sagunbuilds

Joined 3 months ago

Karma
2
Submissions
11
Member since
Jul 28
Mechanical engineer, building Clicked, an AI text explainer extension: https://tryclicked.app
View on Hacker News

Recent submissions

Comment · 3 months ago
That finishing a website and getting Google to index it are two completely different projects. I shipped a few hundred pages, all technically clean (now), and most of them have been sitting in "Discovered, currently not indexed" status MY AI coding assistant tells me a new domain with no external links just doesn't get crawl priority, and no amount of extra on-site work will change that.

Two specific mistakes I did fix along the way: my landing page force redirected visitors to their browser language, which can hide the other language versions from crawlers. Google's own docs say offer, don't force. And my internal links pointed at page.html while my sitemap said /page, so Google was discovering every page at one address and being told the real one was another. This could have also contributed to the limbo my website is in (although Gemini says that that doesn't happen). Fingers crossed

Comment · 3 months ago
I learned by building instead of studying. I'm a mechanical engineer working in finance, zero software background, and over the last couple of months I shipped a browser extension with an AI assistant writing the code while I made the product decisions.

The surprise was a) it has been much harder than I expected b) I ended up learning a lot. I thought it would may be pick up a bit of programming. But it was mostly everything around programming: how OAuth works, why my sign in kept expiring after an hour (a code deletion took out the token refresh route with it), how Stripe webhooks fail silently, why emails land in the Promotions tab, how Google decides what to crawl. Each broken thing forced me to understand a system I didn't know existed the week before. Learning to work with the AI model has also been a huge lesson on its own:)

Comment · 3 months ago
Sure. For consumers, we will get there sooner. For enterprises, it is tough to see that happening any time soon. Even enterprises with fully staffed development teams will struggle to have the operational expertise required to replace legacy systems / providers. But you never really know.
Comment · 3 months ago
Tough to go through what you are describing. The only positive thing I can add is all this experience is making you better skilled / positioned for your next project. When the cycle repeats enough times, eventually you get success. Best wishes.
Comment · 3 months ago
At the highest level, the only fundamentals are supply and demand. Price goes up because people believe it will go up and vice -versa. One (a bit generalized) way of looking at investing is - Indexing helps preserve wealth, investing in individual stocks can help increase wealth. Risk and reward go together. When you invest in an individual stock, the expected return from that individual stock could be higher than the expected return from an index (let's say one tracking S&P500), but the risk is also higher. Remember that there is a lot of money (pension funds, index funds, ETFs, 401(k)s) who continuously invest in all types of index funds. This drives continuous demand and offers some support to the price for the index (and the stocks within). As with your point with volatility in the stock markets, derivative trading around key events drives a lot of that. There are specific shops set up just to do that. The derivative volume can also influence the underlying stock prices (think call options volume exploding forcing option sellers to buy stock to cover their positions, which in turn pushes the price up)
Comment · 3 months ago
Niche domain expertise still matters. A person / team with little experience of banking workflows and requirements can't produce software to replace existing vendors. That offers a moat. Obviously, not true for every case.
Comment · 3 months ago
As I read through the list, I found so many things that became part of my work flow over a period of two months or so. I started by just telling the agent to build what I wanted. Once, I tried to use the product and had to add more features , workflows into it, I quickly realized the limitation of context within a chat window. Also, realized that I could not rely on the model to tell me when something was done or if it was even done the right way. My most important take-away was to document everything important, provide that document at the begining of the chat and have it confirm that the rules were followed / specs were aligned after a task was done

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton