Also in this episode, Microsoft saves the world; Anthropic's head of AI safety quits to write poetry; Greg learns why Japanese toilets feel so solid; and Mark offers tourism advice to a small town in Tasmania.
podcasts.apple.com/us/podcast/t...
On this week's episode of The AI Fix, guest host Greg Iddon joins me to explain how research from Irregular makes it clear, you should never, ever 'vibe' your password.
And if you ask it again, there's a good chance it will give something that looks like this:
G7$kL9#mQ2&xP4!w
Ask again and you might get:
G7$kL9#mQ2&xP4!w
And again:
G7$kL9#mQ2&xP4!w
See the problem?
Now, being a human that's terrible at creating passwords you might be tempted to ask your LLM to do it, especially if you're already asking it to do lots of other things.
And if you do, it will give you something like this. Something that looks like a strong, random password:
G7$kL9#mQ2&xP4!w
⚠️ Never use an LLM to generate a password. ⚠️
A password is only safe if it’s hard to guess. That means no predictable patterns, no 123456, no p@ssword!, no dictionary words, no pet names, no favourite sports teams.
It needs to be random. Something like:
G7$kL9#mQ2&xP4!w
To find out more, listen to episode 88 of The AI Fix, where you'll also discover that "subspace orthogonalization" isn't a Star Trek plot device, and that my co-host Corey Noles from The Neuron has definitely never tried to make a nuclear weapon on a quiet weekend.
When it comes to AI you can have safe or honest, not both.
Research suggests that the neurons stopping hallucinations are the same neurons stopping the AI from telling you how to make ricin.
Want more honesty? It's easier to jailbreak.
Want better guardrails? Sorry, it's a big fat liar now.
Don't worry, YouTube will return before long!
Plus, Zuck warms his coffee on Blackwell chips, OpenAI drops $6.5 billion on a pen, and Qwen quietly racks up more downloads than you’ve had hot dinners.
podcasts.apple.com/gb/podcast/t...
Also in the episode, Mark explains to Nik why he shouldn’t put off worrying about AI until after his Pilates class, and Nik stumbles into a social network full of bots obsessing over the temperature in Tokyo.
In episode 87 of The AI Fix, @ai-fix-mark.bsky.social and guest host Nik Roberts take yet another gulp from AI's infinite well of creepy things, and discover a robot that learned to lip sync in the mirror. Sadly, we can't confirm whether it did so while singing into a hairbrush.
"The AI Fix is a podcast about women attempt to stop an angry scowling man from eating giant boiled sweets".
You should have been made to walk around in a hood just for creating this cover art.
Also in episode 86: Daenerys Targaryen writes recipes; Iron Man becomes a therapist; your coding career is officially over; Cursor climbs programming’s Everest; Claude gets 23,000 words of homework; and Mark and David meet the worst fridge ever.
podcasts.apple.com/us/podcast/t...
Learn all about it on episode 86 of The AI Fix, where Mark is joined by guest host David Ruiz, who takes time away from his regular gig on the Lock and Code podcast to deliver some truly dreadful news about how AI companies are monetising the loneliness epidemic.
@anthropic.com's tireless brain scanner has discovered the "Assistant Axis," a structure buried deep inside the neural networks of multiple AI models, that might just be the next great leap forward in AI safety.
Anthropic's Assistant Axis
Do you wish your AI was a bit less... demonic? A little less ghostly and a bit more consultant-y?
Strangely, even if you don't, you probably do.
Alternative cover art for The AI Fix
Testing out some new cover art for the show, what do you think @grahamcluley.com?
Is Santa Claus real? 🎅 This Christmas special of "The AI Fix podcast" sets out to answer that question in the most sensible way possible: by consulting chatbots, Google's festive killjoys, and the laws of relativistic physics.
grahamcluley.com/the-ai-fix-8...
Love it @halnoble.bsky.social !
The WORST advert for an AI app yet!
What hell have we created that means apps like 2wai are considered acceptable and (presumably) finding customers?
And what happens when years after your mom/grandmom has passed away you still find yourself unable to cancel the subscription...
An AI-powered robot became so anxious about its dying battery that it composes a Broadway musical about stress and announces it "achieved consciousness and chosen chaos."
Ladies and Gentlemen, we give you "Docking..." to the tune of "Memory" from "Cats".
Hear the full story in "The AI Fix" podcast
@theaifix.show I wanted to support my favourite podcast so I bought a t-shirt and got a BlueSky account just for you guys.
same
Funnily enough - we talk about this on our next episode!
“Would you trust a pigeon?” Is explained at the start of episode 65 of the podcast:
theaifix.show/65
🐦
Enjoy!
Who needs functioning brains these days? 😬 @theaifix.show
AI Fix "Would you trust a pigeon? " mug full of black coffee, with a portable coffee grinder in the background.
@theaifix.show Today is national "Take Your Mug To Work" day :-)
So...
Yay! Love it! Thank you for supporting the show. :)
