Cheers to another AI breakthrough

Let me know if you think GPT-3 wrote today's newsletter

Hey everyone! Good morning and happy Tuesday! 

This week’s newsletter is dedicated entirely to OpenAI’s GPT-3 AI API. I’ll cover what OpenAI and GPT-3 are and my thoughts on its impact on fintech and society.


What is OpenAI?

OpenAI is a non-profit, AI research and development organization based in San Francisco, California, and its financial backers include prominent techies like Elon Musk, Reid Hoffman, and Peter Thiel.

OpenAI’s mission is to ensure that artificial general intelligence (AGI)—by which we mean highly autonomous systems that outperform humans at most economically valuable work—benefits all of humanity. We will attempt to directly build safe and beneficial AGI, but will also consider our mission fulfilled if our work aids others to achieve this outcome.

What is GPT-3?

GPT-3 is a pre-trained natural language model that’s been built using 175B parameters. Comparing GPT-3 to other pre-trained models shows just how robust it is.

Image for post

GPT-3 was created to be a task-agnostic solution, which mimics the way humans learn. This means that the model needs zero to few examples to learn what the input is asking for and to predict the most relevant output.

Image for post

So what’s all the hype?

Here are some early experiments run by folks in the tech community:

As you can see, GPT-3 is able to generate intelligent responses and useable code with simple text input. In the above examples, GPT-3 can auto-design layouts and auto-write a blog post with little to no training examples. That’s crazy!

While I don’t envision this model replacing humans overnight, it will at least speed up the deployment of new technologies and product iterations. For example, a designer can spin up a new mobile dashboard in seconds just by typing in one or two sentences! To thrive in the future economy, we must all adapt to the new tools that we’re blessed to play with. This means thinking of creative ways to use a fancy tool like GPT-3 to our benefit to unlock monumental breakthroughs instead of incremental improvement.

Here are the experiments that I would like to see run with regards to fintech:

  • User types in the career, living location, desired lifestyle, and current bank and brokerage accounts. GPT-3 outputs a life path with actionable guidance on job roles to take, cities to live in, and budgets to maintain.

  • User types in a job description, benefits, and location. GPT-3 outputs a unique contract with state-specific regulations and guidelines.

  • User types in current economic conditions including unemployment rate, inflation rate, consumer debt to GDP ratio, etc. GPT-3 outputs recommended benchmark central bank interest rates.

What are your creative ideas for fintech? DM me on Twitter if you want to bounce ideas!

If you want to work on your own experiments, sign up for the beta access here.

Sources and Additional Reading:

  • https://openai.com/blog/openai-api/

  • https://venturebeat.com/2020/06/01/ai-machine-learning-openai-gpt-3-size-isnt-everything/

  • https://www.zdnet.com/article/openais-gigantic-gpt-3-hints-at-the-limits-of-language-models-for-ai/

  • https://towardsdatascience.com/gpt-3-the-new-mighty-language-model-from-openai-a74ff35346fc


Thanks for reading!

Jaimin Desai - CEO & Co-Founder of Reconcile

www.getreconcile.com

*P.S. If you have any thoughts on our content, please leave a comment or email me at jaimindesai@reconcilemoney.com!