As part of our DALL·E 2 research preview, more than 3,000 artists from more than 118 countries have incorporated DALL·E into their creative workflows. The artists in our early access group have helped us discover new uses for DALL·E and have served as key voices as we’ve made decisions about DALL·E’s features.

In order to share the magic of DALL·E 2 with a broad audience, we needed to reduce the risks associated with powerful image generation models. To this end, we put various guardrails in place to prevent generated images from violating our content policy.

We trained a neural network to play Minecraft by Video PreTraining (VPT) on a massive unlabeled video dataset of human Minecraft play, while using only a small amount of labeled contractor data. With fine-tuning, our model can learn to craft diamond tools, a task that usually takes proficient humans over 20 minutes (24,000 actions). Our model uses the native human interface of keypresses and mouse movements, making it quite general, and represents a step towards general computer-using agents.

We trained “critique-writing” models to describe flaws in summaries. Human evaluators find flaws in summaries much more often when shown our model’s critiques. Larger models are better at self-critiquing, with scale improving critique-writing more than summary-writing. This shows promise for using AI systems to assist human supervision of AI systems on difficult tasks.

Large neural networks are at the core of many recent advances in AI, but training them is a difficult engineering and research challenge which requires orchestrating a cluster of GPUs to perform a single synchronized calculation.

Cohere, OpenAI, and AI21 Labs have developed a preliminary set of best practices applicable to any organization developing or deploying large language models.

Codex is now powering 70 different applications across a variety of use cases through the OpenAI API.

Early users have created over 3 million images to date and helped us improve our safety processes. We’re excited to begin adding up to 1,000 new users from our waitlist each week.

We’re happy to announce several executive role changes that reflect our recent progress and will ensure continued momentum toward our next major milestones.

Goodhart’s law famously says: “When a measure becomes a target, it ceases to be a good measure.” Although originally from economics, it’s something we have to grapple with at OpenAI when figuring out how to optimize objectives that are difficult or costly to measure.

We’ve released new versions of GPT-3 and Codex which can edit or insert content into existing text, rather than just completing existing text.

We describe our latest thinking in the hope of helping other AI developers address safety and misuse of deployed models.

Call for expressions of interest to study the economic impacts of large language models.

Announcing GPT-NeoX-20B, a 20 billion parameter model trained in collaboration with CoreWeave.

We built a neural theorem prover for Lean that learned to solve a variety of challenging high-school olympiad problems, including problems from the AMC12 and AIME competitions, as well as two problems adapted from the IMO.