Everyone’s Building Agents. So Why Do They All Feel the Same?

Everyone’s Building Agents. So Why Do They All Feel the Same?

After reviewing a bunch of new AI startups, we noticed a pattern.

Each one launched some flavor of an “agentic AI” system — spawn agents, run them in parallel or sequence, then either merge the results or pipe them downstream. Smart stuff.

But behind the scenes? It was all glue code.

Custom connectors. Prompt tuning. APIs duct-taped together. The real focus wasn’t the solution, it was just getting the plumbing to work.

That got us thinking.

What if we stopped rebuilding infrastructure and started building solutions?

What if there were a crowdsourced way to share:

• Agents (LLM + persona + RAG context)

• Methods (high-quality prompts for specific tasks)

• Flow logic (how agents pass data and collaborate)

We imagined a world where writing a resume, generating a strategy deck, or building a custom email campaign wasn’t 20 steps of manual trial-and-error — but a fieldable AI app made from reusable blocks, ready to go.

That’s the project we’re building.

Interested in contributing? Send us your background and LinkedIn profile at dwc2425@utexas.edu & sbrunner@zylergy.com or message us at David Chen & Stefan Brunner and tell us why you want to be part of AgenticBlox! We welcome both technical and non-technical innovators.