🤔 McDonald's is closing their AI powered drive thru

"If it's flipping hamburgers at McDonald's, be the best hamburger flipper in the world. Whatever it is you do you have to master your craft." - Snoop Dogg

In today’s newsletter:

• McDonald's is closing their AI powered drive thru

• Papers

• Learn

Penses

"BA-DA-BA-BA-BA I ain’t lovin’ it.” The tune going through customers heads as they trialed McDonalds AI-powered Drive-thrus.

Back in 2022, The Register posted this headline:

Then these TikToks start coming out:

@resinsbiren

#greenscreen @mcdonalds #nine #sweettea #mcdonalds #robots #friday #caffeine

@typical_redhead_

I thought TikTok would appreciate this 💀 #fyp #foryou #foryoupage #comedy #fail

It wasn’t surviving in the wild. They were losing money.

This got me to thinking about how many real-world tests went into to this? Knowing the tech world and the speed to solutions that happen, I don’t think that many. I would love to interview someone from that team. But tech often fails we rarely consider how it would be used by the end user.

• Did that document drive through scenarios across the busiest locations?

• Did they stage reproductions of those scenarios and game plan how to reduce risk?

• How about accents? My southern drawl is rarely understood by chat bots.

Development doesn’t stop because we tuned the Natural Language Processing (NLP) model and added an interface. As Data Scientists, we need to understand our user issues just as much as the customer facing team members or we can’t build a great product.

Paper

History of generative Artificial Intelligence (AI) chatbots - A through guide of the history of chatbots so you can chat up your colleagues. It goes from rule-based systems to AI models, how it’s been integrated into different sectors, and the ethical considerations. (Total pages: 31) Read more here.

BioImage.IO Chatbot: A Community-Driven AI Assistant for Integrative Computational Bioimaging This chatbot implementation is kinda cool. I would’ve never thought of this use-case. They are streamlining computational bioimaging workflows. It uses a curated knowledge to make it all happen. Read more here.

Learn

How it works

Contextual Chatbot with LLMs