Interview with Priyanshu, Founder of Hyperspawn Robotics

Written by feverishhh | Published 2022/01/06
Tech Story Tags: startups-of-the-year | robotics | ai | human-psychology | telepresence-robots | artificial-general-intelligence | open-source-technology | startup

TLDRPriyanshu is a 22-year-old CS student based in India. He's working on understanding and reducing the human brain into a humanoid robot by using the information we currently know about our brain. He also has realized that the humanoid’s body movements can be controlled remotely which can be beneficial in many different areas and hence, it as a starting point of the company. He says he's excited about the amount of data we can process and training models like GPT-3 is just the beginning.via the TL;DR App

HackerNoon Reporter: Please tell us briefly about your background.

I’m Priyanshu, a 22-year-old CS student, based in India.

I’m interested in robotics, AI, and human psychology. I’m working on understanding and reduplicating the human brain into a humanoid robot by using the information we currently know about our brain.

I also have realized that the humanoid’s body movements can be controlled remotely which can be beneficial in many different areas and hence, it as a starting point of the company.

What's your startup called? And in a sentence or two, what does it do?

It’s called Hyperspawn. Initially, we’re developing humanoid robots that precisely mimic human movements and can be controlled remotely using VR gear from anywhere in the world. We will also be working parallelly on the humanoid’s brain to be capable of understanding humans’ psychology.

What is the origin story?

Humans are interesting. If we scope out from our daily lives and notice, our brains have evolved to be self-conscious and super-intelligent.

Our brain does a lot of tasks subconsciously. It's really hungry and always searching for patterns. A basic eg.: Seeing a stranger's face and automatically calculating all sorts of parameters like estimated age, background, ethnicity, etc., and compartmentalizing this data efficiently to make decisions off of them.

Coincidently, (or not), neural networks work similarly with different complexities of layers and stacks used to build them. Especially, algorithms already into existence using transformers for contextual training of data, like GPT-J, Codex, etc. that can mimic the working of our brains but efficiently. Making the humanoid super-intelligent is the end goal.

What does your team look like, and how do you plan to solve this problem?

We’re just 4 college graduates building the hardware (humanoid) as of now.

But as far as the software is concerned, anyone interested can be a part of the team. It is an open-source, completely transparent project and we have chosen to build it in public. Our board room is a subreddit, and the meetings are streamed on YouTube.

This is to facilitate data contributions from all human brains about the human brains.

If you weren’t building your startup, what would you be doing?

I don’t know!

At the moment, how do you measure success? What are your core metrics?

We are still in the prototype phase. Our key metrics are very technical (in terms of building the hardware.) Some of them are:

  • No. of DOF mapping of each humanoid body part.
  • Communication Latency from tracked human movement to VR controller to sensors and actuators.
  • % efficiency of each moving part that uses actuators.

What’s most exciting about your traction to date?

We don’t have traction yet. However, this idea was validated by bosch over an email with a letter of interest, which is very exciting.

What technologies are you currently most excited about? And why?

I’m excited about the amount of data we can process. Data indexed to train OpenAI’s GPT-3 is huge. With the technology we have and will have, training models like these is just the beginning. Using more relevant data from a sophisticated system and training contextual models will be the future, I would say.

It is also exciting to combine VR with hardware and robotics for a complete shift of localization.

What drew you to get published on HackerNoon? What do you like most about our platform?

We need as many people to know about what we are building for their feedback and support in the project.

HackerNoon has a good, “hacker-y” user base that could understand this idea fairly easily.

I like how developer-friendly and geeky the whole platform is. Makes a good comfort zone for the geeks and nerds.

What advice would you give to the 21-year-old version of yourself?

Those $800 bucks in your bank account are enough.

What is something surprising you've learned this year that your contemporaries would benefit from knowing?

Relax. We’re all the same. Try not to get caught up and stressed out so much, ultimately just for money.

And also, age does not matter.

If you find our startup interesting, vote for Hyperspawn here.


Published by HackerNoon on 2022/01/06