OpenAI could be about to drop Project Strawberry

Rumours are swirling that OpenAI is about to drop a bombshell this week, and it's got the AI world seeing red — strawberry red, to be precise.

 By the CogX R&I team

August 14, 2024

A cryptic photo of four strawberries shared by CEO Sam Altman has set off a frenzy of speculation about the company's next big thing: Project Strawberry.

What's the big deal about Strawberry? If the leaks are to be believed, Strawberry could be a quantum leap in AI capabilities. Imagine ChatGPT not just answering questions, but independently solving complex problems and verifying its own output. This could revolutionise scientific research, supercharge decision-making in fields like healthcare and finance, and accelerate technological innovation across the board.

Is this just another hype train? Social media has been awash with strawberry emojis, and even Altman himself is fanning the flames. When a X user cryptically tweeted about a "level two" breakthrough, Altman's response? "amazing tbh". If that's not a wink and a nod, I don't know what is.

More than just a new model launch: OpenAI's roadmap suggests we're currently at the chatbot stage, the first step towards AGI. Strawberry could usher in level two: reasoners.

While it's still uncertain if we’ll see a dedicated reasoning model, it certainly feels like something new and exciting is on the horizon.

GPT-4o is also making headlines

Its new voice mode is apparently so realistic it could potentially make users ‘emotionally attached’ during conversations. In the latest safety assessment for GPT-4o, OpenAI warned users could become overly reliant on AI for companionship, potentially impacting their real-life relationships.

Now read the rest of the CogX Newsletter

Machine learning is not “just” statistics

Guest Author: Anil Ananthaswamy

“Once you encounter all these algorithms, methods and models, you’ll be hard-pressed to dismiss machine learning as just glorified statistics.”

You might have heard arguments that machine learning is nothing but glorified statistics. I beg to differ. My perspective comes from having been a former software engineer, and now having researched and written my book, Why Machines Learn: The Elegant Math Behind Modern AI. For the book, I had to relearn coding after a 20-year hiatus. Two decades ago, I used to be a distributed systems software engineer, in the pre-ML/AI days. As I learned python and ML, I was intrigued by the change in thinking warranted by machine learning, when it comes to solving problems: ML-based techniques are distinctly different from non-ML methods.

From a software engineering perspective, you have to flip your instincts about how to solve problems: from thinking algorithmically to learning how to pose questions of the data you have in hand, and use machine learning to, well, 𝘭𝘦𝘢𝘳𝘯 the model to represent the data, which can then be used for inference/prediction/generation. You have to learn how to 𝘴𝘦𝘦 data differently: 𝘢𝘴 𝘢 𝘳𝘦𝘱𝘰𝘴𝘪𝘵𝘰𝘳𝘺 𝘰𝘧 𝘢𝘯𝘴𝘸𝘦𝘳𝘴 𝘵𝘰 𝘺𝘰𝘶𝘳 𝘲𝘶𝘦𝘴𝘵𝘪𝘰𝘯𝘴.

Machine learning-based techniques are not just statistics. Yes, statisticians build sophisticated models of the patterns that exist in data and use these models to infer/predict. But ML is so much more than simply writing code to automate what statisticians do.

… want to keep reading? Check out the full OpEd here on the CogX Blog

Midjourney v6.1 VS Flux. 1: which AI image generator wins?

Midjourney Image generated from Midjourney under prompt: TedX Speaker presenting to crowd

Flux.1 Image generated from Flux.1 under prompt: TedX Speaker presenting to crowd

Closed Source image generation platform. Better at skin rendering and illustration capabilities Has stricter censorship, and you cannot train Midjourney with your own image input Paid access available by signing up through Discord or on the Midjourney website - gives access to unlimited generations, concurrent project processing and more Tune images by upscaling, select specific images to adjust or re-running the prompt

Open-source image generation model Better at generating dynamic compositions, backgrounds and text rendering Can generate images using copyrighted characters, celebrities, politicians  Free access through different platforms such as BasedLabs, HuggingFace, Poe and more Paid subscription allows better performance generation; with Flux. Pro generations available for commercial use

One cool thing
A groundbreaking robotic system has completed the world's first fully automated dental procedure. This AI-powered dentist promises faster, more accurate treatments, from fillings to crowns. Could this be the future of dental care?

Also in the news

Think you got what it takes to beat a robot at ping pong? You might want to think again. Google DeepMind has trained a robotic arm that’s giving human players a run for their money. According to the company, It's the first time a robot has been taught to play a sport with humans at a human level.

AI's leap in healthcare: Researchers have developed a model capable of predicting heart disease with an astonishing 95% accuracy. By effectively identifying true cases while minimising false alarms, this AI tool holds immense promise for improved patient outcomes.

Humane AI Pin has hit a snag: The $699 AI Pin, touted as a revolutionary wearable, is facing a tidal wave of returns. After a glowing launch in April, negative reviews and user frustration have led to a situation where daily returns are now reportedly outpacing sales.

In case you missed it

California-based robotics company, Figure, has unveiled its second humanoid robot model, the F.02: