If you’re building with agents:
Start simple. Test well. Use standards.
You can explore the Standard Agent and Arazzo Engine on GitHub.
Link to full write-up below!
If you’re building with agents:
Start simple. Test well. Use standards.
You can explore the Standard Agent and Arazzo Engine on GitHub.
Link to full write-up below!
The future: “Sovereign stacks.”
Internal agentic workflows that keep data centralized and useful, instead of siloed SaaS.
The future: “Sovereign stacks.”
Internal agentic workflows that keep data centralized and useful, instead of siloed SaaS.
Advice for new developers:
* Be better than the AI and go beyond average
* Build projects, contribute to open source
* Use AI for routine tasks, focus on unique value
Advice for new developers:
* Be better than the AI and go beyond average
* Build projects, contribute to open source
* Use AI for routine tasks, focus on unique value
On context:
“Don’t overload LLMs with tools you don’t need.”
Just-in-time tooling keeps reasoning clear.
On context:
“Don’t overload LLMs with tools you don’t need.”
Just-in-time tooling keeps reasoning clear.
The problem with prompt-driven glue logic?
Unrepeatable results. Hard to test.
Recorded workflows, via Arazzo, are deterministic and easy to audit.
The problem with prompt-driven glue logic?
Unrepeatable results. Hard to test.
Recorded workflows, via Arazzo, are deterministic and easy to audit.
Reliability comes first.
Run agents in test environments. Capture successful runs as auditable, repeatable workflows before going to production.
Reliability comes first.
Run agents in test environments. Capture successful runs as auditable, repeatable workflows before going to production.
On multi-agent frameworks:
“Most teams can’t deploy a single agent reliably. Start small.”
Granular agents like a “receipt filing agent” are more valuable than premature orchestration.
On multi-agent frameworks:
“Most teams can’t deploy a single agent reliably. Start small.”
Granular agents like a “receipt filing agent” are more valuable than premature orchestration.
“An agent is simple code: it calls an API and an LLM that decides when to call it.”
Simplicity is the point. The Standard Agent reflects this principle.
“An agent is simple code: it calls an API and an LLM that decides when to call it.”
Simplicity is the point. The Standard Agent reflects this principle.
🔗 discord.gg/TdbWXZsUSm
🔗 discord.gg/TdbWXZsUSm
⚡️ Try it now on jentic.com
⚡️ Try it now on jentic.com