CrowdStrike, Pickle Jars & Billionaire Bland
5 in 5 - Brave & Heart HeartBeat #210 ❤️
This week we'll be doing a CrowdStrike deep dive, why Meta are pulling their Llama AI from the EU (three guesses and no prizes) and what actually happens if you set up a blank Meta account?
Plus, Open AI have come up with a way to set AI against each other (for the good of research, apparently...) and is the pickle jar theory the original productivity hack?
Let’s get into it.
Were you forwarded this? Not a subscriber? 👉 Sign up here
#1 - The CrowdStrike Outage
Unless you have been living under a rock, and we mean literally, you will know that there was a global IT outage last week that brought the world to it’s knees.
And unless you’re an IT expert or have a lot of time on your hands, you might now quite understand what happened, or why.
So, strap in…
Last Friday morning TV broadcasters, banks, and even airlines were brought to a standstill by the infamous Blue Screen of Death.
This outage was caused by one single company – CrowdStrike. The most important company you’d never heard of, CrowdStrike helps companies find and prevent security breaches, describing themselves as the “fastest” threat finders in the business.
They serve around 29,000 customers, over 500 of which are on the Fortune 1000 list.
Many of which were unable to do pretty much anything this Friday.
One failed update, which appears to have installed faulty software onto the core Windows operating system used by CrowdStrike, caused the systems to get stuck in a “boot loop” – with systems showing an error message saying “It looks like Windows didn’t load correctly”.
Error messages are frustrating enough when you’re trying to watch Netflix, never mind when the whole London stock exchange has gone offline, the airport is full of incensed travelers with no planes to get on, or the 911 emergency call center you’re calling drops off the radar.
And we can only imagine how frustrating it’s been for the IT team working behind the scenes at CrowdStrike to try and fix this enormous mess.
They say any publicity is good publicity, but we bet the CEO is wishing he could go back to a time when we didn’t all know the company’s name…
#2 – Meta Pulls Llama From EU
While the EU continue their crusade against big tech, Meta have joined Apple in fighting back, kind of.
Meta have just announced that they won’t be launching their new multimodal AI model, Llama, in the EU due to the “unpredictable nature of the European regulatory environment.
It’s not like the EU are begging them, please please give us a multimodal AI model we’re desperate for it, but they’ve brought them into the fight anyway.
This is such a level of diss, that The Guardian said that it “underlines tensions between big tech and brussels amid a tougher regulatory environment”.
This follows on from Apple, who deprived EU customers of some of their Apple Intelligence rollouts due to their rules around 3rd party apps and how that would apparently affect the security of the options.
We wouldn’t want to give any kind of opinion on the subject but it COULD be suggested that Apple MIGHT do such a thing to try and get back at them for making them give up their Apple Store app monopoly. Allegedly.
Big Tech seem to be giving the ultimatum “if you’re going to have tough privacy rules and regulations then you can’t have fun AI features”.
Maybe that’s a good thing? Maybe the onus should be on the AI features to be secure enough, rather than the rules in the countries you’re releasing them to be lenient enough?
#3 - What Happens If You Set Up A Blank Meta Account?
The Guardian Australia wanted to find out how Facebook and Instagram’s algorithms affect what you see in your home feed, and to do that they unleashed them on a blank, unused smartphone, linked to a new, unused email address.
Three months later, with no input at all, they were “riddled” with sexist and misogynistic content.
The profiles were set up as generic 24-year-old males, with no Facebook friends and a phone that opted out of ad tracking. Instagram forces you to follow at least five accounts, so to keep it neutral they made the accounts follow popular recommended accounts like the Australian prime minister.
It was an interesting test, considering that Meta say the algorithms rank content according to what people show an interest in, so what would an account with no input at all show you?
Well, for Facebook it turns out the answer is offensive memes, sexist and misogynistic images, and for Instagram, women with very few clothes on and foodie influencers.
One of the misogynistic memes on Facebook referenced the Japanese character for “woman” and the Japanese character for “noisy” which looks like three of the character for women”, the others we cannot even repeat.
An associate professor of digital media at the University of Queensland said this experiment showed that serving up this kind of content to young men on Facebook is “baked into” the algorithm model, basically Meta’s default is to show sexist content to young men, and that’s concerning.
Meta says that they assess posts for signals on who would find a post “meaningful” such as who posted it and how popular it is. This is what they assume a young man would be interested in, with no other input from him.
There are some important questions. Why? What are they trying to achieve? And do they have a responsibility not to serve up this sort of content even if they think the person will like it…
#4 - Open AI Are Researching Again…
But is it enough?
AI has faced criticism, notably from us, for seemingly rushing too quickly and too recklessly with their AI development. One of their ex-safety researchers famously said that they prioritized “shiny new products” over safety.
OpenAI are here to prove the haters wrong, and show that they DO take AI safety seriously.
Last week they showcased research that they claim could help researchers scrutinize AI models as they become more capable and useful. One new technique involves having two AI models have a conversation that forces the more powerful one to be transparent with it’s reasoning – basically explain how it’s doing what it does so that humans can understand what it’s “up to”.
It’s basically a technique to use one AI to make sure you can’t get tricked by another AI, which is innovative if nothing else.
One of the new research crew said that this kind of research is core to their mission of building a “safe and beneficial” AI.
And all this time here we were thinking that their core mission was churning out the most advanced AI possible to make the most money possible to keep Sam Altman in mega sports cars and Hawaiian estates!
#5 - The Pickle Jar Theory
The pickle jar theory is kind of like the original “productivity hack” - but is it really useful?
The pickle jar theory of organising your day involves imagining a pickle jar, or really any jar, and filling it up with all the events in your day.
The big stones are the important and urgent things you need to get done, the smaller stones the less urgent tasks. Then, you fill the jar up with sand, which signifies smaller or more beaurocratic tasks like emails, and then water, which represents your personal life.
The point seems to be to help you realise that you can only fit a finite amount of tasks in your day - and you need to pick which ones you need to put in.
However, if you’re no good at time management or prioritisation in the first place, it’s not really that useful.
And how big is this jar meant to be anyway?
it’s a no from us… no better than a to-do list in our opinion, sorry.
Brave & Heart over and out.
Bonus
Billionaire Bland
Why do the tech billionaires dress like that?
You know exactly how we mean, plain t shirts and jeans, boring limp hair and nondescript trainers.
Are they trying to look inconspicuous, or do they literally have no style?
To find out more on how you can retain your top talent, or how we can help you with digital solutions to your business and marketing challenges, check out our case studies.