DEV Community: David Ohnstad

Enterprise AI Governance: Why Post-Deployment Fails

David Ohnstad — Mon, 20 Jul 2026 10:58:27 +0000

Most organizations celebrate their ML deployment—then face unexpected costs and performance failures within weeks. David Ohnstad breaks down the post-deployment governance gap that catches enterprises off guard and provides the framework to prevent it.

Originally published on David Ohnstad

ML Models in Production: Why Enterprise AI Projects Stall

David Ohnstad — Mon, 20 Jul 2026 10:58:06 +0000

Six months in, fourteen ML models were live—but none had automated retraining. Three still used hardcoded paths from a laptop. When compliance flagging hit 40% false positives, an eleven-day investigation traced it back to an undocumented schema change. This is where enterprise AI projects actually break.

Originally published on David Ohnstad

BWCA Closures: How to Plan When the Forest Service Shuts Down

David Ohnstad — Mon, 20 Jul 2026 10:57:25 +0000

When Superior National Forest closures threaten your permits and food prep, what’s your move? David Ohnstad breaks down the real framework that turned a July closure announcement into a salvaged trip—and how to apply it next time.

Originally published on David Ohnstad

Minnesota Outdoors Guide: Planning Adventure When Conditions Change

David Ohnstad — Mon, 20 Jul 2026 10:57:04 +0000

Jake’s six-month permit meant nothing when the Forest Service closed 55,000 acres. Minnesota’s wilderness doesn’t follow your spreadsheet. Learn how to build outdoor plans flexible enough to survive the state’s unpredictability—from BWCA closures to sudden weather shifts.

Originally published on David Ohnstad

Woodworking Joints: Why Strength Isn't Always Better

David Ohnstad — Fri, 17 Jul 2026 06:57:42 +0000

The woodworking internet preaches that bombproof joinery like mortise-and-tenon and through-dovetails are always superior. But this hierarchy ignores a crucial truth: the right joint depends on your application, not just raw strength.

Originally published on David Ohnstad

Woodworking Joints: Why Popular Choices Often Fail

David Ohnstad — Fri, 17 Jul 2026 06:57:41 +0000

Woodworking internet culture pushes dovetails and mortise-and-tenons as universal solutions. But experienced builders know a dado joint cut in seconds often outperforms techniques taking forty minutes. Learn when to break the rules.

Originally published on David Ohnstad

Why ML Models Fail in Production: Beyond POC

David Ohnstad — Fri, 17 Jul 2026 06:53:56 +0000

A 91%-accurate predictive maintenance model built in four months sits unused six months later. The problem wasn't the algorithm—it was missing pipelines, documentation, and ownership. Enterprise ML stalls not because science fails, but because infrastructure and operations do.

Originally published on David Ohnstad

Data Mesh Implementation: Why Platform Thinking Fails

David Ohnstad — Fri, 17 Jul 2026 06:53:55 +0000

Most data mesh initiatives fail because organizations build platforms instead of products. Fannie Mae's success with Amazon Redshift proves that treating data domains as products—with clear ownership and cross-functional teams—is the difference between stalling and shipping.

Originally published on David Ohnstad

Summer Woodworking: Why Hot Weather Helps Your Projects

David Ohnstad — Fri, 17 Jul 2026 02:21:03 +0000

Most woodworkers avoid their shops in summer heat. But faster glue curing, dust-free outdoor spraying, and perfectly timed projects for outdoor use make these months your most productive season.

Originally published on David Ohnstad

System Drift: Why High-Performing Organizations Slowly Lose Effectiveness Without Realizing It

David Ohnstad — Fri, 17 Jul 2026 02:20:00 +0000

Sustained performance may depend less on adding new capabilities and more on preventing valuable systems from quietly drifting off course.

Originally published on David Ohnstad

Leadership Skills Gap: What 2026 Data Really Reveals

David Ohnstad — Fri, 17 Jul 2026 02:19:37 +0000

Harvard’s ‘power skills’ argument sounds compelling until you examine what’s actually breaking in organizations. Gartner data reveals 68% of mid-level managers report direct reports lack critical tech competencies. The real leadership challenge isn’t soft skills—it’s the gap between what organizations teach and what they actually need.

Originally published on David Ohnstad

Enterprise AI Success Metrics: Why Most Projects Fail

David Ohnstad — Fri, 17 Jul 2026 02:19:14 +0000

A 94% accurate model means nothing if nobody defined what business outcome it should drive. David Ohnstad exposes the critical gap between data science validation and executive expectations that derails nine months of enterprise AI development.

Originally published on David Ohnstad