Andrew Mayne

The question of how much compute it takes to replace humans in the workforce is unnecessarily reductive.

At this year’s OpenAI Dev Day, I think we saw the next chapter of software unfold.

Seventy-two hours ago, OpenAI launched Sora, an invite-only app that has already climbed to the top of Apple’s App Store.

For the latest episode of the OpenAI Podcast I sat down with OpenAI president and co-founder Greg Brockman and Code engineering lead Thibault Sottiaux to talk about the release of OpenAI’s new GPT-5-Codex model.

With GPT-4 now stepping back from its starring role in ChatGPT, I want to share a few of my favorite memories from its launch.

In this video, we delve into the concept of AI hallucinations by comparing them to human cognitive errors.

tldr: Yes.

TLDR: There are multiple ways you can use GPT-4 with Vision to power robotics and other systems.

TLDR: How to boost GPT-4 with Vision’s capabilities with a simple prompt addition.

A recent paper The Reversal Curse points out an apparent failure in large large language models like GPT-4.

tldr: Custom instructions let you provide ChatGPT with relevant information of your choice about you and how you like to communicate.

tl;dr: One of the tricks to using GPT-4 well is understanding how it sees text and instructions When we started testing GPT-4 internally last summer one of the first tests people would try is getting it to play the game Wordle.

tl;dr: OpenAI is testing the ability to run code and use third-party plugins in ChatGPT.

tl;dr: You can use OpenAI’s ChatGPT to bounce ideas around and write story outlines Since I got into the field of AI and started working at OpenAI, it’s been interesting to see how things have accelerated.

OpenAI recently released Whisper, an open source automatic speech recognition model that’s incredibly powerful.

TL;DR: OpenAI has a new code generating model that’s improved in a number of ways and can handle nearly two times as much text (4,000 tokens.)

Large models like GPT-3 can perform a variety of tasks with little instruction.

GPT-3 is an exceptional mimic.

GPT-3 can remember hundreds of items and perform completions with them.

OpenAI’s GPT-3 is a highly capable general language model able to talk about almost anything.

We recently added three new endpoints to the API for GPT-3.

TL;DR: In an API call GPT-3 can recall details from a 1,500 word article and even repeat passages verbatim.

TL;DR: For many tasks you don’t need to provide GPT-3 with examples because it already understands what you want.

TL;DR: GPT-3 is much more capable than people realize when you utilize advanced prompt design that shows it what you want performed in a task then show it how to perform this task with a list of inputs.

Here’s a fun fact: OpenAI’s GPT-3 is actually a family of models, Ada, Babbage, Curie and Davinci that have different capabilities and speed.

When the first demos of GPT-3 content started to circulate it showed the amazing potential for a really smart language model to generate text and do cool things.

OpenAI’s API allows for incredibly intelligent text comprehension and transformation.

OpenAI’s API is incredibly versatile at comprehending text and completing passages, finding data and generating new information.

I’d like to show one of the API’s other superpowers: Its ability to expand a small thought into a larger expression.

Will artificial intelligence ever write a novel?

A few years ago I built an app for dictation that was intended to solve one of the biggest frustrations I had with other dictation applications: They didn’t handle punctuation.

OpenAI’s API is incredibly adept at picking up patterns in text.

A few months ago a remarkable thing happened: I got a glimpse of the future.

1.

TL;DR: I wrote a 50,000 word novel in less than 24-hours using dictation.