Stay curious.building.learning.connected.
Stay connected.

Fast vs. thinking models.

Fast models answer instantly. Thinking models plan and check their work first.

Try thisUse fast for drafts. Use thinking for analysis and multi-step work.

Some AI models answer instantly. Others, often called reasoning or thinking models, take extra time to plan, work through steps, and check themselves before answering.

Agents act, so manage them like a new hire.

An agent doesn’t just answer. It reads files, sends messages, and changes systems.

Try thisGive every agent an owner, a clear job, limits, and an off switch.

An agent doesn’t just answer questions. It takes action: it reads files, sends messages, and updates systems. Anything that can act needs an owner, a clear job, and limits.

Loop engineering.

Agents work in loops: plan, act, check, adjust, repeat until the job is done.

Try thisDefine “done” first, and add a human check where mistakes are costly.

Agents work in loops: plan, act, check the result, adjust, and repeat until the job is done. The quality comes from how the loop is designed, especially what it checks and when it stops.

Rent the model. Own the harness.

The harness is everything around the model: instructions, tools, data, guardrails, and tests.

Try thisKeep your instructions and tests in systems you control.

The harness is everything around the model: your instructions, tools, data connections, guardrails, and the tests that prove it works. Models change every few months. A strong harness lets you swap them without starting over.

Your data is the advantage.

Models are becoming a commodity. Your organization’s knowledge is not.

Try thisUse enterprise AI that commits not to train on your data.

Models are becoming a commodity. What makes AI valuable to you is your organization’s knowledge. Satya Nadella recently called this the reverse information paradox ↗: to get value from AI you share your expertise, and that expertise can end up improving someone else’s product.

Sources: The New Stack ↗ · CXOToday ↗

Zero trust, now for AI.

Verify every request, grant the least access needed, and assume breach. Agents too.

Try thisClean up overshared files and give each agent its own identity.

Zero trust means nothing is trusted by default, inside or outside the network. It rests on three rules: verify explicitly, use least-privilege access, and assume breach. Agents need the same treatment as people: their own identity, only the access the task requires, and logs of everything they do. Enterprise platforms that carry your existing permissions, labels, and audit logs into AI give you a head start.

Sources: NIST SP 800-207 ↗ · Zero Trust overview ↗