GPUMonk

Don't take my tweets seriously!!! I am here just for fun

[GPUMonk @GPUmonk]

https://t.co/lb89edpbId

[GPUMonk @GPUmonk]

AI can make you code faster. AI doesn’t make you a genius software engineer

[GPUMonk @GPUmonk]

I spent approx 1400 USD in cursor in last 7 days. I think I should stop using Cursor 🥲

[GPUMonk @GPUmonk]

@kathanvakharia5 @anjalinirwal02 People don’t understand how valuable these things are in adulthood

[GPUMonk @GPUmonk]

This thing actually works. One of our team shipped the crazy product in just 3 months and we already have thousands of customer actively using it and haven’t broken once

[GPUMonk @GPUmonk]

No tech lead, no solutions architect, no product lead, no tester, no QA engineer.

Just AI engineers and vibe coding on steroids is the new norm

[GPUMonk @GPUmonk]

@GetTestably Users becomes the test suite

[GPUMonk @GPUmonk]

@kathanvakharia5 It just doesn’t make any sense

[GPUMonk @GPUmonk]

One of the useless thing in corporate is mandatory training which is BS

[GPUMonk @GPUmonk]

@Yuchenj_UW That’s what make the life interesting. What’s the point of life if you don’t get bored or unhappy with something

[GPUMonk @GPUmonk]

@NandanNilekani @rvenk @EconomicTimes These AI pundits who claim that India can beat US and China in AI, just tell me how.

[GPUMonk @GPUmonk]

@Kalshi Including bitcoin

[GPUMonk @GPUmonk]

@Dr_Singularity It's funny how you can play with words!!!

[GPUMonk @GPUmonk]

@dom_lucre Bro does not need to work his full life!!

[GPUMonk @GPUmonk]

This is far too optimistic and frankly naïve. China is not just ahead because of “talent”; it has spent decades building the full stack: fabs, electronics manufacturing, batteries, robotics, cloud, telecom, supply chains, state funding, and now domestic AI chips.

Chinese AI companies are already training and deploying on Huawei chips, while India is still largely dependent on imported GPUs, foreign cloud providers, and external semiconductor supply chains. That is not a small gap. That is a structural gap.

India has strong engineers, but talent alone does not create strategic technological power. Without domestic compute, hardware depth, manufacturing scale, research commercialization, and execution capacity, comparing India’s next decade to China’s last decade is more wishful thinking than serious strategy.

Trust me, India is nowhere near China in terms of domestic research depth.

India has talented individuals, but much of the highest-impact Indian tech talent succeeds inside US or global institutions, not through a deep domestic research ecosystem. That distinction matters.

Until India can produce comparable domestic R&D output, patents, frontier labs, compute infrastructure, hardware supply chains, and research-to-industry translation, it is not serious to claim India will disrupt the world the way China did. It is ambition without the machinery behind it.

Trust me, India is nowhere near China in terms of domestic research depth. India’s patent activity is improving quickly, but China is operating at an entirely different scale. In 2024, China-based applicants filed around 1.8 million patent applications worldwide, while India-based applicants filed around 76,000. That is not a small gap; it is a structural difference in domestic R&D and industrial depth.

China has about 5.7 million active patents in force, while India’s active patent base is far smaller and not remotely in the same tier. India is improving fast in patent filings, but China’s accumulated stock of enforceable patents reflects decades of deeper R&D, industrial policy, and commercialization.

[GPUMonk @GPUmonk]

@swyx They are really doing research to innovate at architectural level unlike others

[GPUMonk @GPUmonk]

We need cheaper chips!!! We need cheaper chips!!! We need cheap chips!!

[GPUMonk @GPUmonk]

Too many models are being open sourced today!!! I can’t read these many technical reports in a day

[GPUMonk @GPUmonk]

@TheAhmadOsman Open-source AI will win only if we somehow come up with a model which is under 32B, yet performs good enough for daily coding tasks

[GPUMonk @GPUmonk]

@Zai_org Nothing is better than twitter for learning/sharing cutting edge technologies

[GPUMonk @GPUmonk]

@SelflessCricket Look at Pat Cummins