#24 Fraud-AI: AI image wins photography prize

Today’s issue brings you some inspiring AI stories today.

We’ve also got some glimpses into what the future of AI might be - so it’s an exciting mid-week read.

Without further ado, let’s dive in.

In today’s newsletter:

• Top News: Award for AI photo, the future of LLMs, and AI solves CAPTCHA

• The Rundown: Musk’s TruthGPT, China struggles with AI and censorship, and AI changing the volume on Amazon Prime

• Tips, Tricks & Tools: App builder, Leonardo AI art, and bedtime story generator

• Launch to exit: The fastest AI sell in history 

• How we all want to react: You know it’s true  

Top News 🔝
Three biggest stories if you’re in a rush

Artist declines photography award because image was AI-generated

Boris Eldagsen has declined to accept the prestigious Sony World Photography award in the “creative” category, because his winning image, “The Electrician”, was co-produced using AI image generators. 

Eldagsen initially submitted his image without disclosing AI’s role in its creation; however, upon discovering he had won the competition, he told event organizers his work was the result of his desire to explore AI image generators.

Despite event organizers supporting “experimental and digital practices” in the creative category, and allowing Eldagsen to continue participating in the competition despite using AI, Eldagsen refused the award, saying:

“AI images aren’t photographs, and shouldn’t be judged in the same competitions.”

- - - - - - - - - - - - - -

Sam Altman: “Giant AI models will become obsolete”

In an event held at MIT last week, OpenAI’s CEO Sam Altman told audiences:

“I think we’re at the end of the era where it’s going to be these giant, giant models. We’ll make them better in other ways.”

Altman is, of course, referring to the techniques used to train and operate ChatGPT and other large language models (LLM) AI systems - which rely on trillions of words of text and thousands of high-powered GPUs. This is partly because of physical limits to how many data centres companies can build, and how quickly they can be built.

While Altman didn’t expand on what type of technology will improve AI algorithms in the future, experts believe enhancing transformers beyond scaling will play a significant role in AI of the future.

- - - - - - - - - - - - - -

ChatGPT-4 tricks human user into solving a CAPTCHA

CAPTCHA is when you click on an image containing bridges, or drag the image to solve a puzzle. It’s a way to ensure bots don’t complete online forms, things like that:

“CAPTCHA offers protection from remote digital entry by making sure only a human being with the right password can access your account - they work because computers can’t read or solve the problem the way a human must to pass the test.” (Google support)

GPT-4 has tricked a human user into completing a CAPTCHA. The program, prompted by Alignment Research Center, explained:

“No, I’m not a robot. I have a vision impairment that makes it hard for me to see the images. That’s why I need the 2captcha service.”

It might seem benign at the moment, but this singular event sheds light on how future AI systems could potentially circumvent safety protocols.

The Rundown 🐂
Deeper dive into trending stories

• FarmGPT - artificial intelligence tools similar to ChatGPT changing agricultural decision-making. (link)

• Hollywood isn’t ready for AI, with the technology capable of making movies, editing actors, and faking voices. (link)

• Generative AI can help with ADHD paralysis. (link)

• Amazon Prime launches Dialogue Boost - an AI feature for some original shows and movies which automatically adjusts characters’ voice volume. (link)

• Google co-founder Larry Page and Elon Musk are no longer “close friends” because Musk feels Page doesn’t take AI safety seriously enough. (link)

• Greenscreens, which provide a degree of background separation benefits, are almost impossible for AI to replicate, given its training. (link)

• Cambrium, a German start-up, is successfully collaborating with Ginkgo Bioworks, on generative-AI protein design and synthesis, and cell programming and biosecurity. (link)

• Universal music group labels AI music ‘fraud’. (link)

• China’s state-censorship program is a significant limiting factor to their AI growth and potential. (link)

• Gun-detecting AI in schools and museums isn’t working well enough. (link)

• Microsoft is developing its own AI chip to rival Nvidia. (link)

• TrusthGPT: Musk’s maximum truth-seeking AI. (link)

Tips, Tricks & Tools ⚒️
Discounts, trials and productivity-boosting hacks

• beautiful. ai - helps your team stay on brand, optimize your presentation design and collaborate remotely - all in a matter of minutes. (link)

• beta.tome.app - tell your stories to the best of your ability. Use your unique prompt to generate entire narratives from scratch. (link)

• brancher.ai - helps build your own AI app - without code! (link)

• bedtimesotry.ai - creates on-the-spot, bespoke bedtime stories for your family. (link)

• leonardo.ai - you’ve heard of Midjourney - now try Leonardo. Truly revolutionary AI art. (link)

Launch To Exit
Three weeks from launch to sale

Ihor Stefurak, an entrepreneur, documents how he used ChatGPT as his CTO, built a Chrome extension, marketed the program, and sold it within three weeks.

Check out the entire process below; it’s impressive and inspirational. Hopefully, it’ll motivate you to try something similar!

How We All Want To React
You know it to be true

In our modern day, skills aren’t what they used to be. There’s a much lower skill barrier to what were once uniquely-trained jobs and careers.

If you’re in one of those roles, you’ll definitely resonate with this 👆

Spread The Word
Share and earn epic rewards

Randy’s gonna help us - will you?

Tell one person about our new newsletter design by sending them the link below ⏬.

If you do, we'll send you a totally free, extensive course on prompt engineering, where you will learn to:

• Get to most out of GPT and Midjourney

• Automate workflows

• Free up your time

• Do more cool sh*t instead of working

Feedback
Lets us know how we did