Our principles, and why they matter now

There are a handful of moments in a technology era where the prevailing wisdom turns out to be wrong. We think we’re living through one of them.

The current consensus says you should hand your data to a SaaS vendor, accept their pricing pivots, deploy their AI features the moment they ship, and trust that the platform will be stable, that the security will hold, and that the line item won’t quietly triple at renewal.

We disagree. And we think the consequences of the prevailing wisdom are already showing up in your inbox and on your status pages.

What we’re seeing

Two trends are running in parallel, and they’re getting worse together.

First, product quality and reliability are falling off a cliff. AI-accelerated development cycles have made it cheaper to ship, and shipping faster does not, it turns out, mean shipping correctly. We’re seeing more regressions, more outages, more “we’re aware of an issue” tweets, more silent failures, more silly bugs in products that used to be solid. The reliability tax has been passed from the vendors to their customers, and the customers are paying it in lost revenue and operational burn.

Second, security vulnerabilities are multiplying. AI-assisted coding produces code that looks right and is wrong in subtle ways. The SaaS supply chain expands the attack surface every quarter. Credential-stuffing, session-hijacking, and ransomware attacks are easier and cheaper to run than ever, and the defenders are working with the same tooling that’s accelerating the attackers.

These are not isolated incidents. They’re a pattern. And the pattern is going to get worse before it gets better.

What we’re betting on

Our response is to lean the other way. Not because we’re anti-AI, AI is doing real, useful work, and refusing to acknowledge that would be silly. We’re against the structural posture the industry has adopted: hand over your data, accept the lock-in, hope the vendor stays solvent and competent.

So we anchor everything we do, every recommendation, every system we build, every platform we operate, to four principles.

1. Data ownership

You own your data. You own your encryption keys. You own your audit trail. When a vendor pivots their pricing, you have leverage. When a vendor gets acquired and the new owners shut down a feature, you can walk. When a vendor gets breached, you’re not in their blast radius.

Data ownership is the most underrated form of business insurance in modern operations. It’s also the most direct hedge against vendor price swings, when you control the data, you control whether the vendor’s pricing power applies to you.

2. Open source

We build on software that is open: readable, forkable, and durable past any one vendor’s lifespan. This is not ideology, it’s risk management. Open-source software cannot be unilaterally relicensed against you, cannot quietly add telemetry, cannot disappear because a parent company restructured. You can audit it. You can fix it. You can run a fork yourself if you need to.

There are excellent commercial products built on top of open-source cores, and we use them. But the substrate stays open.

3. Self-hosted

Your workloads run on infrastructure you control. Not in a “managed multi-tenant cluster” you can’t inspect. Not in someone’s serverless platform with surprise egress fees. On servers, in containers, in a configuration you can read.

Self-hosting is harder than clicking through a SaaS signup, for about a week. After that, it’s the same operational discipline, except the bill stops surprising you and the data stops leaving the building.

4. Decentralization

No single vendor as a single point of failure. Portable architectures, well-defined boundaries, escape hatches at every layer. When a platform pivots, gets acquired, sunsets a feature, or just becomes unreasonable, you should have somewhere to go that doesn’t require a rewrite.

This is the deepest principle of the four. Centralization is seductive, it’s simpler in the short term and almost always cheaper at the start. It’s also the structural condition that makes the first three principles necessary in the first place.

And underneath the principles: the values

Principles describe what we build. Values describe how we work.

We work in relationships, not transactions. Engagements are small, contact is direct, and we don’t have an account manager layer between us and you. Most of our clients have been our clients for years; we like it that way.

We work with honesty. If we think you’re about to make an expensive mistake, including one we’d directly benefit from, we’ll tell you. The only way the relationship works long-term is if both sides trust each other to call it straight. So we do.

And we believe in decentralization at the operating level too. Small teams, distributed work, no gatekeepers. The way we operate is the way we recommend.

The boring stack

When it comes time to ship something, the toolchain is boringly conventional: Linux, Docker, Kamal for deploys, Kamal Proxy for routing and TLS. No Kubernetes by default. No managed-Lambda lock-in. No “platform” obscuring what’s actually running.

This stack is chosen because it’s small enough to fit in one head, atomic and rollbackable by default, and stable enough to leave alone for weeks. Most outages happen during deploys or right after them, Kamal’s design eliminates entire classes of those.

Stability and reliability are not features we add at the end. They’re the product.

What this means for you

If you’re a business deciding where to put your trust over the next five years, here’s our pitch: bet on the side of the trade that the consequences of the current era are favoring.

Own your data. Run on infrastructure you can read. Pick tools you could survive without. Build relationships with people who’ll tell you the truth.

We’re here when you’re ready to talk.