Startupification
OpenAI was founded in 2015 by such Silicon Valley glitterati as Sam Altman, Elon Musk, and Peter Thiel. They assembled The Avengers of software – some of the leading talent who could create cutting edge AI infrastructure. In June this year, they released their threequel, GPT-3, and got a select group of beta tests to play around with it and see what they could create. And with this great power, these pioneers created buttons for websites that look like watermelons.
GPT-3 is another step in the evolution of AI, and the release reignited many of the questions that have been smouldering for years. What are the limitations of what it can create? What are the consequences for society? Who are the gatekeepers to the creations – and who gatekeeps the gatekeepers? And perhaps mostly importantly, what motivates OpenAI?
In this episode Matt and Steven give an overview of all these considerations. And don’t worry, they also explain just what GPT-3 is.
Timestamps:
1.45 What is GPT-3? The ultimate autocomplete tool.
2.45 A bit about the creators of GPT-3, OpenAI
7:50 How does GPT-3 work?
12:15 What are they spending all that money on?
16:30 What’s the driving force behind it?
21:00 A few examples of what GPT-3 can create
26:15 GPT-3 and the Turing Test
35:00 Understanding analogies
36:45 The consequences for society (and developers)
42:00 GPT-3 and NoCode
47:30 The future of work for data scientists
49:30 GPT-3 in the Stack
53:00 Ethical concerns