Nikhil Shahane

I no longer use AI to write for me. Summaries for research with citations that I can go after and code - where deterministic tests work well. I find most writing output is very predictable and without soul. It's great for stuff I don't care about, but anything that I'd want to put out (like this reply here) has no input from AI at all. To be fair, I've also been writing much less than I did...

The money in oss has not only been in creating tools and libs that anyone can use for free - but providing a competitive service that can be monetised. The way we do oss will evolve to account for the use of LLMs. Eg: I use Codex a lot to build stuff and a lot of my repo is os - I've built a video editing tool that anyone can use / copy. Of course it already uses the ffmpeg and yt-dlp libs -...

This is super impressive. Not just because of the visual design but the new site loads faster and performs better. I ran a quick SEO checker (using a free tool) and the new site scored a 86/100 while the old one scored 54/100 - I didn't bother looking at the details but it's still a massive jump. Clearly @Astro did a fab job. Definitely interesting.

We're in the process of reinventing our agency to be AI Native. It's the future of Agencies of all kinds and it's the perfect combination of being 'Techno-Human'. The humans will bring the 'taste' and the AI will do the heavy lifting. As I've mentioned, the ultimate barrier to any progress/evolution is the rate at which people adopt anything. I work primarily in advertising and marketing - and...

This is just lame. This is done to trick algos, but the trust it erodes is not worth the temporary spike. I like how a16z think about it - they're in the business of building a reputation. It's done over time and is incredibly hard to build and maintain. One misstep can send you back to the ground floor. I typically ignore such content and my brain flags them as untrustworthy.

Marketing is very important for sure. Google is a great example of great products (generally speaking), but terrible Marketing. Most brands that sell fall in the other bucket - where they have mediocre products, but the marketing is what gets people to try and sample them. But good marketing (and it's downstream cousin, advertising) can also make a terrible product look great. Though, once the...

I thought problem 2 has been solved with many note taking tools, right? Fireflies and Otter do a good job from what I've experienced. Problem 5 is a tough one. There are many influencer management tools in India - but as the problem author pointed out requires a service + product focused on that end of the market. Wonder how monetizable it is.

This is an interesting problem. I haven't worked with stripe, but I'm assuming that if there's an API it'll be relatively straightforward to integrate an agent that can do the stripe onboarding, right?

The proposal one is interesting. There's different types of proposals. Financial, creative, etc... Both equally important. But the creative one is harder to get right. We're trying to solve for that - but not so much from the proposal pov - but from an ideation pov.

I am absolutely paranoid about stuff that is truly personal or with high fall-out potentially. So as far as possible, I don't give it access. I'm actually contemplating running stuff on a local RaspberryPi where I can have much more control. I'm experimenting with building a few skills that make interactions more deterministic so I can give it gated / limited access to personal finances or...

Super late to this party, but here goes: Work Goals: >> Ship our company's AI product this month - get real world feedback and scale that product >> Ship a 1 solopreuner product and see if I can get anyone to pay me something for providing them value - any money, is my first step. >> Find and befriend other solo/small team founders builders who will keep me on the good and straight path (see...