Where does the "DX movement" go in a Post ZIRP world?

Zero Interest Rate Policy (aka ZIRP) was the defining economic policy in the States from 2008 until 2022: Money was stupidly cheap to borrow and as a result easy to invest in tech. ZIRP was one of the main drivers of hypergrowth tech in the last decade, and now it's over. Today, instead of "easy money", tech investments need to show returns on shorter timescales.

A lot of our engineering practices are going to change post-ZIRP. No more throwing money at problems or hiring our way out of them. There will be fewer "hyperscaler" startups and many, many fewer unicorns. Tech companies now have to show profits and growth without lighting investor money on fire.

When this change began last year, the first reaction across the industry was belt-tightening: many (most?) tech companies conducted layoffs to cut labor costs. Everyone in tech is also hiring slower, cutting and renegotiating vendors, and slowing down investments in longer-term projects.

During the ZIRP good times, developers were ridiculously expensive and rare, and as a result, there was an intense focus on "developer productivity" and DX in general. You had to squeeze everything you could get out of your investment in hiring devs, and you wanted to keep them happy. You bought them whatever tools they wanted and whatever fancy coffee they asked for. "DX" and "platform/internal tooling" teams were very popular. I still think they're important and very positive ROI, but because they're seen more as cost centers or as longer-term investments, they are getting less love.

Post-ZIRP and in light of industry-wide cost cutting, where does that leave the DX movement?

I don't have the answer but I'll think out loud in this post. I think there will still be buyers for sleek, Linear-style Saas tools for devs, Aeron chairs, and top-of-the-line Macbook Pros. But we'll re-envision DX to make it work with our new post-ZIRP work values.

"Productivity Bundles" in, specialized tools need to prove value

At my job, it turned out we were paying for both Notion and Coda (both are knowledge mgmt tools), and there was a big push towards a single platform. We removed several other tools we weren't using and reduced the number of seats we bought for others. We still use some specialized tools, but the value has to really be demonstrated to convince us to pay for it. In this cost-cutting age, higher-ups want to see real value for a specialized tool, and they definitely prefer a lower-quality one bundled into something they already pay for.

Can Cloud-Based Development Environments be a cost-saving play?

Will super-overpowered Macbook pros go out of fashion? (Anyways Apple silicon is so fast, I could probably code on an M2 air). If so, could this push more development towards Cloud Development Environments (CDEs)?

See this primer: https://dx.tips/the-end-of-localhost

CDEs like Github Codespaces and Gitpod let your infra team manage development resources: no more fiddling with your local environment! I love the idea and have been pushing them at my job.

On top of the time savings, they have security benefits since data never leaves the company's network. For now, these require a lot of work to setup (cf Shopify's journey), are usually implemented by expensive internal platforms teams, and usually result in increasing cloud spend by renting the fastest machines for devs to run tests on.

Maybe for a new company, doing the initial setup for Codespaces could help justify buying cheaper laptops? A fancy Mac is ~$5k and devs are much more expensive thna that, so I doubt we are saving much here. This trend was slowly playing out and has a ton of DX value, but maybe now the cost-cutting angle will become more salient. On the flip side, if it doesn't find a good cost-cutting angle, this trend will see slower adoption because "higher dev productivity/happiness" isn't visible on a spreadsheet and bean counters won't pay for it.

Simpler tech?🙏🏻

What about the frameworks we use? The Web/JS world is currently in a simplifying wave, after a Cambrian explosion of tools over the last decade, the ESM/CJS split, Typescript and SSR React now somewhat stabilizing. Now we're working on ways to bring that into a single tool like Bun or Nextjs + turbo.

During ZIRP there was a mantra of growth at all costs, and extremely complex architectures/frameworks were tolerated if they allowed for more short-term growth (eg Uber has a bajillion microservices, but that allowed them to scale really fast). Now that the money faucet is off and "new features that make money today" are the focus, I could see us returning to simpler, stable frameworks that make it easier to churn out features to get money now (Rails/Django?). In JS-land, less battle-tested frameworks like Remix might get less love as devs stick with the Nextjs devil they know. Within Next, probably app router adoption will be slow (I for sure am not rushing into it at my job).

Idk where that leaves Svelte and Vue. These are loved for their simplicity, and they've reached some critical mass, so they'll probably keep chugging along?

Will monorepos be worth the headache?

The idea of keeping all of your code in one repo instead of many has been gaining popularity (outside of Google) over the last decade. My job has a monorepo and it's great as a dev. It does require more infra work as many tools assume one-project-per-repo, and the build tooling (for us yarn, then npm workspaces, then pnpm iirc) has been an ordeal. In general, monorepos promise long-term "ease of change" and benefits for larger companies, but require a ton of upfront setup and fiddling with tools like Bazel and Nx. If we are doing simpler tech now, I could see slightly fewer of these getting started?

At the same time, maybe the tooling has improved enough to avoid this?

Conclusion: It doesn't look like a big shift

I doubt any of these will change the course of history, they look more like slight pulls on existing industry changes that were already underway. I doubt company cost-cutting practices will be enough to make us switch tech stacks wholesale for now. Devs are still really expensive, and companies still make a lot of money from digital stuff.

Also, it is entirely false to think of "The DX Movement" as zero-interest phenomenon. As devs, building tools for ourselves is still one of our best sources of leverage. DX makes devs cheaper, and if balanced well, does actually result in more "short term features that make money now" delivered than without.