He introduced GPT-3 in 2020 was a tipping point for artificial intelligence. In 2021, this technology will strengthen the launch of one thousand new startups and applications.
The GPT-3 and similar models bring the power of AI to the hands of test seekers – and the results have been remarkable. Trained on trillions of words, the GPT-3 is a 175-billion-parameter transformer model – the third of its kind published by OpenAI. GPT-3 is notable for its ability generates human-like text, responses – in some cases, it forgotten. When requested by a user with text, GPT-3 can return consistent and timely emails, tweets, trivia, and more.
Suddenly, text emails, customer interactions, social media exchanges and even news stories can create automatically – at least in part. While larger companies are thinking about the problems and risks of generating text (remember Microsoft’s catastrophic Tai bot?), Startups have already begun volunteering for novel applications – and they will continue to lead in transformer-based innovation. OpenAI researchers first published the paper introducing GPT-3 in May 2020, and the use of some Nifty on Twitter became the focal point of the fast-paced activity.
Companies have formed at the top of GPT-3 using the model to create email trackers or marketing copies to create nutrition trackers or chatbot and more. Let others AIAD pass first when writing your emails or try using Broca or Snazzy for your ad copy and promotional material. Other young firms are accelerating their existing efforts, increasing the capabilities of their technical teams with a power of 175 billion parameters and on the other hand bringing solid products to market faster and with more speed and data than possible.
With some clever prompt engineering (a combination of instructions on a model with a sample output to help guide the model), these companies lift the built-in GPT-3 system to improve or expand the capabilities of the existing application. Of course, a text broadcaster can be a useful tool for shorthand notation – but powered by GPT-3, that shorthand can transformed into a product that creates contextually conscious email in your own writing style.
As early stage technology investors, we have been inspired to look at AI more broadly, and natural language processing in particular is becoming more accessible through the next generation of large-scale transformer models such as the GPT-3. We expect that they will unlock new usage cases and skills that we have not yet thought of.