Salesforce, Python, SQL, & other ways to put your data where you need it

Minnebar 2026 recap

04 May 2026 🔖 professional development
💬 EN

Table of Contents

Session notes from the Minnebar conference’s 20th anniversary, May 2, 2026 below.

Photo editing

Editing Photos On Your Phone with No Apps, by John Wilson.

He really did cover what every photos app setting does! I learned about the vertical and horizontal straighteners built into the iPhone, for example. I’d only used the rotational straightener, but it turns out you can also un-trapezoid that picture you took of a painting from not-straight-on.

Plus, we learned basic “photography 101” composition and lighting jargon, and why to care.

Tech and ICE

What tech lessons should we learn from the ICE invasion?, by Paul Cantrell.

A great antidote to software brain, pointing out that while noodling around with computers is fun, and “research and development” might one day turn out useful, neither is inherently “helping people.” To help people, approach them and ask what they need, and then do that, even if it’s not your hobby.

All tech problems are user experience problems, it seems – most of what was actually relevant for techies to help with, on the ground, was reducing usability friction in tech suddenly being leveraged for volunteer coordination, to ensure that nontechnical retirees with ample free time could participate without compromising the safety of themselves or of the immigrants they hoped to helped. There was very little “new technology” to create (pretty much just a map and a few simple data analytics visualization charts to accompany it).

Pricing your worth

Pricing Office Hours: How (and How Much) Should You Charge by Garrick Van Buren.

No question destabilized him – always a good answer to off-the-wall ones.

Also, he was holding office hours about pricing; he was not lecturing how to price the act of holding office hours. I was not the only person to misunderstand that. 🤣 Either one would’ve been cool, but I liked what it actually was even better.

Instructional design and inclusion

Learning Together to Build LLMs at Scale with Open-Source Coursework by Sona Maniyan and Jason A. Grafft.

Sona and Jason addressed challenges around organizing an enterprise study group of very academically challenging material, when everyone’s super busy.

A fascinating tidbit they found as they combed research about how humans learn: apparently, on average, humans get a bit shy about being wrong in front of other people, starting around our 40s, and it gets worse as we age. So they had to account for that. Another factor that pushes youth to be more bold, get up in front of a group, and say “stupid” thing and ask “stupid” questions is that they can’t afford to hire the elder “panel of experienced experts” sitting before them with their own cash – “engagement-seeking” boldly and potentially getting steered toward a more correct answer by elder peers, for free, is the deal of the century on early-career training that could have decades of return on investment.

On the other hand, when asked how LLMs might help, they found that it usually just helped them design their next moves as study-group organizers. It turns out that while novices are awesome at asking questions of other humans, they often get very little out of asking machines questions because they don’t yet know the right questions to ask.

Linguistics nerd programmers and crayons

Every Corpus Was a Body Once: Putting the (body) language into LLMs by Kelly Heitz

You’d better bet I didn’t miss this one.

I got a fun crayon partner for back-to-back drawing. “Lorem ipsum chicken chicken” forever, Kevin. 👊

Two cool points from Kelly’s session:

  1. She’s beginning to hear people romanticize the alleged “hard work” of StackOverflow to warn of the dangers of LLMs for newbies just learning, but really, those of us who learned to code once StackOverflow already existed were incessantly warned by older developers, who’d had nowhere to learn from but paper books and product documentation (Kathy Sierra called it “just-in-case” learning), that we weren’t really learning anything when we got help on forums (“just-in-time learning”), but merely mindlessly cloning “copypasta.” (I own a laptop sticker with the sarcastic “O’Rly” book cover to rebelliously flaunt my pride about how much I actually did learn that way.) In general, Kelly suspects (and excellent book “Index, A History of the” would probably agree), we humans seem not to trust newcomers to effectively use a brand new time-saving tool as an entry path into good learning-journey destinations (despite history repeatedly showing that actually, humans are curious, and newbies totally figure out how to start from the time-saved place and still learn a lot).
  2. A fellow attendee shared that being forced to code by throwing prompts and prayers at a nondeterministic system was taking all of the psychological safety out of their job description they’d so loved before LLMs started getting mandated. Authoring deterministic code had felt like a safe retreat from the cognitive toil of having to write emails, and this was making all day that bad. Yikes, and I’d thought LLMs were making me tired at the end of each day. I can’t imagine how painful it’d be to be under some sort of nondeterministic-LLM-generated code minimum, if programming were your way to afford food on the table and a roof over your head because it offered a path to excellence through many hours a day spent on deterministic activities. Now I want to ask around if this is becoming a severe employment accessibility / disability justice problem for, say, neurodivergent people whose brains best “rest” through plentiful breaks for deterministic repetition. 🤔

AI-generated video

AI Films Are Real. Here’s How I Use LLMs, Diffusion Models & Voice AI to Build an End-to-End Film Pipeline by Venkata Sai Praneeth Sirigiri.

Honestly, I mostly went to this one to check out the competition for a musician friend. Figure out where LLM-based generative AI in artmaking is going, so perhaps people who prefer to work without it can steer clear and keep finding an income in other aspects of the field.

Ironically, the most bleeding-edge-exploration session I went to had audiovisual problems. But being a professional creative before being a professional technologist showed, though, and Sai looked across the room, said, “Oh, there’s a whiteboard,” and gave us an old-school classroom lecture until A/V help arrived. It was lovely and meditative at the end of a long day.

The talk didn’t end up offering a glimpse into which skills are taking over which corners of the paying industry, but apparently, people who prompt-generate movies organize themselves into film festivals. The more things change, the more they stay the same.

Also, prompt-based filmmaking, for people who like storytelling through film, reminds me a lot of the relationship experienced coders and other software folks are having to LLM-based generative AI: the ability to noodle around and quick-sketch ideas, alone, in ways that otherwise would’ve cost prohibitive amounts of other experts’ time, and therefore probably never would’ve made it out of a diary and into a mockup at all.

I mean, are we all going to go up in flames because we software developers and apparently also filmmakers want to stretch our creative-spirit wings? Sigh, probably. But for 40 minutes, it was fun to set that aside for a moment and immerse into the delight of “making.”

Security for end-to-end testing

End-to-end tests considered harmful (securing credentials for E2E and synthetic testing) by me.

You should talk next year

Want to make sure you get a ticket? Watch for the 2027 session submissions to open.

You don’t even have to present about tech, if you’re not ready to, and feel more comfortable/engaged talking about something else. A fantastic session I attended years ago was a primer on Korean skincare products.

There’s no judged “call for proposals” process to worry about putting your heart and soul and confidence into – just an open signup form. They cut off signups when they have as many speakers as there are timeslots multiplied by conference rooms on the first floor of Best Buy headquarters. Otherwise, if you volunteer to speak, congratulations, you’re a speaker.

--- ---