Wolfram Research just announced a ChatGPT-powered [Wolfram] Notebook Assistant for their technology stack. Stephen provided an extensive writeup of the product on his blog: Useful to the Point of Being Revolutionary: Introducing Wolfram Notebook Assistant. Stephen’s introductory comment in that article is noteworthy:
Personally, I thought I was a pretty efficient user of Wolfram Language—but Notebook Assistant has immediately made me not only significantly more efficient, but also more ambitious in what I try to do. I hadn’t imagined just how useful Notebook Assistant was going to be. But seeing it now I can say for sure that it’s going to raise the bar for what everyone can do. And perhaps most important of all, it’s going to open up computational language and computational thinking to a vast range of new people, who in the past assumed that those things just weren’t accessible to them.
Through the years, I’ve read a bunch of Stephen’s code. He is a consummate and voracious user of his company’s product. It is strange to hear him use the phrase “raise the bar” referring to his own coding, but he did just that today. He is genuinely excited about this product – the possibility of mere mortals easily generating code in his fabulous language. Their success with this product says something interesting about LLMs: the specificity of the Wolfram Language API (and its documentation) was fertile training grounds for the OpenAI LLMs.
The Wolfram Language software is rather pricey: over 1/3 of an Alex for an annual hobbyist subscription and almost 3 Alexs for a commercial license. OTOH, Wolfram has provided a free version of their product on every Raspberry Pi for over a decade. AFAICT, this code generator will be bundled with that product. All that’s needed to run with this is access to the OpenAI “Plus” tier. With the vast improvements in capacity of the RPi, I’ve grown nervous that Wolfram would deliberately cripple the capability of this platform; that has not happened. An 8GB Raspberry Pi 500 sounds like a superb place to play with this software, and it looks like a 16GB iteration of the RPi 5 versions will happen early in 2025.
@Leo and Tom Merritt talked with Stephen in Triangulation #7 back in 2011. Mikah also recorded Triangulation #425 back in in 2019. Anyone curious about a TWiT-centric focus on this 1981 MacArthur Fellowship winner should check out those episodes.