DEV Community: Siddharth Mittal

So Astra the latest frontier model by OpenAI was able to drive a car.

Siddharth Mittal — Wed, 23 Sep 2026 22:12:11 +0000

So Astra the latest frontier model by OpenAI was able to drive a car.

A very short distance at a very slow speed sure, but the interesting part as mentioned in the report, Astra retained context from it’s first failed attempt and “learnt” to avoid the same mistake in the second one.

Originally published on https://webxsid.com/.

Why I Built Crona

Siddharth Mittal — Sat, 19 Sep 2026 00:00:00 +0000

It started about 10 months ago, end of 2025, however it wasn’t a terminal app, or even a programming project.

Back then as I started taking on more responsibilities at my day job, it was getting increasingly difficult to keep track of everything I was working on. Especially being in a remote role, my day was starting to blend together—the day job, personal projects, daily chores and other small things that make up life.

I did keep a To-Do list fairly regularly but it wasn’t enough, I wanted a better picture of my time, what was getting done & what I was neglecting.

At this point you’re probably thinking, “but there are already a million tools out there to solve this problem” & you’re right, there are, and I did try a few of them. Some were too simple, others wanted me to adapt the way I worked around the way they thought I should work.

But there was one thing I kept coming back to: I wasn’t particularly comfortable handing over such a detailed record of my life to yet another service.

Tasks themselves might not seem particularly sensitive but over time the data logged could paint a surprisingly detailed picture. What I’m working on, when I’m working my habits, my projects. This might sound paranoid but the more I thought about it the less convinced I got that this data should ever leave my system.

So I did what any reasonable developer would do and started abusing the one app I knew keeps everything offline—Obsidian.

It was the perfect choice at the time, completely local and extensible enough to conjure up macros that suited me. So tasks became md files, I spent days writing macros with JS and building “dashboards” with Dataview & DataviewJS, and it was insanely powerful.

I had heat maps, trend lines, tables and graphs—powered by a growing pile of Markdown files and a bit of JavaScript. Looking back this was the first primitive version of Crona, it just didn’t have a name yet.

It wasn’t the most efficient system, of course, but I knew that going in. After about a month of using it though, the inefficiencies became impossible to ignore. I was constantly switching between my main work setup and Obsidian just to track things, and as the number of .md files grew, the macros started slowing down too.

I could’ve spent some time optimizing this further to tackle the performance issues, but Obsidian was still another app I had to constantly switch to. Most of my work already happened in the terminal, and I realized that’s where the tool needed to live too.

So I went back to the drawing board and started figuring out what a purpose-built version of this system would look like. After some thoughts & discarded ramblings I ended up with three core parts that would become Crona:

The Daemon — the always-running background layer that owns the data and does the actual work.
The TUI — the richer, interactive interface for when I actually want to sit down and use Crona.
The CLI — the quick interaction layer for doing things without leaving whatever I’m already working on.

The idea was simple: Crona should always be there, but I shouldn’t always have to have Crona open.

Daemon was probably the most important piece of this. Timers had to keep ticking, notifications had to fire & data had to be processed regardless of whether I had the interface open. It owned the state and gave everything else a single source of truth.

The TUI and CLI were simply two ways into that same system. The TUI could give me the dashboards, history and richer interactions I had grown used to in Obsidian, while the CLI handled the quick stuff—start something, mark it done, check what’s running and move on. Neither needed to stay open for Crona to keep working.

From v0.0.0 through v1.9.0 (the latest stable at the time of writing), the implementation underneath all of this has changed quite a bit, but that basic split hasn’t.

And that’s how, a slightly over-engineered collection of Markdown files eventually became the tool I needed: something that stays out of your way, is easily accessible without demanding attention & still tracks virtually everything—without sending that information off to someone else’s servers.

Crona has obviously grown much beyond the small tracker I sketched out at the end of 2025. There are more features, a macOS companion and plenty of things I still want to build. But I don’t want Crona to become another productivity system that tells you how you’re supposed to work, so I’m being very careful about what gets added—and whether something should be removed.

I built this because I wanted a record of what I was doing without changing how I was doing it.

Ten months later, that’s still the idea

Website · GitHub

Originally published on https://webxsid.com/.

The Dev Withdrawal

Siddharth Mittal — Sat, 18 Jul 2026 00:00:00 +0000

The CI pipelines were green, release workflow finished with the same few warnings I’d been ignoring for months. Crona was finally stable. I’d done it, built the app. I lean back in my chair, equal parts exhausted and ecstatic, tell myself “good job,” and call it a day.

Next day, I sat down at my desk, slipped into the same routine I’d followed for months.

Coffee. Editor. Terminal.

I opened the repository and just stared at it. There was nothing left to build.

Sure there are bugs, A feature request here and there. A bit of tech debt I’d been ignoring, but majority of the work was done.

For months, every spare minute was dedicated to this. Every commute, every walk somehow turned into an architecture discussion with myself. Suddenly all of that fell silent, I’d gotten so used to the daily chaos that the stability felt wrong.

I’d started looking for problems to fix.

Maybe the warnings are worth solving now, maybe I can refactor the codebase to be “cleaner”, maybe SQLite wasn’t the right choice, maybe, maybe, maybe.

It took me a while to realise, I wasn’t looking for work, I was looking for the feeling of having work. I don’t know if withdrawal is the right word for it, but it wasn’t boredom. It felt like finishing a really good book, instinctively reaching for the next chapter only to find About the Author.

I’d gotten attached to the routine more than the codebase. At some point, building Crona became more than a chore, it became a part of my daily routine and to some extent me. So when the codebase stopped demanding that of me, I suddenly had to answer a question I hadn’t asked myself in a long time.

What do i do now?

To be honest, I don’t have an answer, over the last few days my mind has been working overtime to find the next thing the becomes my routine, new tools to build, weird experiments I can do, rebuilding my portfolio, even writing this blog.

Part of me realise it’s just another way to avoid the silence. To replace one obsession with another. Perhaps I’ll find another project to occupy my time, perhaps I won’t.

For now I’m trying to learn something I haven’t considered before:Sitting in the quiet.

Originally published on https://webxsid.com/.

Owning your data again

Siddharth Mittal — Sun, 05 Jul 2026 00:00:00 +0000

Not too long ago, buying something meant owning it perpetually. Your files lived on your machine, your photos on your hard drive, your music collection wasn’t tied to a subscription. If the company disappeared tomorrow you still had the data. While today almost everything sits behind an account, a subscription, your data is stored on the “cloud”.

Recent events have made that increasingly difficult to ignore. Sony recently announced that all new PlayStation games will become digital-only from 2028, marking another step away from physical ownership. Around the same time, PlayStation reminded us how fragile digital purchases can be when hundreds of previously purchased movies disappearedfrom their libraries after licensing agreements expired.

Neither of these decisions is particularly surprising in isolation. Licences expire. Businesses change direction. Digital distribution is undeniably convenient. But together they highlight an uncomfortable reality: in many cases, we haven’t been buying ownership—we’ve been buying permission.

That distinction seems small until the day that permission changes.

It made me ask myself: how much of my own digital life do I actually own? If I stopped paying tomorrow, lost access to an account, or a company simply decided not to renew a licensing agreement, how much of what I consider mine would still be mine?

The answer, unsurprisingly, was very little.

I could wake up tomorrow and find that my favourite songs have disappeared from the playlist I’ve curated over the years. The games I paid for could vanish from my library with no way to recover them. The applications woven into my daily workflow could shut down, taking years of data, habits, and accumulated context with them.

Some of this might sound far-fetched. After all, the books you’ve purchased digitally aren’t going anywhere… right?

Now, to be fair, I don’t think the companies making these decisions are necessarily wrong or even acting maliciously. More often than not, they’re operating within the law, responding to changing markets, and building products that millions of people genuinely prefer.

And it’s hard to argue with the convenience. I stream almost all of my music. I can’t remember the last time I bought a movie instead of watching it through a streaming service. My photos are backed up to the cloud, I collaborate through shared documents every day, and I happily pay for software that genuinely saves me time.

Cloud computing solved real problems. It made collaboration effortless, backups automatic, and software easier to access than ever before. I wouldn’t want to go back to manually syncing files between computers or carrying around external hard drives just to move a document.

The issue isn’t the cloud. The issue is when cloud becomes the only medium.

I don’t believe every application should be local-first, nor do I think every service should be self-hosted. Collaborative tools, communication platforms, and anything fundamentally multiplayer benefit enormously from centralized infrastructure. But I do think every piece of software should begin by asking a simple question:

Who should own the data this application creates?

If the answer is the user, then ownership shouldn’t be an afterthought. It should be the foundation the rest of the system is built on. Cloud sync, collaboration, backups, and AI can all be layered on top, but they should enhance ownership rather than replace it.

That one question has shaped how I think about software over the last few years. It’s why projects like Crona are local-first by default. It’s why I’m increasingly drawn to software that works without an internet connection, degrades gracefully when services disappear, and treats exporting your data as a basic expectation rather than a premium feature.

See I’ll be honest, I don’t think we’re going back to the days of boxed software and physical media (at least on a global scale). We’ve gained far too much convenience to abandon it now. But I do hope we can move toward software that gives people both: the convenience of the cloud, and the confidence that the things they buy or create ultimately belong to them.

Because if software is meant to help us build our lives, it shouldn’t require us to hand ownership of those lives over in return.

P.S. There’s a certain irony in relying on external links to make an argument about digital ownership. If some of them stop working in a few years, consider that an unintended—but rather fitting—demonstration.

Originally published on https://webxsid.com/.