A Room Full of Builders & Bridge‑Builders

Check‑in was quick, the taproom buzzing by 6 p.m.  Digital‑health founders clinked glasses with IoT hobbyists, and a surprising number of philosophy grads compared notes with game‑dev veterans.  Wild AI’s “no slides, no gatekeeping” rule flattened every hierarchy: just five‑minute fire‑talks, name tags, and unstructured networking.

Why it matters: The group isn’t trying to be another conference.  It’s a sandbox where API tinkerers, educators, and sales‑ops strategists collide—and leave with phone numbers, not business cards.

Deep‑Dive: Large Language Models, Semantic Kernels & IoT Mash‑Ups

Semantic search done right. Our kickoff discussion explored how Microsoft’s Semantic Kernel is fast becoming a universal adapter for LLMs—now with Java SDK support and a healthy stable of open‑source “skills.” The crowd zeroed in on vector databases as the secret sauce behind lightning‑fast retrieval of resumes, conference transcripts, and other semi‑structured text.

From Next.js to full desktop automation. Front‑end devs shared tales of wrestling (and eventually making peace) with Next.js, while a live demo showed an LLM orchestrating mouse clicks and keystrokes on a Windows machine—powered by YOLO‑based object detection over the screen buffer. The big takeaway: abstract the low‑level API calls so your prompt can read like a checklist, not a stack trace.

IoT, meet AI. A show‑and‑tell of personal projects ranged from a Raspberry Pi–driven geothermal controller (yes, using the old thermistor lines) to ESP32 wireless relays. The conversation inevitably landed on when to embed tiny AI accelerators on‑board versus streaming sensor data to the cloud for heavier inference.

AI for Revenue Teams, Courtesy of TSO Consulting

• Market universe → ICP: ML filters every potential account down to “worth calling.”
• Intent scoring: BANT signals feed a probability model; reps see a ranked list, not a spreadsheet.
• Outreach automation: Python micro‑services spit out cold‑call scripts and LinkedIn sequences on demand.
• Modular by design: Swap in your CRM, keep only the steps you need.

Bottom line: Sales Ops is finally getting the same ML love that marketing enjoyed five years ago.

Hard‑Data Reality Check Dr. Matt Phillips (Healthcare AI)

Most ML tutorials glorify “easy data”: perfectly labeled, open‑sourced, and irrelevant to clinics.  Real healthcare projects, he reminded us, live in “hard data” messy EHR exports, half‑filled Access tables, and sensors that drift.  His playbook:

1. Modernize the pipeline (Power Apps + SharePoint > Access).
2. Marry data scientists with MDs for iterative labeling.
3. Validate relentlessly; patient risk tolerance ≠ ImageNet accuracy bragging rights.

When Aristotle Meets GPT Dr. Mark Collier’s Ed‑Tech Experiments

A Socratic Tutor GPT prompts students with “Why?” until they articulate reasoning.  In VR, avatars of Confucius and Karl Marx spar over economic theory while spectators earn points for spotting logical fallacies.  Next up: a fully explorable metaverse “library of ideas.”  Unity/C# contributors welcome.

The Corridor Conversations

• Jobs & layoffs: Yes, Big Tech trimmed fat—but small teams here are hiring for niche AI roles.
• Multi‑agent orchestration: Everyone’s hitting context‑window walls; expect shared tooling soon.
• “Boring” automation: The group’s mantra: automate the drudgery so humans can chase originality.

What’s Next?

• RSVP for the June 19 meetup same DayBlock taproom, new lightning‑talk lineup.
• Pitch a 5‑min talk: Email jake@attentio.ai with a one‑sentence teaser.
• Demo days coming: Expect deeper dives on the YOLO desktop‑control stack and the VR philosophy sandbox.
• Sign up link: Wild AI | Meetup

TL;DR: Wild AI is equal parts tech dojo and coffeehouse debate.  Show up, ship a 300‑second demo, and find your next collaborator over a pint.  See you on the third Thursday!

‍