Hello! Lots of stories this week, aren’t you lucky?

This week was kind of like a $5 Amazon gift card (what the hell am I supposed to do with that and why is it in dollars?) 🤷‍♀️. Scroll down mindlessly for these things:

• There’s an open letter going around calling for a pause on AI development — excited to see how ineffective it is

• OpenAI are now letting certain companies develop plugins for ChatGPT to make your able-bodied life even more frictionless than it was before

• France decides to ban more than just TikTok from government devices

• Twitter Blue will fully replace the legacy verification system by 15th of April

• Something something the pope in a coat…

⏸️ Can we all just PAUSE for a sec?

The Future of Life Institute has called on all AI labs to pause experimentation for six months — and there’s even an open letter with over a thousand names on it. I am ALMOST convinced that, in the world of over hyped technology, six months is just about long enough for everyone to forget that generative AI even exists. But, in reality, all that will happen (I mean, if we do get the six month pause), is a brief respite from all the tech commentators going on and on about the risks and benefits of this new technology — yes I am very aware that I am basically describing myself, leave me alone.

I am also confused about what this group of people expect anyone to achieve in just six months. Regulators take six months alone just to open an email that might be warning them of the dangers of AI. Then they take another six to ponder over the email carefully. Then another six to maybe draft a white paper about it (but probably not).

The open letter itself uses alarming language and makes ridiculous claims, including that the developers of AI should essentially lobby policymakers to regulate it. This is such an unrealistic expectation; usually technologists engage in aggressive lobbying to make it easier and faster for them to deploy their rancid products — so good luck with that!

Tbh the Future of Life Institute look like bullshit artists; they say that they “believe that the way powerful technology is developed and used will be the most important factor in determining the prospects for the future of life.” — and so therefore we should do everything we can to ‘improve’ those prospects. Not only is this a horrendously woolly thing to say, but also it seems to take humans (and anything that isn’t technology) completely out of the equation.

If you want a more comprehensive breakdown of how misleading this open letter is, read this post from AI Snake Oil.

🔌 ChatGPT has plugins now

What does this mean? Basically, if/when these are deployed beyond the alpha and beta stages, you won’t have to interact directly with certain apps or websites — you can get ChatGPT to interact on your behalf. But more broadly it means that ChatGPT is finally going to be able to connect with the wider internet, instead of regurgitating answers from a corpus that only goes up to 2021.

Once again I would like to point out that this is a weekly newsletter and as such I ingest tech news basically every day — and yet I am still struggling to keep up with generative AI’s developments, possibilities, and unsolicited productivity advice from tech bros on Twitter. But don’t worry, I’ve managed to wring out my brain to get some thought-juices out 🧃.

First of all, without sounding like an irrational evangelist/alarmist, ChatGPT plugins have the potential to change the way we interact with apps and websites f o r e v e r:

• Consider this: you no longer have to go to a website to compare insurance packages; ChatGPT will just do it for you. It provides a ‘meta-layer’ between you and the actual app or service that you wish to use

• So in this sense, the service itself becomes less important, and the only ‘thing’ you really ‘use’ is ChatGPT. Interaction with digital products and/or the internet could be reduced to a single interface, such as the OS voiced by Scarlett Johansen in the movie Her.

• This shift can be likened to the gradual digitisation of crucial everyday services we’ve seen over the last twenty years: banking, finding a room to rent, applying for benefits, and paying bills, are all things that are ‘easier’ to do online.

I often think that the New and Convenient way of doing every-day tasks eventually becomes The Only Way — and when you only have one way of doing something, you often leave behind those who exist on the fringes of the key demographics that most technology is designed for; the people who maybe have accessibility needs that technologists only think about as an after-thought (or never think about at all!)

This is exactly what drives digital exclusion and it needs to fucking stop. When you look at the list of companies currently developing plugins for ChatGPT (e.g. Expedia, Slack, Instacart, OpenTable) it is abundantly clear who this technology is made for, and who it will inadvertently exclude.

Employing the use of an effective virtual assistant to book flights sounds perfectly useful and convenient to me, but I don’t have any specific accessibility needs and I also already know how to do that myself. There are 10 million people in the UK who don’t have the digital skills needed to do even the most basic things online — do we really expect everyone to become a prompt engineer just so they can check their bank balance?

🚫 The banning of apps on government devices continues, and nothing materially changes

Last week I wrote about the ridiculous intellectual somersaults the US government were doing to position TikTok as a major threat to national security — and this week the fear-mongering continues: just like many jurisdictions before them, France have proudly announced that they will ban TikTok from government devices, as a way of maintaining national security.

Oh except France will also ban streaming services and gaming apps. This might sound completely unhinged but to me it makes more sense than only banning TikTok if the logic is that these apps collect way too much data (they do), are not very secure (probs not), and use opaquely governed algorithms (yep).

Of course, the overall problem with banning these apps from government devices is that it will materially change nothing; apps like Candy Crush have obscene data practices because governments — who now want them banned — have failed/refused to put in place any meaningful laws/regulations around this. Instead, they wilfully created the conditions for large scale data processing, behavioural ad targeting, and algorithmic amplification — and now they don’t have the power to do anything except complain, apparently.

🐣 Twitter Blue For You

In a classic case of Being Accidentally Funny, Elon announced on Monday that by the 15th of April, the legacy verification system for Twitter will be gone, and the only people with a blue tick will be the ones who were dumb enough to pay the $8 a month. Furthermore, the ‘For You’ page will only contains posts from Twitter Blue subscribers, and only Twitter Blue subscribers will be able to vote in polls.

Excuse me, but let’s zoom out:

• Twitter Blue has proven to be extremely unpopular. Less than 0.2% of Twitter’s monthly active users pay for it.

• So that means that your feed of recommended tweets will only come from about 300k people

• That ALSO means that all algorithmic amplification for non-paying users will be gone.

This will either drive even more people away from the site, or begrudgingly force people to pay to even get slightly noticed on what is already the hardest platform to get attention on if you are a small account.

I just find it SO FUNNY that these seething, middle-aged, Silicon Valley men are failing to understand that the main draw of social media is that it provides an endless stream of free content without having to do any piracy. Since 2020, it feels like all they’ve been doing is trying to take that away.

In 2021/22, before AI was being injected into every empty feature, it was NFTs. They were trying to take money from users and make it seem cool — it was hilarious. Nearly all of those projects have shut down now. The new trend is to literally just charge you a monthly subscription to get any meaningful use out of the platform

✝️ The first AI image to trick even the most aggressively online people

Midjourney has had an upgrade, and people have worked out that with the right prompts, it can spit out really good photo-realistic images. Here is the pope in a puffa jacket that (apparently) looks like it was made by a brand called Balenciaga. I’m not someone who cares about fashion so I’ve never heard of them before now, but from what I can gather, they sell really ugly trainers that are already covered in dirt? Good for them!

Literally, the only dead giveaway is the weird coffee cup type looking thing in his left hand — everything else is basically perfect. Should we be afraid? No; you should have already been afraid. You have no idea what’s going to happen to you; one day an AI will memorise your likeness, and inject a much better looking version of yourself into unskippable package holiday adverts. You won’t be able to even go on holiday because your job would have been replaced with a very good NLP engine. Are you prepared? Of course you’re not! You’re just sitting there trying to recreate an image which is as good as the one above.

Anyway, have a good weekend!