artificial intelligence Archives

midjourney rendering of a speculative city environment with colorful infrastructure

Imagining cities and their infrastructure using artificial intelligence, and playing around with some of the weirder parameters in Midjourney to see what we can squeeze out of the generative AI platform.

Artificial Intelligence

July 17, 2023July 17, 2023 Nat M. Zorach

Prose Rife With Burstiness & Perplexity: AI & Its Jargon

Artificial intelligence models can generate outputs we can understand from conversational input. But how does an AI narc on itself? We look at a couple of the rationales for why one author says that an AI detector “thought” that the US Constitution and the Book of Genesis were almost certainly written by AI– and unpack some more details about the mysterious innerds of the model.

friendly chatgpt robot traveling with human friend

Culture & Language Travel

July 13, 2023September 5, 2023 Nat M. Zorach

How ChatGPT Saved Me While Traveling

Hiking up an Icelandic glacier is already pretty daunting, but it can be made a bit less daunting with an AI assistant. It can’t help you climb, but it can explain some missing puzzle pieces.

artificial intelligence ai and the demise of the usable internet

Artificial Intelligence Business & Economics Tech

July 12, 2023July 12, 2023 Nat M. Zorach

Two Things Making The Internet Unusable In 2023

The tech sector is trying to both expand the use of artificial intelligence while clinging on to the dying days of an epic business cycle. Both are contributing to making the internet, IoT, and any consumer product even remotely connected to the internet, increasingly unusable, increasingly resource-intensive, and increasingly frustrating.

midjourney ai rendering cities city neighborhoods

Art & Design Artificial Intelligence

July 3, 2023July 3, 2023 Nat M. Zorach

Unpacking AI Training Biases in Midjourney

Artificial intelligence can produce imagery of whatever you ask it for. The question is: How accurate is the AI’s interpretation? In some cases, it sounds like it’s based more on the prejudices and stereotypes held by the people who design and train the model. That’s a big problem.