Autonomy is a runtime decision, not a model decision
Vamsi Krishna
Command Palette
Search for a command to run...
Archive
66 posts published so far
When MCP earns its overhead, and when a direct API is better
Vamsi Krishna
A clean connector does not make a reliable agent
Vamsi Krishna
Tools, resources and prompts are not interchangeable
Vamsi Krishna
Where your MCP server runs is a production decision
Vamsi Krishna
What the MCP protocol actually standardises
Vamsi Krishna
Day 60: From model demos to mission-ready AI systems
Vamsi Krishna
Day 59: Why production AI is coordinated infrastructure
Vamsi Krishna
Day 58: Why governance belongs in the architecture
Vamsi Krishna
Day 57: Why retrieved content must stay untrusted
Vamsi Krishna
Day 56: Why evaluating the model is not enough
Vamsi Krishna
Day 55: Why observability is not optional for AI systems
Vamsi Krishna
Day 54: Why multi-agent systems need decision rules
Vamsi Krishna
Day 53: Why handoffs need contracts
Vamsi Krishna
Day 52: Why shared memory becomes a permission problem
Vamsi Krishna
Day 51: Why agents need a shared language
Vamsi Krishna
Day 50: Why role design matters more than agent count
Vamsi Krishna
Day 49: When decentralized agent behavior makes sense
Vamsi Krishna
Day 48: Why AI debate needs a reliable judge
Vamsi Krishna
Day 47: Why shared state needs structure and attribution
Vamsi Krishna
Day 46: Why peer agents need protocols, not vibes
Vamsi Krishna
Day 45: Why pipelines make AI handoffs inspectable
Vamsi Krishna
Day 44: Why dynamic dispatch needs ownership rules
Vamsi Krishna
Day 43: When hierarchy helps multi-agent work scale
Vamsi Krishna
Day 42: Why supervisor agents help and where they bottleneck
Vamsi Krishna