AI can no longer be stopped

Welcome to edition #10 of the “No Longer a Nincompoop with Nofil” newsletter. I crossed 2000 subs recently and I just wanted to say thank you for all the support and kind messages. I appreciate it. You can now comment if you’d like to show support or discuss anything at the bottom of the post 👇

Here’s the tea ☕

• Too little, too late 🚀

• The one thing that will redefine society as we know it 💼

• This is what the future will look like 👀

The train that can’t be stopped

Some of the most respected people in the AI field have signed a letter to stop the training of large AI models like GPT-5 for at least six months, and I’m not talking about just some rich guys like Elon Musk and Steve Wozniak. I’m talking about legitimate researchers and engineers, many of which are recognised experts and award recipients in their respective fields. This whole AI situation has become quite peculiar.

Here’s the deal. I agree with the sentiment. I think we are definitely going too fast. But this letter is useless. It will do nothing. In fact, it will do even worse than nothing. It’s going to make AI look even more hyped than it already is, and in a really bad way. The fact of the matter is this - even if (it’s a big if) the US government does impose regulations to stop AI advancements, how exactly will they enforce this? What happens when other companies catch up? Will we ask them to stop as well? The letter has good intentions in that it hopes that governments can come up with suitable frameworks and legislation to mitigate the impending impacts of AI. Since when did any government get anything impactful done in a mere six months? The letter goes into detail about what it hopes we can achieve during our little AI interlude, and let me just tell you straight up, it is fantastical, dreamy and incredibly unrealistic.

Don’t get me wrong, I agree with the letter on the necessary regulations and oversights that are needed. But do I think they can actually be implemented? Not a chance. The letter says AI will have dramatic economic and political implications. I wholeheartedly agree. But there’s no stopping this train now. There’s only one way we see how this plays out, and that’s by diving head first into it.

The elephant in the room

There are a lot of unexplored issues with AI - privacy, security, scamming, democratisation, but the biggest elephant in the room is jobs. Jobs are kind of important, not only in the moral, ethical or financial sense, but in the we need these for the world to function the way it does sense, and AI is about to change the way the world functions forever.

Recently a former employer reached out to me and asked about using AI for his product (can I call myself an AI consultant now??). Here’s the thing. After some consideration and thought, it’s game changing for his company, but it also made me realise something. The job that he hired me for previously won’t exist anymore. In fact, a lot of jobs in a lot of companies won’t exist anymore. Millions of people are going to be jobless much sooner than we realise. ChatGPT was released four months ago, its plugins were released just over a week ago. They can do things we can’t even dream of yet, the possibilities are literally endless.

I’m not alone in my prediction either. Goldman Sachs predicts that over 300 million jobs will either be lost or degraded because of AI. This is across essentially every single field that uses a computer. People still don’t quite understand how AI is going to make so many jobs obsolete. Let me show you.

The future of work

There’s a concept I have purposely not mentioned even once in any of my newsletters - AGI (Artificial General Intelligence). AGI refers to an AI as intelligent as a human, and with the same ability to learn how to do things. The reason I don’t like speaking about AGI is because of a simple question we don’t really have an answer for - what is intelligence? How “intelligent” does an AI have to be for it to be considered intelligent? We don’t have a concrete answer to this. But could something like GPT-4 be considered a proto-agi? An AI that is not in any way conscious, but is capable of thinking and doing a large number of tasks just like a human. Let’s take a look at what that might look like.

This is an AI agent who’s only task is to figure out what its next task should be, given an objective. The AI first creates an entire task list and prioritises the tasks. Then it goes through and analyses how it will solve each task. It looks eerily similar to the clip I showed from Terminator 2 a few weeks ago. By the way, the objective can be absolutely anything, and yes, that includes anything bad, harmful or illegal. This is the type of agent you connect to the internet and suddenly you have an AI that can do literally anything.

Seems crazy right? What if you could use something like this right now? A research paper released a few days ago does exactly this, and its open source. You can literally use it right now. It essentially connects numerous AI models together to complete a task. Just take a look at this image.

This is the Jarvis I wrote about last week. An AI that can create its own tasks, plan its steps and find the appropriate means to execute them. In this case, a person still has to go and give the AI a task to complete, an objective to carry out. What if you even got rid of that step?

What if all you had to do was turn on your computer and AI would do the rest? What if AI could control your mouse and keyboard and write your emails, your documents, your code and do anything with the internet for you? Turns out they already can, and they’re already quite good at it.

Every single job that requires a computer is at risk. We’re about to live through the greatest change in human history. The future is already here.

Every week I compile a list of the crazy things happening in the AI space on reddit. You can check last weeks post here.

As always, Thanks for reading ❤️

Written by a human named Nofil