AI Is Moving Fast. Is the Enterprise Keeping Up?
Four observations from ODSC West: the gap between AI vendors and enterprise adoption, BigQuery’s ML ambitions, AI sovereignty, and the emerging discipline of agent reliability.
LangGrant Windocks
You Can’t Stop Employees from Using AI. You Can Make It Safe.
You cannot stop employees from using AI. An independent safety layer lets them use Claude Dashboards while the enterprise controls database access, query execution, and reusable reasoning.
Would You Put Your Enterprise in the Hands of an AI This Powerful?
A lesson from the boatyard: powerful enterprise AI needs human visibility, judgment and control—and a memory for the reasoning people and AI create together.
Sam Altman Says Humans Still Need to Stand Behind Decisions. But How?
Meaningful human judgment requires more than approving AI output. Humans need to understand, shape, and reuse the reasoning behind consequential decisions.
Knowledge Flow and AI Point to Two Sides of the Same Transformation
Knowledge flow and human/AI collaborative reasoning create an opportunity for compounding knowledge and continuous organizational learning.
Jensen Huang Says There’s a 0% Chance AI Will Destroy the World by 2030. Can We Sit Back and Relax Now?
Even if AI will not destroy the world, there is work to do. Human judgment must actively shape the reasoning behind business decisions.
Dario Amodei and Sam Altman Want to Slow Down AI. Are We Just Going Down the Cliff a Little Later?
Slowing AI down buys us time. Getting humans ahead requires AI, software, and humans to reason together.
Google Says AI Can’t Replace Human Judgment. Does Anyone Have Time to Use That Judgment?
Keeping humans around is not enough. Human judgment must become part of AI reasoning—and part of the organization’s reusable intelligence.
You Define the Intent of Autonomous Agents — Not Them.
Autonomous agents can decide how to execute, but the enterprise must continue to define what they are executing and why. They need the equivalent of a skipper’s standing orders—not just permissions and guardrails.
Your AI Agent Shouldn’t Own Its Memory. You Paid for It. You Should Own It.
On a boat the navigation log belongs to the boat, not the autopilot. Agent memory works the same way: the organization paid for the model, the data and the expertise, so the reasoning and the lessons should outlive whichever agent happens to be steering.