let's play factorio and build a manufacturing system for a factory. for free.

i don’t want to get into blackpills, but yes - while india designs and builds its own car engines, it still lacks a fully indigenous automotive supply chain, even after several decades. it still relies on global suppliers like bosch, avl, honeywell etc for technology and components. when you don’t own the underlying tech, you can’t control costs or independently drive any r&d. you’re basically hostage being a low value integrator meanwhile, china has developed domestic alternatives across every price point and quality tier. blackpill ? yes. whitepill - also yes. there’s nothing you can’t do except if you weren’t retarded.

founder exits

My slightly contrarian, slightly in-agreement view is that India doesn’t have a Shenzhen. And without a dense manufacturing ecosystem, you won’t have the seeds of hard tech. So debating vc investment is largely moot. Indian hard tech startups are chasing high iq problems and that’s fine…but there’s still no Indian Anker. Even simple consumer electronics like battery chargers involve incredibly complex supply chains at cost. China has hundreds of Anker copycats operating at unit economics India simply cannot match. And that gap keeps widening as China pumps out more volume. So founders naturally gravitate toward niche areas like rockets where the problem is more about engineering brilliance than supply chain economics. I understand the ambition, but very few startups can play that game. Indian founders could build a saas ecosystem because it didn’t require govt support or a physical industrial base. Engineers, laptops, cloud infrastructure, and global distribution were enough. Startups could emerge first and the ecosystem followed. Minimal govt interference - minimal infra dependency. Hard tech works in reverse. You need suppliers, tooling, factories, skilled labor and volumes before startups can compete. These things take decades to compound, and VC money cannot shortcut that process. In saas startups created the ecosystem. In hard tech, the ecosystem has to create the startups. India needs a Shenzhen circa 2012 before it can have its Anker circa 2016.

Openrouter solves a discovery problem. There are 100s of inference providers. 100s of endpoints...all serving the same model with varying latency, reliability, throughput (and pricing) How do you know which endpoint is actually good ? Openrouter solves that.

these are all the models I tested this month on Proto business agent deepseek v4 flash (1 tn tokens overall) continues to be our daily driver.

this model looks really good. we need to put it to the test live. see where it cooks and where it falls apart.

Codex usage on Openrouter makes more sense. Top model is Luna, followed by Deepseek v4.1 flash, followed by Astra...so top 7 are all GPT except Deepseek

saas ceo ❌ x poaster 😢 youtuber ✅ pls like and subscribe my channel: youtube.com/@shashark