Published 6 Feb 2023
Michael and Michael start things off for this edition of Games at Work with some follow up stories about smart & connected appliances, as there was a flurry of articles on this topic in the prior week. First up, is an article from The Atlantic, where the author explores the idea of subscription services for ink, and how this changes the nature of the purchase of the printer to something that is no longer owned, to pages as a service. Then, there’s a post by Dan Goodin on a self-driving automobile with no safety driver that was causing problems for firefighters in San Francisco.
Then, turning to all things AI, Michael and Michael launch in with a New York Post article about AI driven Armageddon. Next up is a story about companies decoding your brainwaves, which causes Michael R to reminisce about the Emotiv headset, and Michael M to remember cameras being used to identify micro expressions (both of which were part of e361, link below!) There are additional articles about GAN networks for determining vulnerabilities, concerns with uses of ChatGPT that may train the corpus with corporate secrets, ChatGPT word problem problems and running large language models on the new Apple M2 MacBook Pro.
Quite a big episode on AI – with a few links that Michael and Michael did not get a chance to cover in the podcast. Meta testing members only Horizon Worlds metaverse experiences, a superb jigsaw puzzle web app (thanks, Andy!) and a Super Mario infused random text generator.
What appliances do you want to connect (or disconnect!) from the Internet? 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. That’s our story and we’re sticking to it.
Selected Article Links
“Smart” Appliances redux
The Atlantic article: My Printer Is Extorting Me
Self-driving automobiles are a solution in search of a problem. People don't want them, but companies like Uber, Amazon, Cruise and Waymo keep pushing them because there's so much money to be made. I hate the way these companies are compromising our safety and the way government regulators aren't pushing back.
Firefighters were battling a major house fire near the intersection of Hayes and Divisadero streets early in the morning of Jan. 22 when a Cruise vehicle with no safety driver started to creep its way into the emergency scene.
Two firefighters stood in front of the car to prevent the vehicle from driving over hoses used to douse the growing inferno, but that didn’t work. As the car continued to inch forward, one firefighter took quick action and smashed the vehicle’s front window, finally bringing the car to a stop. First responders contacted Cruise, who sent workers to move the vehicle out of the way.
UPDATE: I have already walked back the statement above that no one wants this. Fair enough. Many of you want this.
Lots of you pointed out the number of accidents human drivers have. There is zero evidence that self-driving cars will have a better safety record. And anecdotal evidence like the dangerous incidents in the very limited San Francisco trials suggests AVs may be less safe.
If you want AVs, you should volunteer to have your city be the guinea pigs for this potentially fatal tests. San Francisco ought to ban AVs until there's data they are safe.
The San Francisco Standard article: SF Officials Describe Chaos From Cruise, Waymo Cars as They Try To Slow Their Rollout
AI
New York Post article: Rogue AI ‘could kill everyone,’ scientists warn as ChatGPT craze runs rampant
Wikipedia article on The Terminator (movie)
Popular Mechanics article: Companies Already Have the Ability to Decode Your Brainwaves
Games at Work e361 Ancient Games & Ancient Texts
Games at Work blog post: Getting All Emotiv(tional)
Schneier on Security article: AIs as Computer Hackers
Amazon Begs Employees Not to Leak Corporate Secrets to ChatGPT
This issue seems to have come to a head recently because Amazon staffers and other tech workers throughout the industry have begun using ChatGPT as a "coding assistant" of sorts to help them write or improve strings of code, the report notes.
The Byte article: Amazon Begs Employees Not To Leak Corporate Secrets To ChatGPT
Wall Street Journal article: ChatGPT Needs Some Help With Math Assignments
I got this to work! https://til.simonwillison.net/python/gtr-t5-large
I managed to run a language model on my laptop!
I ran https://huggingface.co/sentence-transformers/gtr-t5-large – a sentence transformers model, which can turn sentences of text into a 768 dimension vector, suitable for things like related content searches
Here are my detailed notes: https://til.simonwillison.net/python/gtr-t5-large
See also my explanation of embeddings from a few weeks ago: https://simonwillison.net/2023/Jan/13/semantic-search-answers/
Bonus links
TechCrunch article: Meta starts testing ‘members-only worlds’ in Horizon Worlds
Podcast: Play in new window | Download (Duration: 32:26 — 37.7MB) | Embed
Subscribe: Apple Podcasts | Spotify | Amazon Music | Android | Podcast Index | Youtube Music | RSS | More
Great show guys! Sorry I missed the recording.