📬 Getting this in Promotions? Move it to Primary so you never miss an edition.
IT HootClub — AI Community Newsletter
Hands-on. Career-focused. Future-ready.
Issued 2026-06-21
Choosing the Contenders
Six open-weight contenders for the Milwaukee showdown, chosen by a search I had to get wrong first — and the curated database that now governs the events list itself.
Announcements
Milwaukee is hosting one of the biggest AI events in the region this week, and it's happening in our backyard: Summerfest Tech AI runs June 23–26, 2026, anchored at The Westin Milwaukee and spilling across downtown. Four days, more than a dozen tracks — advanced manufacturing, healthcare and biotech, fintech, martech, food and ag — keynotes, pitch competitions, founder features, and the kind of hallway conversations that actually move a tech community forward. Hosted by Milwaukee World Festival, Inc., it's the clearest sign yet that Milwaukee's AI ecosystem has arrived.
If you've been waiting for a reason to plug into this city's tech scene, this is it. Passes range from free to paid depending on the track and day, so there's an entry point at every level — register and see the full options at summerfest-tech.com/register.
Come say hello — there's $25 in it for someone.
I'll be at Summerfest Tech AI across several of the events, and I'd genuinely like to meet the people reading this. So here's the deal: the first person who walks up to me and says "hello" gets a $25 gift card on the spot. Look for the black hardAIs shirt. No catch, no pitch — just an excuse for the two of us to actually talk.
Want the full picture before you go? The complete four-day, session-by-session agenda is laid out in this week's Events section below — every keynote, track, and panel, so you can plan exactly which days and rooms are worth your time.
Building Intelligence
This is the edition the whole Building Intelligence series has been walking toward. The plan, stated plainly: take the curated Milwaukee tech inventory I've been building, hand it to an open-weight model I run myself, and find out whether that grounded local model can answer Milwaukee questions as well as — or better than — a frontier model like Claude leaning on its native web search. Curated knowledge versus the open internet. My data versus their tooling. That's the showdown.
But a showdown needs contenders, and this week is about choosing them — not scoring them. The scoring comes next, when I put these models on rented GPUs and grade them against the rubric I built back in April. Today is selection: how an engineer actually picks a model for a job. And the most useful thing I can show you is that I got it wrong the first time.
The search I ran — and why it failed
hardais.com has a model catalog I built: a few hundred AI models you can search in plain English. It returns only models that are actually in the catalog, ranked, with a note on why each one matched — and it never invents a model. So I typed in what I thought I wanted:
"an open-weight model that excels at data retrieval from a postgres database and well suited for fine-tuning"
What came back was a list of coding models. Nearly every one: a code specialist, a formal-proof agent, a model that operates software by reading screenshots. Not a single natural fit for the job I actually had. And here's the thing — the catalog wasn't broken. It answered my question faithfully. The problem was my question.
Here's what I'd forgotten about my own setup. In this showdown, the model never has to find the facts — I hand them over. The entire Milwaukee inventory is small enough to drop into the prompt whole, so the model isn't querying a database at all; it's reading a complete copy of one and answering from it. There's no retrieval step and no script translating my question into a targeted data pull, because none is needed: I give it everything and ask. That's the simplest form of grounding there is — don't make the model fetch, just feed it the whole thing — and it works precisely because the inventory is still small.
So the model's only real job is reading comprehension: take the supplied facts, answer the question in clean prose, follow the instruction, and say "I don't know" when the facts don't cover it. Not retrieval. Not querying. Reading.
Which is exactly why the catalog handed me coders. "Retrieval from postgres" describes a model that writes its own database queries to go fetch data — a skill called text-to-SQL (SQL being the language databases speak), and it's coding-adjacent, so coding models surfaced. The catalog answered faithfully. But here's the subtler truth, the one I didn't see at first: I wasn't asking the wrong question so much as asking it too early. A model that fetches its own data only earns its keep once the inventory grows too big to feed in whole — and I'm nowhere close to that. I went shopping for next year's model with this year's problem. ("Well suited for fine-tuning" was the same mistake twice — smuggling a later decision into this week's.)
That's the thesis of this whole series, turned back on me. AI is downstream of its data — and a search is downstream of its query. A faithful tool will still hand you the wrong thing when you give it the wrong job, or the right job a year too soon. The skill was never running the search. It's knowing what to actually ask for.
The search I should have run
So I re-asked. I dropped the plain-English box and drove the catalog's filters straight at the real job: Knowledge base Q&A (answering from provided knowledge), Open weight (I can download it and run it myself), and Commercial use: Yes — a hard requirement, because hardAIs is a business. Then I deliberately spread the results across size, because the single most interesting question in this entire showdown is this: once every model is handed the same facts, what does being bigger actually buy you?
That produced a real field. Here are the six I'm putting forward as contenders:
SMALLMinistral 3 3B· France · Apache 2.0 · the featherweight
MIDGemma 4 26B-A4B· United States
MIDgpt-oss-20b· United States · Apache 2.0
LARGEMistral Small 4· France · the heavyweight (the name's misleading — it's the biggest one here)
ALTERNATESQwen2.5-14B&Seed-OSS-36B· China · held for testing
And notice who made these. Two of the six come from the most recognizable names in AI: gpt-oss is OpenAI's open model, and Gemma is Google's. "Open-weight" doesn't mean fringe or second-rate — the same labs behind the frontier products you already use also give away models you can download and run yourself. This isn't a scrappy-underdog story. It's frontier labs handing you the keys.
And if the names read like alphabet soup — gpt-oss-20b, Gemma 4 26B-A4B, Ministral 3 3B — that's the other thing worth defusing, because every piece of those names means something. "oss" is open-source software. "20b" is twenty billion parameters, a rough measure of the model's size. And a suffix like Gemma's "A4B" actually tells you something genuinely useful once decoded: it's a mixture-of-experts design with about 4 billion active parameters out of 26 billion total — which is to say it delivers roughly 26B-class quality while running at closer to 4B-class speed and cost, light enough for a single consumer GPU. The scary-looking label was quietly advertising the model's best feature. Decode it and the intimidation drops away; it's a compact spec sheet, not a secret code. Making exactly that legible — turning this field's overwhelming sprawl into a plain-English resource you can actually browse — is the whole reason I built the hardAIs model catalog this search ran on.
A few notes on how this field got built, because the cuts teach as much as the picks. The one hard filter — commercial license — did real work: it knocked out a powerful 123B Mistral coding flagship (a "modified MIT" license with commercial restrictions tied to revenue) and another strong model whose license came back conditional. A capable model that's no longer maintained — "feature frozen," in catalog terms — got cut too; I'm not building on something already abandoned. The four active finalists run from a 3-billion-parameter featherweight to a Large heavyweight, and I deliberately set two US Mid-class models — Gemma and gpt-oss — head-to-head at the same weight. That asks a sharper question tucked inside the big one: at identical size, with identical facts, does the lab that built it still matter?
A word on the two from China
Two of the alternates, Qwen and Seed-OSS, come from China — and for a Milwaukee audience, that's a fair thing to raise. So let me be direct, because it's a useful point. An open-weight model running on your own isolated server cannot phone home. The worry about a foreign model is almost always about data leaving or someone steering it remotely — and a model whose weights you downloaded, running on hardware you own with no connection back to its maker, has no channel to do either. That's the deeper argument this whole edition keeps making: when you control the data, the model, and the box it runs on, where the model came from stops being a vulnerability. I'll be honest about the limit, too — isolation stops data from leaking and stops remote control; it does not scrub whatever bias or slant is baked into the weights, which travels with any model wherever it runs. But on the "is it spying on me" question, a self-hosted open weight simply can't.
How complicated does this actually need to be?
Today's approach works because the whole inventory fits inside the model's context window — the amount of text a model can read in a single pass. If the Milwaukee inventory ever grows past that, the job changes: I'd add real retrieval (fetching only the rows that matter) and probably a more capable model to match. So I'm not marrying a model this week. I'm choosing the right component for the job as it's shaped right now, and building so a better one can slot in cleanly if the work ever demands it. Select for today; keep upgrading cheap.
But sit with that if for a second, because it's the real lesson hiding in this whole edition. I keep catching myself assuming this system is destined to get bigger and more complicated — more data, retrieval, a heavier model, fine-tuning to shape its behavior, each one a road this series might walk in a future edition. Maybe. Or maybe a curated list of Milwaukee tech organizations and events is simply never that complicated, and a small model reading a complete copy of it is the entire answer, permanently. Building the elaborate version anyway — because elaborate feels like real engineering — would be a disservice to the actual problem. The honest question running under all of this isn't "what's the most powerful model I can deploy?" It's "is the need as complicated as I think it is?" I suspect that, far more often than anyone in this field likes to admit, the answer is no.
So at the most basic level, here's all I'm really out to prove right now: that the concept holds — that a small open-weight model I run myself, fed a curated source of truth, can answer Milwaukee questions as well as a frontier model searching the open web. If that's true, the simple version already wins, and the elaborate version stays on the shelf until something actually requires it. That's the field: six open-weight models, chosen on the right criteria for the real job, with my wrong turn left in because the wrong turn is the method. Next they go to rented GPUs and get scored — same April rubric, same Milwaukee questions, each one grounded against the inventory. The contenders are set. The fight is next.
Under the Hood
If you've read this newsletter for a while, you know the Events list has a recurring problem: bloat. Every week the pipeline scrapes Meetup and Eventbrite for local AI events, and every week it drags in things that technically match but aren't what you came for — a one-day "Generative AI Skills" corporate training, a bootcamp's sales webinar, an "AI networking party." Last week I trimmed the list from 22 events down to 13 by hand. That hand-trim was the tell: my filters could catch the obviously-off (a wine tasting, a body-sculpting class), but they couldn't tell a real community meetup from a polished commercial pitch — because both say all the right keywords.
So this week I stopped filtering by keywords and started filtering by authority. Instead of asking "does this mention AI and happen near Milwaukee?" — a question keywords answer badly — the pipeline now asks "is this run by an organization I've already vetted?" That question has a real answer, and it lives in a database.
That database is the Milwaukee tech inventory I've been building across this series: a curated list of the orgs, venues, and events that actually make up this ecosystem. As of this edition, the newsletter pipeline opens a read-only connection to it on every run — read-only meaning the pipeline can look but never change anything, so even a bug on my end physically cannot corrupt the inventory. From that connection, two rules now govern the Events section:
1. Seeded events are required intake. Anything I enter into the inventory by hand always appears, no filtering. I put it there on purpose; that's the editorial decision, made once, in the data.
2. The scrape is subordinated to the inventory. Meetup and Eventbrite still run — I don't want to lose the breadth — but now a scraped event survives only if its organizing group is one already in my inventory and flagged both trusted and local. Skelora isn't in my inventory. Neither is the bootcamp. So they're gone — not because a keyword filter finally guessed right, but because they were never authorized in the first place. The bloat dies by construction.
In the interest of transparency: right now the scrape authorizes nothing. I haven't yet registered the source links that tie a given Meetup group to an inventory organization, so every event in the inventory right now was seeded by hand. As I register those links, the scrape turns into a live monitor of groups I already trust — it'll catch their new events automatically. But finding a new organization stays my job. The lesson there is one this whole series keeps landing on: breadth doesn't come from scraping wider into unvetted ground. It comes from me widening the inventory.
Which is why I should set an expectation about this week specifically: the Events list below is long — far longer than usual. That's not the bloat coming back. It's the opposite. Summerfest Tech AI lands in Milwaukee this week, and it's a four-day conference with a packed, multi-track agenda — dozens of distinct sessions, each with its own time, room, and host. Because Summerfest is a vetted local organization in my inventory, every one of those sessions is authorized, so every one flows through. A normal week would be a short, tight list; this week a single trusted host happens to be running an entire conference, and the system faithfully surfaces all of it. The point was never that the list is short. The point is that every event on it earned its place — whether that's three events or sixty. Not one Skelora slipped in alongside the real ones. A long list of all-vetted events isn't a failure of curation; it's curation working exactly as designed.
Which brings me to a messier kind of call — and to an organization regular readers already met. Last week, the Association of Technology Professionals (AITP) was my cautionary tale: the grounded model had reported AITP as a Milwaukee organization, I traced that error to my own data, and I fixed it by flipping AITP's local_eligible flag — a plain true/false field — to false. The lesson read clean: AITP is headquartered in Appleton, it isn't a Milwaukee organization, so it shouldn't surface in a Milwaukee answer. Flag false. Done.
This week the same organization pulled me the other way — and the fact that one org can produce both calls is the whole point. Earlier this month I seeded an AITP event into the inventory by hand: the AITP June Tech Meetup, held in Pewaukee. (It's since passed, so you won't find it in the events further down the page — but the decision behind it is what matters.) By the strict letter of that same flag I set to false last week, the system should reject anything AITP touches. I overrode it and seeded the event anyway.
So which is it — does AITP belong, or not? Both, and that's not a dodge. The flag answers one question: is AITP a Milwaukee organization? No — its home is Appleton, so as an org it stays flagged out, exactly as I decided last week. The override answers a different question: is this particular event in metro Milwaukee? Yes — Pewaukee is metro, and my definition of "local" is the whole regional corridor, not the city limits. An Appleton organization can still run a metro event worth listing. The flag governs the org; my hand governs the event. Last week's exclusion and this week's inclusion were never a contradiction — they're one overworked boolean meeting two different questions it was never built to tell apart.
Here's the part worth being honest about: that override was easy. Could AI have made the same call? Yes — hand a model my rule ("local means Milwaukee metro, not city limits") and it will correctly reason that a Pewaukee event qualifies despite an Appleton headquarters. The model does the reasoning; I supply the standard. One event, one rule, one answer. Prompting handles that fine.
So that is not why I built the database. The database exists because of the other problem — the scrape — and that one prompting could never solve. I have a year of evidence. I spent it prompting frontier models to read scraped Meetup and Eventbrite listings and keep the real community events while rejecting the commercial filler. The output was the bloat you've watched me trim by hand, week after week. Not because the model couldn't reason about any single listing, but because "is this a genuine community event or a sales pitch wearing AI keywords?" has no rule I can write down. The standard lives in my judgment as pattern, not as a statable definition — so the model substitutes its own boundary, that boundary drifts from one listing to the next, and a bootcamp's polished copy reads as legitimate. Worse, the inputs are adversarial: marketers optimize their listings to pass for real events, so any heuristic I prompt into the model gets gamed. A fuzzy standard, applied at volume, against inputs designed to beat it — that is the losing game I finally stopped playing.
The grounded approach wins because it refuses to make the model judge at all. I didn't write a better "is this legit?" prompt. I replaced an unanswerable per-event judgment with a question that has a factual answer: is this event's organization in my vetted inventory? A fact doesn't drift, and it can't be gamed by better marketing copy. The judgment didn't disappear — it moved to the vetting step, where I decide once, deliberately, whether an organization earns a place in the inventory, instead of asking a model to re-litigate that call on every messy row it scrapes. AITP is the rare case where I reached in and overrode the vetting by hand. The bloat is the common case where I took the model out of the decision entirely.
That's the same argument the rest of this edition makes, aimed at a different target. Structured data — a flag, a trusted list, a vetted inventory — is about the closest thing there is to a natural enemy of AI hallucination, because it converts a judgment into a fact the system can enforce. A model can make a judgment I've fully specified. What it can't do is hold my standard consistently across scale and adversarial input — which is exactly why I stopped asking it to judge events and built something that answers a factual question instead. The reasoning can be the model's. The standard is mine — and so is the decision about when to take the model out of the loop. The database is the control surface. The curator is the authority. That's not a limitation of the system — it's the design.
The Learning Loop
DEFINITIONHallucination
Hallucination refers to when an AI model confidently generates information that is factually incorrect, fabricated, or entirely made up. It occurs because language models predict statistically likely text rather than retrieving verified facts, meaning they can "fill in gaps" with plausible-sounding but false details. Always verify AI-generated facts against authoritative sources, especially for medical, legal, or statistical claims.
Source: Wikipedia
TIPPersona Stacking
Layer multiple professional perspectives into a single prompt to get richer, more balanced output. For example: "Respond as both a seasoned UX designer and a data privacy lawyer reviewing this app feature." This forces the model to surface tensions and trade-offs between disciplines that a single-role prompt would miss, saving you multiple rounds of follow-up.
Source: PromptingGuide
TOOLOtter.ai
Otter.ai is an AI-powered transcription and meeting assistant that provides real-time voice transcription, automatic meeting summaries, action item detection, and speaker identification. It integrates with Zoom, Google Meet, and Microsoft Teams, making it ideal for professionals who need searchable, shareable records of conversations without manual note-taking.
“The strength of the team is each individual member. The strength of each member is the team.”
— Phil Jackson — Phil Jackson is a former professional basketball coach who won eleven NBA championships with the Chicago Bulls and Los Angeles Lakers. He is widely regarded as one of the greatest coaches in sports history, known for his use of mindfulness and team-centered philosophy. His leadership approach has been studied extensively in business, psychology, and organizational behavior.
The Nest Jest
Upcoming Events
Event06/23/2026 4:00 PM
Opening Event: Registration & Badge Pickup Opens — The Venue, 275 W. Wisconsin Ave., Milwaukee, WI
Join us for the opening event at Summerfest Tech! This panel and networking event will take place at 3rd Street Market Hall and their event space "The Venue" (275 W Wisconsin Ave Suite 100, Milwaukee, WI 53203). Discounted parking will be available onsite! Pre-register for this event through our website and your badge will be available to be printed and picked up onsite for ease and efficiency! You may also register onsite.
Summerfest Tech began in 2018 with the goal of leveraging the Summerfest brand to showcase technology and provide a forum for attendees from the tech and business community to network, educate, and celebrate the evolution of technology. Summerfest Tech continues to champion the regional business community, while shining the spotlight on Wisconsin’s growing reputation as a tech hub, highlighting the growth and opportunity in the region. Programming from local, regional, and national speakers, as
Opening Panel: From AI Hype to Operational Impact: How Wisconsin Companies Are Putting AI to Work — The Venue, 275 W. Wisconsin Ave., Milwaukee, WI
Location: 3rd Street Market Hall event space "The Venue" (275 W Wisconsin Ave Suite 100, Milwaukee, WI 53203)
This opening panel highlights how the Microsoft AI Co‑Innovation Lab is helping companies move from AI exploration to real operational impact. Operated in partnership with UW‑Milwaukee’s Connected Systems Institute and TitletownTech, the lab provides a hands-on environment for building and deploying applied AI solutions. Panelists will share examples from companies that have worked in th
Registration Opens & Breakfast — The Westin Milwaukee, 550 N. Van Buren St., Milwaukee, WI
Registration will be located on the first floor
Breakfast will be located outside of The Grand Ballroom on the third floor
Join us ahead of the keynote to explore our Exhibit Hall, networking with attendees, and snag a quick bite before the day!
Evening badge pickup available at Discourse at Radio Milwaukee from 5-7pm.
Keynote - Return on Intelligence: A Strategic Enterprise Playbook for Scalable AI Agents — The Westin Milwaukee, 550 N. Van Buren St., Milwaukee, WI
Keynote Presentation
Westin Grand Ballroom
Kristin will keynote on her book, “Return on Intelligence: A Strategic Enterprise Playbook for Scalable AI Agents.” The first 100 attendees will receive a free copy of Kristin's book and have the ability to have her sign the book and attend a meet and greet immediately following her keynote.
Speaker:
Kristin Milchanowski, Chief AI and Data Officer, BMO Financial Group
Advanced Manufacturing Track - Software-Defined Vehicles — The Westin Milwaukee, 550 N. Van Buren St., Milwaukee, WI
Rewriting the Rules of Mobility: Software‑Defined Vehicles, Built Through Engineering Innovation
The Westin Ontario/Erie rooms
Sponsored by Husco
Software‑defined vehicles (SDVs) are reshaping the future of wheeled machinery, where engineering‑led innovation and the people behind it are the true driving force. Built to evolve through software rather than remain locked at launch, SDVs unlock new capabilities, features, and performance over time. Off‑highway machines are entering a software‑defin
Ecosystem Builder - From Small Town to Scale-Up — 833 East, 833 E. Michigan Street, Milwaukee, WI
833 East Potawatomi Ventures Boardroom
From Small Town to Scale-Up: Midwest Startup Success Stories
Innovation is thriving across the Midwest—not only in major metros, but in smaller communities often overlooked in the national conversation. This session brings together founders who have built and scaled successful startups from small towns, navigating limited resources, attracting and retaining talent, and accessing capital beyond traditional hubs. Speakers will share candid insights on lever
MarTech Track - Build, Buy, or AI — The Westin Milwaukee, 550 N. Van Buren St., Milwaukee, WI
The Westin Milwaukee Michigan room
Build, Buy, Vibe or Agent?
The question of build or buy has gotten complicated in 2026 for marketers. Let's break down how to think about your options and provoke some discussion. In the process, it will become clear why I bet the farm on building my SaaS startup, Path3, while the media is spouting claims that SaaS is cooked.
Speaker:
Steve Robinson, Co-Founder & CEO, Path3
MarTech - Feedback to Innovation — The Westin Milwaukee, 550 N. Van Buren St., Milwaukee, WI
The Westin Milwaukee Michigan room
Listening in the Age of AI, What Are We Really Looking For?
Organizations have never collected more feedback or trusted it less. Response rates are falling and single-number metrics tell us how much people feel, not why. This session asks what we're really after when we say we want to "listen," and what genuinely changes when AI turns feedback from a form into a conversation. Drawing on PrismScope's conversational survey work, you'll leave able to separate r
Pitch Competition: Advanced Manufacturing, Food/Bev/Ag, & FinTech — Downtown Kitchen, 777 E. Wisconsin Ave. Galleria Level, Milwaukee, WI
Location : Downtown Kitchen
Meet our top ten Finalist companies competing across two pitch tracks: Manufacturing, Food/Bev, and AgTech; and FinTech. Hear from startup founders who are working on the latest and greatest innovations that will shape these sectors in the years to come, get the inside scoop on the latest tech trends and innovations, and find hot investment opportunities with startups that are poised to scale. Plus, cast your vote for the Fan Favorite in each category and help decide
FinTech Track - Building the AI-Enabled Firm — U.S. Bank Center, 777 E. Wisconsin Ave., Milwaukee, WI
U.S. Bank Center Lobby Conference Room
You Can't Spell B-AI-RD Without AI – Building the AI-Enabled Firm: Baird's AI Journey
At Baird, AI isn't a side project bolted onto IT – it's owned by business leaders, championed from the top, and governed at the board level. In this two-part session, Tim Byrne first shares the firm-wide playbook: how a four-year journey went from "first time we asked the question" to clear ROI woven into business-unit strategy, built on federated ownership, cross-functio
Keynote - UW–Madison: Shaping an AI-Driven Future — The Westin Milwaukee, 550 N. Van Buren St., Milwaukee, WI
Westin Grand Ballroom
Sponsored by University of Wisconsin-Madison
During an interactive panel discussion, three University of Wisconsin–Madison deans will share how AI is shaping the professional landscape and the future workforce. You’ll also learn how UW is leading in AI education, innovation, and thoughtful implementation—and giving today’s leaders the technical tools to sharpen their competitive edge.
Speakers:
Remzi Arpaci-Dusseau, Incoming founding dean, College of Computing and Artifi
Milwaukee Innovation District Bar Hop — The Westin Milwaukee, 550 N. Van Buren St., Milwaukee, WI
Innovation District/Walker's Point
Shuttle information:
5-6pm from the Westin to the 4 locations below
6-10pm between all 4 locations
Boone & Crockett
Sponsored by Bayforce
Boone & Crockett located at 818 S Water St, Milwaukee, WI 53204
Station 1846
Sponsored by Light the Hoan Powered by Color Kinetics
Station 1846 located at 215 W Bruce St, Milwaukee, WI 53204
Great Lakes Distillery
Sponsored by Concurrency
Great Lakes Distillery located at 616 W Virginia St, Milwaukee, WI 53204
Di
Summer Sideshow Concert Series — The Cooperage, 822 S. Water St., Milwaukee, WI
The Cooperage - 822 South Water Street Milwaukee, WI 53204
Milwaukee's best in rock/punk/emo head to the Harborfront for Summerfest Tech adjacent evening of fun and tunes!
A night of fun and music to run alongside the Summerfest Tech festival. Milwaukee's own Gold Step's, Outpacer, Barely Civil, Alley Eyes and Fakeout will bring the entertainment while Taco Moto provides the food and the vibes provided by you.
This one is free to Summerfest Tech attendees/badge holders. Doors at 6, Music at 7
Registration Opens & Breakfast — The Westin Milwaukee, 550 N. Van Buren St., Milwaukee, WI
The Westin Milwaukee
The Westin Milwaukee is located at 550 N Van Buren St, Milwaukee, WI 53202
Registration will be located on the first floor
Breakfast will be located outside of The Grand Ballroom on the third floor
Join us ahead of the keynote to explore our vendor tables, networking with attendees, and snag a quick bite before the day!
Just before the keynote, Ryan Baumann, VP of Direct Sales, Power Systems will be representing Rehlko
Keynote - Southeast Wisconsin and Microsoft: Defining a Template for Durable Tech Ecosystems — The Westin Milwaukee, 550 N. Van Buren St., Milwaukee, WI
Westin Grand Ballroom
General Manager of Datacenter Operations for Microsoft’s U.S. Central Region will review the capabilities of our AI super factory here, will discuss how southeast Wisconsin is uniquely positioned to be a tech leader and how it fits into Microsoft’s growth.
Speaker:
Christa Markgraff, Central Region Leader, Data Center Operations AMER, Microsoft
FinTech Track - AI Security Strategy — U.S. Bank Center, 777 E. Wisconsin Ave., Milwaukee, WI
U.S. Bank Center Lobby Conference Room
AI Security Strategy
Sponsored by Guidepoint
Fintech is one of the most targeted industries for email-based attacks. BEC, phishing, and vendor fraud are constant threats against teams that cannot afford slow response times. In this session, GuidePoint Security, Sublime and Torq break down how AI is transforming security operations. Sublime covers how modern email threat detection works, and Torq shows how AI-driven automation closes the loop from detectio
Advanced Manufacturing Track - Industrializing AI — The Westin Milwaukee, 550 N. Van Buren St., Milwaukee, WI
The Westin Milwaukee Ontario/Erie rooms
Industrializing AI Across the Manufacturing Enterprise
Sponsored by Slalom
AI is rapidly reshaping how manufacturers operate, compete, and make decisions across the enterprise. From supply chain and operations to customer experience and commercial strategy, organizations are looking for practical ways to move beyond experimentation and apply AI where it can deliver measurable business value. In this session, Slalom will share perspectives on the emerging
Ecosystem Builder – Collaboration Over Competition — 833 East, 833 E. Michigan Street, Milwaukee, WI
833 E Michigan Building Conference Room
Collaboration Over Competition: Building Regional Startup Corridors
Sponsored by WEDC
The next great startup ecosystems won't be built by a single city - they'll be built by regions that learn to act like one. But that kind of coordination doesn't happen organically. It requires intentional architecture: investors who see their role as ecosystem builders, institutions that treat talent development as an economic output, and cities willing to compete on c
MarTech Track - AI Brand Visibility — The Westin Milwaukee, 550 N. Van Buren St., Milwaukee, WI
The Westin Milwaukee Michigan room
To AI, You're Average.
To an AI, your brand is a number. It's the average of everything the internet has said about you, and most brands look just like their competitors. Ask AI for the best companies in your category. It names a few brands. Most companies have never checked whether they're one of them, or whether what it says is even right. And a lot of what shapes that answer lives outside your marketing team's control. Tyler Einberger co-founded Momentic an
Founder Feature – Beyond the City Bias — 833 East, 833 E. Michigan Street, Milwaukee, WI
Potawatomi Ventures Boardroom
Beyond the City Bias: Rural Entrepreneurship as Strategy
Innovation doesn’t belong to cities—or universities. It happens wherever people are solving real problems. Rural communities are quietly building strong, impactful businesses, offering a powerful counterpoint to urban-centric models of entrepreneurship. This session explores how rural founders lead with tenacity, resilience, and place-based values, leveraging trusted networks, lean operations, and access to m
FinTech Track - Community Lender Transformation — U.S. Bank Center, 777 E. Wisconsin Ave., Milwaukee, WI
U.S. Bank Center Lobby Conference Room Featured Session
This Is the Community Lender's AI Moment
Sponsored by Potawatomi Ventures
Understand it. Execute Correctly. Own the moment.
The technology gap between big banks and community lenders just closed. This session explains what AI actually is, what it takes to execute correctly, and why community lenders are uniquely positioned to capitalize right now.
Speaker:
Aaron Colcord, Founder and CEO, Voyager AI
Healthcare/BioTech Track - Actualizing Biohealth Careers — The Westin Milwaukee, 550 N. Van Buren St., Milwaukee, WI
The Westin Milwaukee
Actualizing Biohealth Careers
Sponsored by ABC Pathways
Join representatives from MATC, Madison College, Employ Milwaukee, the Workforce Development Board of South Central Wisconsin, and the University of Wisconsin-Milwaukee as they showcase how the Actualizing Biohealth Career (ABC) Pathways initiative is strengthening Wisconsin’s biohealth workforce pipeline. ABC Pathways is one of five strategic projects supported by the Wisconsin Biohealth Tech Hub, an initiative funde
FinTech Track - BMO Banking Panel — U.S. Bank Center, 777 E. Wisconsin Ave., Milwaukee, WI
U.S. Bank Center Lobby Conference Room
Information Coming Soon
Speakers:
Andrew Harrison, Head - Emerging Partnerships & Digital Innovation Fund, BMO Financial Group
Amy Chou, Co-Founder and CEO, Additional Wealth
Shannon McGhee, CEO, SEED SPOT
Ecosystem Builder – From Red Tape to Real Opportunity — 833 East, 833 E. Michigan Street, Milwaukee, WI
833 E Michigan Building Conference Room
From Red Tape to Real Opportunity: Navigating Government Collaboration
Building strong partnerships with government agencies can unlock powerful opportunities for entrepreneurs—from funding and pilot programs to new markets. This session dives into how founders can navigate public-sector relationships effectively. Learn how to identify the right partners, align with community goals, and create collaborations that drive both business growth and public good
Healthcare/BioTech Track - Scaling Clinical Intelligence — The Westin Milwaukee, 550 N. Van Buren St., Milwaukee, WI
The Westin Milwaukee Huron room
Scaling Clinical Intelligence in Patient Physiological Monitoring and Therapy
Scaling Clinical Intelligence charts the shift of patient monitoring from isolated boxes to a patient-centric ecosystem. The Internet of Medical Things (IoMT) anchors this vision with a cloud-extended, wireless, and wearable architecture connecting devices, edge compute, and AI at enterprise scale. A hybrid compute model distributes containerized services from patient room to cloud, all
Pitch Competition: Healthcare/Biotech & Disruptive Tech — Downtown Kitchen, 777 E. Wisconsin Ave. Galleria Level, Milwaukee, WI
Downtown Kitchen
Meet our top ten Finalist companies competing across two pitch tracks: Healthcare/Biotech and Disruptive Tech. Hear from startup founders who are working on the latest and greatest innovations that will shape these sectors in the years to come, get the inside scoop on the latest tech trends and innovations, and find hot investment opportunities with startups that are poised to scale. Plus, cast your vote for the Fan Favorite in each category and help decide which lucky entrepren
FinTech Track - Governed AI Integration — U.S. Bank Center, 777 E. Wisconsin Ave., Milwaukee, WI
U.S. Bank Center Lobby Conference Room
Beyond the Chatbot: Securely Connecting AI to Enterprise Data and Workflows
Most people are already familiar with AI chatbots, but the real value of artificial intelligence emerges when it can securely interact with an organization’s internal data, systems, and business processes. By connecting AI to knowledge bases, business applications, and operational workflows, organizations can dramatically improve employee productivity, accelerate decision-making
Keynote - Power Generation and Grid Resilience in the Era of AI and the Gigawatt Data Center Explosion — The Westin Milwaukee, 550 N. Van Buren St., Milwaukee, WI
Westin Grand Ballroom
Sponsored by Generac Power Systems
Erik Wilde from Generac Power Systems will discuss the unprecedented demand for backup power and energy solutions for data centers, how data center growth expands manufacturing in the Midwest and how access to reliable power is more critical than ever before.
Speaker:
Erik Wilde, Executive Vice President, Generac Power Systems
Summerfest Music Festival — Henry Maier Festival Park, 200 N. Harbor Dr., Milwaukee, WI
Henry Maier Festival Park
Thursday Headliners
7:00PM | The Mountain Goats @ Aurora Pavilion
7:30PM | Ed Sheeran, Myles Smith, and Aaron Rowe @ American Family Insurance Ampitheater
NOTE: A separate ticket is needed to see concerts in the American Family Insurance Amphitheater, but your Amphitheater ticket also includes admission to Summerfest on the day of your show
9:00PM | The Roots @ BMO Pavilion
9:30PM | Subtronics @ Generac Power Stage
9:30PM | Halestorm @ Uline Warehouse Stage
10:00PM | K
Summerfest Tech Week – Private Founder & Investor Mixer — BMO Tower, 790 N. Water St., Milwaukee, WI
BMO Tower Pre-Registration Required
Summerfest Tech Week – Private Founder & Investor Experience
Sponsored by BMO
BMO Tower 790 N Water St, Milwaukee, WI 53202
This exclusive gathering brings together a select group of founders, investors, operators, and ecosystem leaders for a focused, high-value evening centered on strategic access, venture intelligence, and meaningful outcomes.
The evening is designed to foster impactful dialogue around:
Founder–investor alignment
Strategic growth and s
Registration Opens — The Italian Community Center, 631 E. Chicago St., Milwaukee, WI
The Italian Community Center
Badges will be available for pickup before during the marquee keynote and panel. This will be the last opportunity to pickup your badge!
Marquee Keynote - Bring Impossible Ideas to Life — The Italian Community Center, 631 E. Chicago St., Milwaukee, WI
The Italian Community Center All-In or Marquee Badge Required
Technology & AI are dramatically changing what’s possible in our world. Our challenge as we create new things is to dislodge from the comfort of what we already know. Chris created Shazam after being told by every “expert” that his outlandish idea was impossible. It was an idea far ahead of its time – eight years before iPhone apps even existed. In this presentation, Chris helps audiences imagine and then create new visions for the fu
Closing Panel: From Vision to Execution: Making Innovation a Business Strategy — The Italian Community Center, 631 E. Chicago St., Milwaukee, WI
The Italian Community Center All-In or Marquee Badge Required
Event for Marquee and All-In tickerholders
An emerging theme of the conference is the growing recognition that technology and innovation are no longer simply an IT function, but a core business strategy shaped by leadership. This session will follow the marquee keynote from Chris Barton, Founder and Former CEO of Shazam, who will explore how bold ideas and purposeful AI can unlock transformative innovation across entire organizations
Networking Event — Lasso Lounge, 100 N. Harbor Dr., Milwaukee, WI
Lasso Lounge, Summerfest
Sponsored by University of Wisconsin-Milwaukee
All Summerfest TechAI badge holders are welcome! Must have picked up badge prior at these registration dates, times, and locations:
June 23 4-7p at Venue 3 at Third Street Market Hall
June 24 & 25 9a-5p at The Westin Downtown Milwaukee
5-7p at Discourse Coffee — Radio Milwaukee
June 26 9-11a at Italian Community Center
Summerfest Music Festival — Henry Maier Festival Park, 200 N. Harbor Dr., Milwaukee, WI
Henry Maier Festival Park
Friday Headliners
7:00PM | Sudan Archives @ Aurora Pavilion
7:00PM | Cody Johnson, and Jessie Murph @ American Family Insurance Ampitheater
NOTE: A separate ticket is needed to see concerts in the American Family Insurance Amphitheater, but your Amphitheater ticket also includes admission to Summerfest on the day of your show
9:00PM | Louis Tomlinson @ BMO Pavilion
7:30PM | Big Wild @ Generac Power Stage
9:30PM | The Revivalists @ Uline Warehouse Stage
10:00PM | Petey
Founders Day — Ward4, 313 N. Plankinton Avenue, Milwaukee, WI
Milwaukee Startup Founders – Founders Day was designed for YOU! We know the entrepreneurial journey can be tough, so we’re hosting a full day of free coworking, mentorship, networking, and more.
Join a network of founders and mentors who come together to foster collaboration, share insights, and help grow the startup space in Milwaukee.
Founders Day — Ward4, 313 N Plankinton Avenue, Milwaukee, WI 53203
Monthly networking and programming for Milwaukee's startup and technology community, hosted by MKE Tech Hub Coalition at Ward4. Recurring last Wednesday of the month.
Source: MKE Tech Hub Coalition (mketech.org)
In the News
News2026-06-21 — TechCrunch
Nobel Laureate John Jumper Is Leaving DeepMind for Rival Anthropic
Nobel laureate John Jumper has announced his departure from Google DeepMind to join Anthropic, marking a significant talent shift in the AI industry. Jumper is reportedly not the only high-profile name leaving Google DeepMind, suggesting broader organizational challenges at the lab. The move highlights intensifying competition among leading AI research organizations for top scientific talent. read more
News2026-06-21 — Wired
The White House Is Making Up Its Rules for AI in Real Time
Anthropic remains unable to distribute its Claude Mythos and Fable 5 models after running into conflict with the Trump administration, yet no clear explanation has been provided for what the company did wrong. The situation underscores a lack of defined regulatory framework governing AI model distribution and export controls. The ad hoc nature of the administration's AI policy decisions is creating uncertainty for AI developers navigating compliance. read more
News2026-06-21 — VentureBeat
Google Just Redesigned the Search Box for the First Time in 25 Years
Google has formally announced a sweeping redesign of its iconic search box at its annual I/O developer conference, retiring the traditional paradigm of a simple text input and blue links after 25 years. The redesign reflects Google's broader push to integrate AI-native experiences directly into its core search product. The change signals a fundamental shift in how users are expected to interact with information retrieval powered by large language models. read more
Tools from Hard AIs
AI Airfare Research
Find cheaper flights using an AI-powered research tool
built by Hard AIs.