generated 31 August 2025
Published 8 September 2025
e528 with Michael, Andy and Michael – stories about AI image editing with Nano Banana, GAN enabled LLM evolution with R-Zero, MentraOS open source smart glasses, automotive software, Making Monsters, Kazeta and a whole lot more.
Michael, Andy and Michael get things rolling with the Nano Banana image editing software from Google. While the generated and altered images are very sophisticated, there are still a few tells that the photos came from AI. An example from the Washington Post article calls out the “AI gibberish” replacement of numbers on the phone keypad – while the replacement of the human in the phone booth with a water buffalo replete with smart ring is ultra realistic. Andy’s ChatGPT generated nano banana is a fun visualization for an atom-sized banana, even though he was “AI-splained” by the chatbot that “a banana at that scale couldn’t exist in any realistic way.” Ha!
The team touches on a couple more AI stories dealing with with the Fast-VLM video captioning model and a generative adversarial network method of self evolving reasoning LLM with R-Zero. Next is a springboard for the MentraOS open source smart glasses operating system that reminds the team of Andy’s experience in 2024 with the Brilliant Labs Monocle.
Then the co-hosts talk about automotive software – and the challenges posed by the need to troubleshoot and correct for the intersection between evolving software and existing hardware. The frequency for software updates for a vehicle, phones and more requires a level of testing and integration that can be very frustrating when things don’t work as expected. Understatement of the year, I’m sure.
Wrapping up the episode are a couple of games – a kickstarter called Making Monsters, Office Job, which has a television sized screen and suitcase sized mouse and Kazeta for cartridge gaming.
What’s been your most frustrating automotive software experience? Have your bots 🤖 drop our bots 🤖 a line at @gamesatwork_biz (our home for now) and let us know!
These show notes were lovingly hand crafted by a real human, and not by a bot. All rights reserved. That’s our story and we’re sticking to it.
Selected Links
AI
Washington Post article: Masterful photo edits now just take a few words. Are we ready for this?
9 to 5 Mac article: You can try Apple’s lightning-fast video captioning model right from your browser
FastVLM-webgpu on Huggingface Spaces
Venture Beat article: Forget data labeling: Tencent’s R-Zero shows how LLMs can train themselves
R-Zero: Self-Evolving Reasoning LLM from Zero Data paper on arXiv
Unacceptable situation to wear camera glasses of the moment: while doing a bikini wax
https://futurism.com/wax-center-meta-glasses
MentraOS Open Source Smart Glasses OS on Github
Games at Work e453: Vision Pro a Pro-Pro (for Brilliant Labs Frame)
Games at Work e436: Squishy Purple Doom (for Andy’s experience with Brilliant Labs Monocle)
Automotive Software
TechCrunch article: BMW, I am so breaking up with you
BMWblog post: Apple’s iOS 18 Update Is Causing BMW Digital Car Key Problems: Solutions Inside from October 2024
Games
Making Monsters on Kickstarter
Games at Work e455: Star Trek vs Douglas Adams (for Spore)
Felix Fisgus post: Office Job from 2022
hackster.io article: Alesh Slovak’s Kazeta Turns Mini-PCs Into ’90s Throwback “Cartridge”-Based Games Consoles
Podcast: Play in new window | Download (Duration: 30:59 — 43.2MB) | Embed
Subscribe: Apple Podcasts | Spotify | Amazon Music | Android | Podcast Index | Youtube Music | RSS | More
[…] « e526 — Beans and Bricks […]