The Real Risk of Google’s Race Riot; Why Microsoft Took a Stake in Mistral; Your Next House Could Be Built With LLMs

First, some news. As generative AI has seeped into a number of industries, from law to healthcare, industry-specific large language model startups have proliferated, as we showed in this analysis. But I was still surprised to come across a startup working to apply AI in a traditionally unsexy field: construction.

Parspec, a startup that uses AI to help distributors of lighting and electrical equipment, recently raised $11.5 million in seed funding led by Innovation Endeavors, CEO Forest Flager told me.

We’ll get back to this lower down. Before we do, though, we have to address Monday’s big developments in AI.

Google’s stock sank 4.5% and prominent technology figures continued to attack the company, as the racially inaccurate results from Google’s AI chatbot Gemini dominated conversations in the field. We’ve seen this drama before: last year, OpenAI suffered from similar criticism when ChatGPT appeared to give answers that were left-leaning. The ChatGPT problems even prompted an apology of sorts from OpenAI president Greg Brockman.

That kerfuffle eventually blew over. So, unless CEO Sundar Pichai gets hauled before Congress to answer for the mistake, this controversy should also be forgotten. Indeed, Google has the opportunity to make fixes while relatively few people are using Gemini. Google’s technology isn’t the problem—it’s how product managers shaped it.