ETHDenver 2026

Missed the previous week due to travelling to Denver and back, so this post covers two weeks, ie, from 2026-02-16 through 2026-02-28. Side note, this is not terrible as I am trying to practice showing myself grace, and letting myself fail without beating the self up, so over all, growth?

That being said, this post largely covers ETHDenver 2026, and catching up post that. See below to read more in detail!

Driving

First off, I drove to Denver all the way from upstate New York, and while Google Maps did say ~24h drive, and I hope to make it in that much time. I had to stop quite a bit and finally made it there in like 40 hours. One stop was because I'm getting fucking old, and I really needed to sleep; but the other stop was because of actually needing to work for a few hours and attend a couple of calls for The khaali Film Festival (more below). Speaking of which, this new workflow that I have with Claude Code and Wispr Flow is really insane! My productivity has gone through the roof, and my core problem of execution is slowly being taken care of! I also had Starlink mounted on the car, so I always had network access regardless of where I was; which meant if I wanted to take a break, or had sudden inspiration, I could easily stop at the next rest stop and act on it! Ofc, I would pick slightly scenic rest stops so I could also take timelapses.

But yes, I enjoy driving very much, and I was glad to take the opportunity to drive to Denver and back for the conference. Fortunately, I also found a rather inexpensive way to take care of the pending issues in my car before I left. Maybe I should talk more about it in a future post. The funny thing was, I was pretty popular at the hacker house I was staying at because I drove all the way from Upstate NY; but little did they know, I live for this shit; the freedom in driving at night, with barely anyone on the road and low visibility blooms my creativity and makes me feel alive! It does take a bunch of time though, and there are still a few things I need as mainstays in my car to actually make it sustainable, especially as the backlog keeps increasing!

Speaking of backlog, claude code is a fucking life saver! For a long time, my problem as been execution; I have a bunch of ideas, some good ones, and a bunch of implementations in mind, but to actually implement these ideas is as painful as dragging my corpse through a bed a nails. Maybe this is ADHD, maybe its just me, but I like mulling over ideas in the idea space or the headspace, and once they have some level of finality, I just wish for it to exist; but of course, one still needs to put in the work and bring their ideas to reality. This is where Claude Code really shines; because now the time and effort required to actually make something real has dramatically dropped. Ofc, there is still some work to be done with the specific tooling and harness, but that's something I am working on and also something that khaaliDimaag can potentially solve. More on this in the coming posts.

But, on the way back, I did end up finishing a bunch of tasks due to speech-to-text and my agentic setup (as minimal it is right now): fixed an issue with my personal portfolio, started a library project using ens subnames, started another project for Story Protocol's PIL terms, and got a little further with tuning baseline for the lighthouse framework. More on these in the Claude Code section below.

The khaali Film Festival

One of the main reasons why ETHDenver was so important was to get a case study out for The khaali Film Festival. This is something I have been meaning to do for the upcoming devcon conference in Mumbai; the core idea is to make this a confluence or a celebration of Bollywood, AI, Blockchains, and also add a hackathon element to this amalgamation! One of the things that this entails is to have a customized platform to manage submissions and also judging and relatedly crowd voting. Ofc, the crowd voting is quadratic and modeled after Gitcoin and Allo Protocol, with other web3 goodies added. So, we ended up partnering with Vibe Check 4 and their film festival to test out the crowd voting feature of our eventual platform; and to great success!

The first version of this involved a simple crowd voting app, but when I was showing the demo to people, somebody suggested that I should make it a live companion app, which truly changed the game. A lot of shows, especially plays, have a booklet that the audience gets and often refers to during the event, to get a sense of what is going on, so at the very least, this companion app lets them do that. More interestingly though, provided submissions also happen on this platform, the organizers of the film festival can run the entire screening through this app, from within the admin section! But even more interestingly, because now that the app is a live companion app, I can also track when the audience votes for a film, thereby providing specific feedback to the filmmaker that "hey, audience really liked the action sequence at 3:23"!! Of course this dynamic has a bunch of other applications (such as use within focus groups), but more on that in the coming posts. In fact, The khaali Film Festival probably needs its own dedicated post, now that I think about it, since there are bunch of other ideas I am exploring within it as well.

Kismet Casa / Based House

Changing tracks slightly, I do want to rave a little about getting access to the Base Hacker House, organized by Luciano. Getting access to this house actually meant that I could actually go to Denver! I was already driving to Denver and there would be expenses there, but having an additional expense of at least 500 USD would have made Denver a no-go. But, more than that, staying at the house helped me make many more friends, many in the Base ecosystem, folks I might have not had the chance to spend time with or would have met for maybe five minutes at a side event and exchanged contacts that went nowhere. Staying at the house, and interacting with web3-aligned people was different. So yeah, while originally I was just grateful to be having a place to crash, by the end, I saw how massively helpful this stay can prove to be. Of course, time will tell if something comes off this, but at least it is a probability now where without this it wasn't even a possibility!

Claude Code

As mentioned, Claude Code was a damn life saver (I should really stop glazing this so much lol), and helped me vibe code engineer a bunch of projects / issues. The first one being the live voting platform mentioned above, along with the quick pivots, and the others being the following:

Social Links on portfolio

A lot of social websites allow you to insert a link in bio so other people can find your portfolio (or anything else) through that channel or medium. I have always thought that's a great opportunity to gauge how a platform is serving me, but I also don't like all those UTM links, all urls with query params look kinda ugly. Besdies, I need to setup special analytics tracking to actually have them be useful, which means costs, even on Vercel. I had also explored tracking referrers and other headers but that gets too complicated quickly and also is not always accurate. So, the solution I came up with was linking to my site as thisispalash.com/[platform], and then eventually having a middleware layer that can route traffic accordingly. This also had the added benefit of just maintaining a json file to keep track of all platforms I am currently on.

So, the issue (may not be visisble as repo may be private) was put up on github way back in September, but I hadn't found time to actually build out the components. Enter Claude Code. Gave it access to the repo and the issue and let it get to work after a few turns of conversation. It even suggested that I should be rewriting the request to / and use an ephemeral cookie so the app knows that a redirect has taken place. For now the build only shows a modal with a message pertaining to the platform (again, a simple mapping in the json), but in the future can also be a dedicated page with content from that platform. Here too, because Claude had access to my codebase, it reused my BaseModal component instead of even suggesting a new package to manage the modals! I know this doesn't seem like a big deal, but a year or even a few months ago, this was not the case..

While a major chunk of the work was done on the drive, I had to still spend about 4 or 5 hours debugging once I got home and fix some stuff so that everything worked flawlessly. For instance, Claude did create a middleware.ts file, but did not place it correctly and as NextJS expects it, so that caused some problems. Another issue was that I wanted custom components to also be shown in the modal and when writing plain JSX in the json, it wouldn't be rendered. So, after updating my json fully, I let Claude convert it to a JSON based DOM structure and everything rendered perfectly after that. It is also quite flexible and scalable! But, this sort of collaborative work made the PR look beautiful (again, might not be visible due to repo's visibility settings). Exciting things are happening in 2026!

PIL Generator

This was another of these long pending issues and projects in my backlog. When participating in Story Protocol's hackathon late last year, I was running around in circles trying make sense of their extensive Programmable IP License terms and wanted an easy way to come up with appropriate terms for a specific project. I then figured this would be a useful project for the community as well and hence opened an issue back in November. Again, no chance to really get to this, and this was also a lot bigger and involved project / issue than the social link one above, although that was complicated for a different reason.

Still, I figured, let's let Claude have a crack at it! and again, a few turns of conversation explaining what I want, a link to the issue, some other links to the specific shadcn/ui template I'd like to use, and it built out the entire thing! Take a look ~ PIL Generator! Of course there's a few things still pending like actually enabling the LLM chat, verification of certain terms (like royalty contracts and the like), building out the more QnA style generator, and maybe one-click deploys. But, it's about 60% of the way there! And in ~5-6 turns and ~3 commits. I also did not use any skills for this yet since I am yet to set those up at a global level. The most remarkable thing about this is that it managed the codegen portion of the project so anyone using this tool can quickly get started with Story Protocol. Also, the entire app is client side with a bring-your-own-keys design, which ofc in the future can be enhanced by actually having a backend and also storing user preferences and such. But, I am not bearing that cost, so it's all in client now with a BYOK design.

This project likely needs two more sessions to get to a ship ready state, and I haven't found time yet for it since there are more pressing things, but again, 2026 is for sure a make-or-break year!

khaaliNames

Now this project is fairly recent and was born out my CS instincts; namely, I had build khaaliSplit a few weeks back where I needed subnames and then needed subnames again during the voting platform. As I thought about it, all my projects were going to need subnames since for a while now I have been looking at how a traditional server architecture can be replaced by blockchains; and ENS is perfect for user management. So, why not create a reusable library onchain that all my projects can reference / call, and that effort is now reduced for every subsequent project? (PS. don't you love when software design starts doing this?!) This also has other benefits such as milestone tracking (ie, how many users have signed up), being permissionless, and also ensuring acceptable usernames. I put down my thoughts on this issue if you'd like to see more.

But, I figured, hey lets let Claude build out a minimal version of this to get over my "cold start problem" and I can take it from there, and that's what I did! Claude successfully pointed out nuances and built out the contract architecture for the library itself, and now all I need to do is update the actual dictionaries, maybe refine some of the functions in the library contract, and provide an example implementation for using this library. And if I am honest, once I think through how control and data should flow, I'm likely going to have Claude write it out for me. It aslo writes out the tests, so that's a huge plus as well. I should really look into the Test Driven Development tbh, something I have historically not been too good at while I'm at it. But, right now the priority is to ship faster since this is upstream of my other projects, so I will have to put a pin in TDD workflows.

Baseline System

Finally, gotta mention 0xTrxnt as well since he has told me that he intends to be featured on every blogpost! But jokes aside, this was the final task accomplished on the way back with the help of Claude Code. As a refresher, the Baseline System is roughly a knowledge management system for a project / org, and its real innovative aspect is a way to route through the context files and skills so that the context window does not get filled immediately (as is the case with a bloated system prompt, AGENTS.md / CLAUDE.md, or MCP servers). I have been using his system right now to first document my projects (and also think through them since there are skills available for brainstorming), especially the lighthouse-framework which is designed to make dApp creators focus only on core business logic, and is the underpinning of all interoperability and composability within khaaliDimaag (more on this in the coming weeks and months!).

Previously I had had Claude document the core system, but there were still some peripherals that needed documenting. I had also not used the system cleanly from the get-go so there was some cleaning up to do. So, I had a multi-turn conversation with Claude to clean up the directories and also include the peripherals and it's almost at a stage of deliberate manual review; there are probably one or two conecpts missing (such as the color palette or theme generation for one).

The goal here is to first document all of this context, and then also structure it to also include engineering specific context (such as architecture or tech stack or conventions, etc.), so that in the future, for a new project or feature, I can hand of the entire thing to Claude (or any agent really), and only have to review a PR (where also effort could be minimized by following a more TDD approach I think). This lets me operate solo, but with the appearance of a cracked team, while also allows me to create those magical moments with users or customers where I can hear out their problems at hour one, and have a fully deployed solution by hour three. Of course, the lighthouse-framework and the architecture of khaaliDimaag is doing a lot of the heavy lifting as well, but having an agentic workflow lets me build the thing (or more accurately, have the thing be built) from anywhere, in any setting. Another point to the magical moments I wish to create!

The Second Week

This section is added due to missing the weekly cadence. Provides a minimal overview of the subsequent week.

I reached back home in the middle of the following week and after reaching it has been playing catch up mostly for all the stuff that was pending, while also trying to juggle new opportunities. One, the part time role I am at was gracious enough to let me take it a little slow during the week of ETHDenver, so most of this week went towards catching up there and finishing my refactor. Turns out in my absence, there was another client demo and there are more features to be added! So, really need to finish up my refactor haha. Also, I keep saying refactor but the tool currently is already more than what was in v1, although we have not changed branches yet, and the tool has only just started working end-to-end, with some new feature additions, on both, the user side and the infra side.

Two, I met my friends from Cracked Labs after a long time, and they have an exciting new project going on around financialized games and a platform + events to go with it. I have been roped into this as well, so been onboarding myself here as well. Three, I am currently participating in Rootstock's builder camp and have been massively lagging behind. I did join late (as a result of getting off a waitlist), but I have also not really been keeping up. So, getting on track with the camp is also a priority. As for why Rootstock? Two reasons mainly I think, (i) I had won a bounty last year at a hackathon from them, and the project was more a proof of concept than fully functioning (~36 hours after all!); and (ii) I have been meaning to get more involved in the bitcoin ecosystem, and Rootstock is perfectly situated with its EVM runtime and BTC merged-mining. I have also been curious as to what a L2 on bitcoin even means, and again, Rootstock seems to be the answer.

Finally, four, the crowd voting / live companion app from The khaali Film Festival above has taken a life of its own to some extent, and I need to polish a few aspects of it; (i) make it a self-serve festival management platform that can take in submissions and also run the screening, along with live voting and maybe also async voting; (ii) in the interest of time, the current iteration of the app is very web2, I need to make it a mroe web3-forward app and fix my contracts to actually handle the different scenarios it can be used under; (iii) fix the bugs and polish the ui as a result of the live testing that happened during ETHDenver; and (iv) translate the current app and layout to the lighthouse-framework (more soon!) and fix the framework as needed.

Actually that reminds me of a five, The khaali Film Festival itself! Apart from all the sponsorship stuff and general business admin, there are tech things also that need to be taken care of like the copywriting on the page (still ai written!), connection to a proper database (currently connected to google-sheets!), more information about the various platforms, faq, and longterm vision, and also finishing up the ai film that's supposed to autoplay in the hero section! God damn I'm going to drown lol.