Journal

The Death of Software as a Service (SaaS)

Denver AI, a local group focused on moving AI out of theory and into execution; bringing together operators, founders, and builders to share real-world use cases, demos, and practical approaches to applying AI inside...

War of The Claws

In 1938, a young Orson Welles captivated and terrified radio listeners with a fictional Martian invasion, inciting a real panic across the nation. Last weekend, more of my family and friends called be about playful AI...

Design: Once a Job, Now a Power Everyone Can Access and Use

In the early days of aviation and space exploration, the word “computer” didn’t refer to a machine. It referred to a person.

Which workflow should you build first?

Bring one workflow, its owner, and three recent examples.

Choose a workflow your team repeats, whose current cost you can describe, and whose owner can help test a first version.

The back office is the bottleneck: applied AI for community lenders

Mission lenders were built for underwriting judgment, but their week goes to packaging, compliance, and impact reporting. Separate the rules from the judgment and adding a program stops meaning adding headcount.

CDFIs exist to make the judgment call banks won't — yet most of their capacity goes to manual packaging, compliance, and funder reporting. An LOS digitizes the form; applied AI absorbs the rules-heavy work on either side of it, keeping the judgment human.

When the note-writing outlasts the visit: applied AI for behavioral-health teams

Documentation has quietly become the job in community mental health. Separate the rules from the judgment and human-in-the-loop AI gives clinicians their hours back — without touching the record's integrity.

Community behavioral-health clinicians spend nearly as long documenting a session as delivering it, right as budgets cut the admin staff who used to buffer them. The fix isn't a bolt-on scribe — it's redesigning the workflow around a rules-vs-judgment split with a human in the loop.

When the backlog is the bottleneck: applied AI for high-volume legal services

Legal-services and immigration-aid organizations are drowning in intake, not law. Here's where AI actually helps — and where a human must stay in the loop.

In high-volume legal aid the constraint is rarely the law — it's the intake, eligibility, and document work upstream of the lawyer. Separate the rules from the judgment and applied AI clears the backlog without automating the calls that must stay human.

When your best asset is judgment that lives in people's heads

Every expert weather and climate-risk firm is built on judgment and haunted by the same ceiling: it scales with the expert's hours. When the founder is on a plane, the firm's best product is offline.

AI weather models are everywhere, but they don't carry the judgment. The forecasting firms that thrive won't have the biggest model — they'll be the ones whose expert judgment finally scales past one person's calendar, with the human firmly in the loop.

The trades don't have a demand problem. They have a throughput problem.

600,000 open trades jobs, 150,000 new entrants a year, and grant money flowing. So why does it feel stuck? The orgs meant to fix it run modern mandates on case-manager-era operations.

The demand and the funding are already here. The variable left is throughput — and the workforce orgs that scale will be the ones that give overstretched staff their hours back with applied AI, while keeping human judgment on the decisions that change lives.

The hardest part of scaling a defense-hardware company isn't the hardware

Investors stopped paying for prototypes and started paying for throughput. Most defense-hardware teams have a shop floor and a group chat — not a coordination layer.

The money moved from prototypes to throughput. The teams that win the ramp are the ones that build the coordination layer above the factory floor — with applied AI applied where judgment is expensive, and a human in the loop where being wrong is costly.

Cheap Software Didn't Solve Your Platform's Problem. It Moved It.

AI made software cheap. What is scarce now is the builder who can turn a partner's need into shipped, trustworthy product.

Cheap software did not solve the platform's problem, it moved it. The scarce layer now is the forward-deployed builder who turns a partner's need into working, trustworthy software fast enough to close the deal.

Veteran workforce training built the program. AI removes the ceiling on delivery.

The veteran workforce sector's bottleneck isn't the curriculum. It's a delivery model that scales one hire at a time.

Veteran training orgs have real programs, real funding, and a population ready to learn fast. What they don't have is a delivery model that scales without headcount. Here's the two-day move that removes the ceiling.