion dollar companies won't be the ones making the AI. They’ll be the ones building the agents that out-execute everyone else.
If you can build an agent that handles a complex workflow better than a human or a generic chatbot, you don’t just have a tool you have the foundation of a massive company.
2. The Hardware Gatekeepers
We talk about software, but the "Netherlands bottleneck" is real. ASML remains the only company on Earth capable of building the lithography systems required for the next generation of chips. As the software race heats up, the physical hardware becomes the ultimate point of leverage. No chips, no agents.
3. Verification is the new Coding
With models like Anthropic’s Opus 4.6 and GPT-5.2 handling nearly all the heavy lifting in coding, the bottleneck has shifted. The skill is no longer "writing" the code it’s verifying it. We are moving toward a "self-verification" loop where AI audits AI, but the humans who know how to architect the system are the ones winning.
The Prompt: The Agent Architect
This is the exact framework needed to move from a "chatbot" to a persistent system. Copy and paste this to map out multi-step agentic workflows:
"Act as a Senior Systems Architect. Analyze the following business process: [Insert Process]. Break this down into 5 distinct 'Agent' roles. For each agent, define: 1. Their specific objective, 2. The required input data, 3. The verification criteria for their output, and 4. The trigger that hands the task to the next agent. Format this as a deployment roadmap."
The Weekly Audit: Verification is the new Coding
With models like Anthropic’s Opus and GPT-5 handling nearly 100% of the heavy lifting in coding, the bottleneck has shifted. The skill is no longer "writing" the code it’s verifying it.
This week, we audited the "self-verification" loop. We found that AI auditing AI is only as good as the human architecting the logic. To stay ahead, you must stop being a "writer" and start being an "auditor" of the systems you deploy.
Till next time,