All you need to know about GPT-3 & Open AI

My twitter feed is on fire with all the cool stuff people are building with the new GPT -3 beta API released by Open AI in June 2020. So I spent my weekend learning about all the use cases people are exploring and their experiences. I knew about Open AI from 2015 when it was launched by Elon Musk, Sam Altman & few other investors (which included Infosys from India) who collectively pledged US$1 billion to start work on AGI (Artificial General Intelligence)

Open AI Stated goal - promoting and developing friendly AI in a way that benefits humanity as a whole.

Open AI turned into for-profit after Microsoft came in with an additional USD$1 billion investment in 2019. OpenAI then announced its intention to commercially license its technologies, with Microsoft as its preferred partner.

Now, this might look like a lot of money but when you are aiming at training a machine with the entire internet of data, the infrastructure cost will make $2B look insufficient.

Of course companies like SpaceX, Tesla, Google, Facebook & Amazon maybe much ahead in using AI but we don't have much visibility into what they have and how they are using it. The precision landing of rockets, self-driving cars, Alexa, Siri & Google Home is all great examples of AI in action.

In the 2017 tax year, OpenAI spent US$7.9 million, or a quarter of its functional expenses, on cloud computing alone. In comparison, Google's DeepMind total expenses in 2017 were much larger, measuring US$442 million [Wiki]

OpenAI brings the powers of AI to startups and small businesses by giving API access to GPT-3, a state-of-the-art language model made up of 175 billion parameters. â€œThe supercomputer developed for OpenAI is a single system with more than 285,000 CPU cores, 10,000 GPUs and 400 gigabits per second of network connectivity for each GPU server.,” the companies stated in a blog.  

AI is becoming a platform

Training massive AI models require advanced supercomputing infrastructure or clusters of state-of-the-art hardware connected by high-bandwidth networks. It also needs tools to train the models across these interconnected computers.

GPT-3 API makes the model available for millions of developers to go figure out how to do interesting and creative things with it.

Microsoft claims that it ranks in the top five machines compared to machines listed on the TOP500 supercomputers in the world. China currently dominates the list with 229 supercomputers, leading the second place (United States) by a record margin of 121. Since June 2020, the Japanese Fugaku is the world's most powerful supercomputer, reaching 415.53 petaFLOPS on the LINPACK benchmarks.

So what happens when the world's top supercomputer which can 'learn new skills & solve tasks as humans do' is given to common people to explore its possibilities, it can be endless and amazing, get a glimpse of what people are building below-

1. Use Case for Design -

NLP + Figma Design to create a design by just typing what you want to create. https://twitter.com/jsngr/status/1284511082019729409

2. Use case for Development -

This Demos's a plain-language input converted to a working React code and an app. https://twitter.com/sharifshameem/status/1284421499915403264

3. Usecase for Strategic writing -

It's not just autocompleting of a sentence but can complete an article, book, etc. Imaging article, legal documents, Ph.D. thesis was written by AI - https://twitter.com/zebulgar/status/1283921361627344896

4. Usecase for an AI Therapist -

Intense Therapist with a human-like conversation - https://twitter.com/nicklovescode/status/1283300424418619393

5. Auto completing the images -

https://twitter.com/xuenay/status/1273898995706998790

Many more here - https://twitter.com/xuenay/status/1283312640199196673

Writing this article was my attempt to have a long conversation with myself on understanding AI. There is still a lot more to explore in the implementation and implication of this powerful technology. If you come across any interesting use-case please share the link in the comments below. Would love to hear your thoughts on how this will impact India and Indian startup space.

Disclaimer - This article was written by a human ;)