e462 — Generating Both Sides

two sides of a caution cleaning in progress sign.
Photo by Oliver Hale on Unsplash

Published 15 April 2024

Michael and Michael get together to talk tech while Andy is away.  In this episode, the team focuses on a variety of AI topics including AI assistants, the Humane AI Pin, LLM training and deepfakes before touching on several AR topics and concluding with a couple of games.

The first story deals with human forgetfulness vs AI assistants ability to remember everything.  Luckily the co-hosts remember dealing with this topic long ago.  One “lifelogging” article from 2014 references Steve Mann’s work, and intriguingly, the recording device looks very similar to the Humane AI Pin.  Humane’s AI Pin has had a number of reviews in the past days, and the one from The Verge gives a very thorough treatment.  Check it out in the show links below.

An article from ZD Net reports that transcripts from YouTube videos were used to train LLMs. Now that the Games at Work podcast is available on YouTube, Michael and Michael wonder if the podcast may be on the ingest list for training the next set of LLMs.  

After considering a WSJ article about financial companies actions to counteract potential identity deepfakes, Michael and Michael take a look at a LEGO powered AI security system.

Switching gears to AR, Michael and Michael enjoy a story about the Dyson CleanTrace which allows you to see in AR where you’ve already vacuumed and where you still have to.  Making cleaning a game is not new for the Games at Work crew but this is a new take on how to combine AR, gaming and cleaning that may have many other possibilities.  After taking a look at CT scans of Apple’s Vision Pro and Meta’s Quest, Michael R shares some insight on the updated Persona for the Vision Pro. 

The team wraps up the episode with a couple of games.  Michael R shares a 8 bit (or 1 bit) game called Nothing Door.  Michael M shares an article about the recent Sims expansion pack that allows the player to take on the role of a landlord or tenant.  

Do you think that tracking vacuuming is the killer app for AR?  Perhaps extreme ironing?  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 Article Links

AI

Wired article: Humans Forget. AI Assistants Will Remember Everything

Games at Work e157: Through the 2016 Looking Glass

SingularityHub article: Lifelogging Gear Is Small, Cheap, and Powerful, So Like It Or Not, You’re Going To Be Recorded

The Verge article: Humane AI Pin review: not even close

ZD Net article: OpenAI and Google reportedly used YouTube transcripts to train their AI models

Wall Street Journal article: Deepfakes Are Coming for the Financial Sector

Upgrade your home defense

— CatSalad🐈🥗 (D.Burch) :blobcatrainbow: (@catsalad) 2024-04-07T09:02:33.639Z

AR / VR

iMore article: Dyson turned this viral Apple Vision Pro vacuuming concept into a feature, and it’s coming soon

Wikipedia article: Extreme ironing

Games at Work e60: Bubbly Bubblers in Gamified Buildings

Chore Wars

9To5Mac article: Apple Vision Pro scans show meticulous array of cameras, sensors, and more

Lumafield article: Apple Vision Pro and Meta Quest Non-Destructive Teardown

Upgrade podcast e507 – Differently Social

Games

NothingDoor on the App Store

Slate article: The Sims: For Rent Allows Millennials to Turn the Tables

The Sims 4 For Rent Expansion Pack

e461 — Lawsuits & Swimsuits

suits in a clothing store
Photo by Alexander Naglestad on Unsplash

Published 8 April 2024

Michael, Andy and Michael get together to talk tech — focusing on a very healthy dollop of security and open source, with some GenAI art, and an updated set of AR swim goggles thrown in for good measure.

First, a little housekeeping – you may now find the Games at Work podcast on YouTube!   If everything worked just right, this very episode will be automagically posted to the YouTube channel, replete with the suits image from Unsplash.

After getting this announcement out of the way, the show focuses on security and open source topics, starting with the xz open source backdoor.  Andy posted on this subject on March 30, and gives an overview of the amazing story in this episode.  A New York Times article also covers the situation, detailing how close it was to have the code included in major Linux distros.  Be sure to check out the XKCD comic “Dependency” from the show notes below.

For those that are intrigued by class action lawsuits, check out the post about the Incognito browser mode action and settlement.  Another suit takes aim at preventing the unauthorized reuse of a creative’s name, image, voice and likeness through generative AI.  Check out the Ars Technica article in the show notes below for the complete explanation. 

Then, sticking with generative AI and the art world, the co-hosts discuss Cody Nash’s post on the output of image generation to create a completely white image.

The team wraps up the episode with the most current version of the FORM Smart Swim 2 goggles.  Michael R shares his experience with the earlier version and how much he likes them.

Do you use incognito mode in your web browsing?  What would be your prompt to create a blank white image through GenAI?   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 Article Links

New way to listen

YouTube link – https://www.youtube.com/@GamesatWorkdotBiz/podcasts 

Security

Catching up on the detailed background to the xz compromise. This is a good summary (with links to the detail) of the long play here. boehs.org/node/everything-i-kn

— Andy Piper (@andypiper) 2024-03-30T12:18:31.069Z

Schneier on Security post: xz Utils Backdoor

New York Times article: Did One Guy Just Stop a Huge Cyberattack?

XKCD comic: Dependency

The Verge article: AT&T confirms data breach and resets millions of customer passcodes

Schneier on Security post: Class-Action Lawsuit against Google’s Incognito Mode

AI

stackoverflow post: Is AI making your code worse?

Ars Technica article: George Carlin estate forces “AI Carlin” off the Internet for good

Cody Z Nash post: Can AI Create a White Painting?

AR

T3 article: Form Smart Swim 2 Goggles review: the world’s best swimming wearable just got better

Form Swim

Games at Work e335: Horsing Around

Games at Work e270: Virtual Bricks

e460 — AskEmilyPost: AI etiquette

dinner place setting
Photo by Karen Sewell on Unsplash

Published 1 April 2024

Andy and both Michaels get together to talk tech — focusing on AI generated music, wearable etiquette and AR user experience with a dollop of security & privacy.  

Starting off with a reprise from last week’s episode, the co-hosts enjoy a discussion about music sparked by Suno, which uses GenAI to create new songs.  This harkens back to the conversation from last week, and extends with Andy raising the Rock Family Trees documentary and Michael R bringing up the Pandora Music Genome Project.  Michael M remembers the discussion on Alex Murell’s The age of average, and how this might apply toward generating music.

Next up, the team takes a look at Harper Reed’s post on using an LLM to parse information captured from a variety of sensors in an office setting and post it to the office discord channel.  This is pretty familiar to the co-hosts, and it brings Andy joy knowing that MQTT still surfaces in context like this one.

The Heinz Remix allows the conversation to flow to creating custom condiments.  Then, the discussion turns to wearables and etiquette of when and how to wear them.  An article from The Verge takes on all manner of potential discretions and indiscretions when it comes to the use of technology.  Passive wearables such as the Oura Ring, to smart watches and other machines that go “bing” get their deserved attention.  Devices with cameras pose special challenges, especially for people that are not familiar with the device and don’t know whether they are being recorded.  

Next up after the etiquette discussion are some further ideas for augmented reality user experiences.  An article from BGR describes how the Apple Pencil could be used by a Vision Pro wearer, and this reminded Michael R of the 3Doodler.  This led directly to a game called Screenbound which presents a sidescroller game combined with a 3D experience – have a look at the link below to see how this works.  

The team wraps up the episode with a couple of stories on security, and retro Flight Simulator & Commodore 64 topics.

What wearables would you wear to a wedding?  Which would you leave home, and why?   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 Article Links

AI

The Verge article: AI-generated blues misses a human touch — and a metronome

Games at Work e459: Pendulums in the Stack

BBC article: Rock Family Trees

Wikipedia article: Music Genome Project

Alex Murell article: The age of average

Games at Work e413: AI, your way

Harper Reed’s blog post: Our Office Avatar pt 1: The office is talking s*** again

Games at Work e60: Bubbly Bubblers in Gamified Buildings

Fast Company article: Kraft Heinz, king of condiments, is now letting you make your own concoctions

Kraft-Heinz press release: Kraft Heinz Unveils HEINZ REMIX™, The First Customizable Digital Sauce Dispenser

AR etiquette

The Verge article: The principles of wearable etiquette

AR UX

Boy Genius Report article: New Apple Pencil might work with Vision Pro, which sounds weird but is actually genius

Post by @tomwarrenuk
View on Threads

Security

Interesting and scary thread over at Xitter describing what appears to be a series of targeted attacks designed to hijack iCloud accounts by doing something that causes the user's device to be inundated with push OTP requests. The idea seems to be that if they send enough requests, the target might eventually click yes — either by accident after denying it the 59th time, or because they just want to make the prompts stop.

It's worth noting that in the end game of this attack, the scammers apparently relied on data from people-search services to gather the target's data and contact the user directly posing as Apple. And when you ask them info about yourself to verify you, they can usually read off enough details to fool people into thinking they're actually talking to Apple. And then they ask you to verify a one-time code, and if you do that, your account is toast.

twitter.com/parth220_/status/1

— BrianKrebs (@briankrebs) 2024-03-24T16:56:12.867Z

Games

I mean, it's just a but you can actually run and 4 on a bloody microcontroller! 😮​

.oO(and it's literally the same experience I had back in the day on my 10MHz PC XT clone)

youtube.com/watch?v=UwWX2_GvtD

— Wintermute_BBS (@Wintermute_BBS) 2024-03-22T07:37:43.767Z

Steam: Screenbound

The Verge article: 8BitDo’s latest retro keyboard is an ode to the Commodore 64

e459 — Pendulums in the Stack

metal desktop pendulum
Photo by engin akyurt on Unsplash

Published 25 March 2024

Andy and both Michaels get together to talk tech — focusing on AI generated music, security & privacy.  

Starting off with the news about Suno.ai, the co-hosts discuss what generative AI music might be like, and what it may mean from a copyright perspective.  This prompts Michael M to remember the Wired CD published in 2004, specifically designed to be ripped, sampled and remixed.  

The team takes a moment to consider the news of the United States Department of Justice bringing a complaint against Apple as outlined by the post from Sixcolors.

Then, the team moves to a plethora of articles dealing with security and privacy.  An intriguing post from IanVisits details how pendulums are used as a source of randomness at the Cloudflare offices in London.  Next up is a story about Glassdoor now requiring real names, and the fallout from that.  Then, a discussion about The Stack, and how you can determine whether your content has been included in it or not.  Last is a TechCrunch article dealing with LinkedIn’s reported plans to add gaming to the platform.

What AI music would you generate?  Are you in the Stack?  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 Article Links

AI

Gadgets360 article: Suno, a ChatGPT-Powered Chatbot, Can Generate AI Music Using Text Prompts

Suno.ai

Free Music Archive — The Wired CD: Rip. Sample. Mash. Share.

Games at Work e458: Generative Podcast (Andy and Michael’s Version)

Apple

Sixcolors post: U.S. versus Apple: A first reaction

Security & Privacy

IanVisits post: Inside the London office where swinging pendulums keep cyber threats at bay

Office Snapshots post: Cloudflare Offices – London

Ars Technica article: Users ditch Glassdoor, stunned by site adding real names without consent

Glassdoor

Fishbowl

HuggingFace bigcode/in-the-stack post: Am I in The Stack?

Techcrunch article: LinkedIn plans to add gaming to its platform

Bonus Links

New World Notes post: BREAKING: Linden Lab Partners With AI Company To Create NPCs To Help New Second Life Users

hackster.io article: This Cardboard Boba Fett Helmet Includes a Raspberry Pi-Powered Thermal Vision Mode and Mini-Map