Going back underground with OpenAI’s latest model Back in 2020 I wrote an article about navigating the London Underground using GPT-3. Back then all kinds of people were getting GPT-3 to solve all kinds of problems, but nobody seemed to be testing it on spatial problems. I decided to see…

Streamlined web deployments of Keras models with preprocessing layers — Three years ago I published a story about deploying a Keras model using Flask. In that story I described how to set up a Flask server and a simple set of HTML pages to create a web deployment for a Keras model trained on a tabular dataset. This approach worked…

The ARC challenge and the future of AI Nobody can predict the future reliably, but history can sometimes give us an idea of how things will unfold. Today we may be on the cusp of an AI revolution, or we could be in for several more years of debate, disappointed…

Can ChatGPT find new life for punch card era code? Back in 2021 I wrote an article about using OpenAI’s Codex to translate COBOL code to JavaScript. Codex did a decent job translating simple COBOL programs. Now that ChatGPT is available, I wanted to see whether it was able to…

Lessons I learned getting this certification — In early 2022 I wrote the Google Associate Cloud Engineer (ACE) exam. I found the process of preparing for this exam extremely helpful. I learned about the key components of Google Cloud and got hands-on experience working with these components to solve problems. Just about every day in my job…

How much have things progressed in 2 years? Since I got access to GPT-3 in the summer of 2020 I have been astonished by how much it is capable of doing. When OpenAI announced the availability of ChatGPT, I decided to compare how it performed on some questions that had…

Comparing the results produced by each system with identical prompts Earlier this year I got access to DALL·E 2 and wrote an article to describe my initial experience with it. When Stable Diffusion was released this month, I was interested to see how it compared to DALL·E 2. Starting with…

Experimenting with images generated from text — I’ve had access to OpenAI’s GPT-3 for almost two years, and access to Codex for about a year. I’m grateful for the opportunity to experiment with these models, and I have blogged and created videos to record my experiences. After a wait of several months, this week I got access…

A new Google Cloud blog shows how to use Vertex AI to run DeepMind’s groundbreaking Alphafold protein structure prediction system — I spent several months in early 2021 interviewing data science candidates. In every interview, I asked the candidate to name two major AI accomplishments from 2020. Here are the two headlines I was looking for: GPT-3 — released in limited beta by OpenAI in June 2020 2. AlphaFold — announced…