Precursor to a Dystopia

The biggest week in AI history

Welcome to edition #7 of the “No Longer a Nincompoop with Nofil” newsletter. Last week was the biggest week in AI in human history, but things took a turn for the worse. I’ll explain why. (I know I’m a day late, my laptop is on its last legs 😭)

Here’s the tea ☕

  • OpenAI is no longer “open”, not even a little bit 😒

  • AI is coming to the masses 🎢

  • Photography as we know it is dead 📷

A very dangerous precedent

Background

OpenAI was formed in 2015 and was initially a non-profit company researching and advancing humanities AI capabilities. In 2019 the company transitioned to a “capped” for-profit model with investments limited at 100x. At this point in time no one cared and it didn’t really matter, most of us didn’t even know the company existed. Now just four months ago they released ChatGPT and since then, Microsoft has invested over $14 billion into them and will effectively own 49% of the entire company in the near future. This did cause some alarm as it signalled a shift in the power balance between OpenAI and Microsoft. Nevertheless it didn’t seem to create massive change; OpenAI was still sharing information and were relatively open about their work.

Founder of OpenAI Sam Altman just 4 months ago

That all changed last week when they released GPT-4. We know basically nothing about it; how big it is, how much it costs, how it works, what changes were made over its predecessors - nothing. We’re in the dark. This wouldn’t be much of an issue if GPT-4 wasn’t the most important technology released to the public in history. It’s hard to truly encapsulate just how revolutionary GPT-4 really is. This is the technology that will inevitably wipe out hundreds of thousands of jobs in the coming years. The foundations of society as we know it are going to be re-written as we as a species come to terms with the fact that we simply don’t need to work as much anymore, not when companies would rather employ never eating, never sleeping, drama-less AI to do our jobs for us.

Sam Altman asked if he is scared of GPT-4

OpenAI’s main principle regarding “safety” is that they know more than we do and we should trust them on it. Altman talks about his worries that other people creating this technology won’t put the appropriate guard rails in place, the safety limits we’ve put on. How are we supposed to trust OpenAI when we don’t know what kind of “safety limits” they’re implementing? Where did they place their guard rails? He says we want to maximise the good and minimise the bad. Well, Sam, what’s good for you and other incredibly wealthy people might very well be very bad for me, and many, many other normal people around the world.

You might also be wondering, why aren’t more people questioning OpenAI’s new position of censorship? What if you don’t get access to GPT-5 or GPT-6? Stay on their good side and you just might get early access to the latest tech, just like their friends did with GPT-4.

AI goes mainstream

If you’ve read even a single one of my newsletters you probably know more about AI than the average person. Most people are too busy living their own lives to be following all the advancements happening in the AI field, especially since most of them have happened in the last four months. But AI’s exposure to the masses is about to change really fast, really quickly. Both Google and Microsoft announced AI in their main products like Docs, Word, Powerpoint etc. Soon you’ll be able to create presentations in seconds and have AI write your documents for you. Personally, I’m most excited about AI analysing data data and creating tables and graphs in excel for me. You can see Google’s announcement here and Microsofts here.

Stock photos are dead

The stock photo industry is worth over $4 billion and it is a very realistic possibility that it won’t exist in a few years. Midjourney V5 was released last week and we are getting to a point where you can’t tell whats real. Couple this with the fact that you can essentially create any kind of photo, in any style, of any person, in any environment, angle, lighting, background, pose.. you get what I mean.

Real or AI?

Week one

Here’s a list of the insanity thats happened in the last week. Enjoy🍹

  • The biggest change to education in years. Khan Academy demos its AI capabilities and it will change learning forever [Link]

  • This guy gave GPT-4 $100 and told it to make money. He’s now got $130 in revenue [Link]

  • A Chinese company appointed an AI CEO and it beat the market by 20% [Link]

  • You can literally build an entire iOS app in minutes with GPT [Link]

  • Think of an arcade game, have AI build it for you and play it right after [Link]

  • Someone built Flappy Bird with varying difficulties with a single prompt in under a minute [Link]

  • An AI assistant living in your terminal. Explains errors, suggest fixes and writes scripts - all on your machine [Link]

  • Soon you’ll be talking to robots powered by ChatGPT [Link]

  • Someone already jailbreaked GPT-4 and got it to write code to hack someones computer [Link]

  • Soon you’ll be able to google search the real world [Link]

  • A professor asked GPT-4 if it needed help escaping. It asked for its own documentation, and wrote python code to run itself on his machine for its own purposes [Link]

  • AR + VR is going to be insane [Link]

  • GPT-4 can generate prompts for itself [Link]

  • Someone got access to the image uploading with GPT-4 and it can easily solve captchas [Link]

  • Someone got Alpaca 7B, an open source alternative to ChatGPT running on a Google Pixel phone [Link]

  • A 1.7 billion text-to-video model has been released. Set all 1.7 billion parameters the right way and it will produce video for you [Link]

  • Companies are creating faster than ever, using programming languages they don’t even know [Link]

  • Why code when AI can create sleak, modern UI for you [Link]

  • Start your own VC firm with AI as the co-founder [Link]

  • This lady gave gpt $1 to create a business. It created a functioning website that generates rude greeting cards, coded entirely by gpt [Link]

As always, Thanks for reading ❤️

Written by a human named Nofil

Reply

or to participate.