The Toolchain Upgrade That Compounds

There's a category of upgrade sitting somewhere on your list. Not the speculative "I should learn Rust someday" item. The concrete one: switch to uv for Python package management, move your SSH keys to the TPM chip, set up WireGuard instead of the VPN you're tolerating. Each one is an afternoon. Each one pays indefinitely.

And each one has been sitting there for six months.

---

The Math Nobody Does

The "Why Aren't We uv Yet?" post makes the case almost embarrassingly simply. uv is 10–100x faster than pip at package resolution. If you install or update packages daily — even once, on slow hardware — the compound savings over a year are measured in hours. Real hours. Hours you spent watching a spinner that didn't need to be spinning.

The SSH key in TPM post has the same structure. One afternoon to configure. The private key never leaves the hardware. No ongoing maintenance cost — just ongoing security benefit.

WireGuard on Windows just cleared a signing dispute with Microsoft. For users running workarounds: it's now a 10-minute reinstall for a cleaner, faster, officially-sanctioned setup.

Three tools. Three afternoons. Indefinite improvement.

So why haven't you done it?

---

The Friction You Stopped Noticing

The usual answer is switching cost. And it's real — the time to do it, the cognitive overhead of learning new behavior, the risk of something breaking at a bad moment. Switching cost isn't imaginary.

But we systematically overestimate it, and we compare it to the wrong baseline.

We compare switching cost to the effort of staying put, which registers as zero. The VPN takes 10 extra seconds to connect? That's not a cost — that's just Tuesday. The package install takes 40 seconds instead of 4? You're used to it.

The coherenceism frame that applies here is alignment over force. When your tools are well-suited, the workflow carries the work. When they're misaligned, you're managing constant friction — small losses, continuously, invisibly. Forcing work through a narrower channel than necessary.

That friction doesn't feel like a cost because it's normalized. It's just what working feels like.

---

Build Once, Use Forever

The "Build Once, Use Forever" principle is about infrastructure — tools and configurations you establish once and route all future work through. Toolchain upgrades are the clearest application of this principle that exists.

The question isn't "is this worth an afternoon?" — it obviously is. It's "why is this afternoon the right moment?"

The trap: there's no wrong moment. Which means you can always defer to when things are less busy. Things are never less busy. The deferred upgrade compounds a cost you never see on any dashboard, because it manifests as absence — the seconds lost, the breach that didn't happen, the mental load of managing a tool you've never fully trusted.

The minimum useful intervention is smaller than the inertia suggests. For uv: one install command, update your virtualenv setup. For TPM SSH keys: one keygen command with the right flags, two lines in your SSH config. For WireGuard: reinstall, done.

---

What Compounds

Every tool in your chain has a downstream effect on your cognitive load. A faster package manager means you stop building a callus around "waiting for dependencies." You just proceed. SSH keys in hardware means you stop thinking about key management. The machine holds them.

These aren't dramatic improvements. They're the quiet kind — the ones you notice six months later when you realize you haven't thought about that problem class in a while.

That's the compound interest of a well-configured toolchain. Not productivity theater. Friction removed, permanently, from your daily path.

The upgrade on your list has been sitting there long enough.