This episode is brought to you by Hubspot. Unlock unparalleled productivity with the ultimate guide to revolutionizing your workflows. Grab For Free Now

How do you eat your strawberries?

When asked that question at Wimbledon a few years ago, most top tennis players stereotypically answered “with cream.”

Except for one - two-time champion, Andy Murray.

His answer? “With my fingers.”

In this week’s episode, we’re going to shed plenty of light on OpenAI’s exciting new “Project Strawberry” and its potential implications. Which, by the way, are huge!

Before we dive in…

We're trying to get our ducks in a row about who's reading our emails. Care to lend a hand? It'll be a laugh, promise! Help us out: 4 questions, 2 minutes, 1 happy team!

Let’s crack on!

First things first:

What exactly is this mysterious project we’re talking about?

Here’s what we know so far:

Project Strawberry, formerly known as Project Q-Star (or Project Q*), is an ambitious initiative by OpenAI aimed at creating a next-generation AI model that can understand and reason about the world in more advanced ways than previous models.

We first heard about Project Q-Star in late November 2023 when Reuters reported that OpenAI researchers had written a letter to the board of directors warning that a new AI discovery the company had made “could threaten humanity.”

This was around the same time Sam Altman was sensationally fired, then reinstated, as CEO.

Could be a coincidence, of course, but we don’t believe in them…

In the more than eight months since Reuters broke the story of Project Q*, there has been plenty of speculation on social media about how it’s actually going to work.

As you’d expect, some of it has been pretty sophisticated. Some of it, not so much. 

With that in mind, here’s the gist of what we know at the moment:

In simple terms, Project Strawberry aims to create an AI that isn't just smart but also a little bit wise. It’s AI that thinks and reasons more like a human.

Speaking of smart, picture this, middle of the workday with a 5pm deadline… ChatGPT to the rescue, right?

But that pesky chatbot just won’t give you what you want. Hubspot’s ChatGPT guide to the rescue…

Revolutionize your workday with the power of ChatGPT! Dive into HubSpot’s guide to discover how AI can elevate your productivity and creativity. Learn to automate tasks, enhance decision-making, and foster innovation, all through the capabilities of ChatGPT.

Download the offer for actionable insights and tips on harnessing AI for a more efficient, impactful professional life.

Back to strawberry

Imagine current AI as being really good at specific tasks, like: 

However, these Level 1 “narrow” models often lack a deeper understanding of the world, especially when it comes to reasoning through complex problems that require a bit more common sense or logic.

Enter the big strawberry.

This Level 2 type of AI, usually referred to as Artificial General Intelligence (AGI), is all about pushing the boundaries by reasoning more like a human. The goal is to create an AI that can not only perform tasks but also think critically about them and solve new problems. 

What’s more, it can possibly make decisions with a level of understanding that is closer to human reasoning.

Unfortunately, the exact specifics of how Strawberry functions remain a closely guarded secret.

According to Reuters, two sources mentioned seeing demos earlier this year that OpenAI staff referred to as Q*. They could tackle tough science and math questions that current commercial models can't handle. 

Another source familiar with the details said OpenAI tested an AI model internally that scored over 90% on a MATH dataset, a benchmark for high-level math problems.

Now, that might not seem like a massive breakthrough at first.

But think about it:

Given enough time and compute, a reasoning AI model can develop exponentially to the point where it reaches singularity - and beyond.

Say what?

To put it simply, singularity is the moment when AI surpasses human intelligence and starts to self-educate faster than you and I can stop it. 

Basically, Terminator stuff. Once AI achieves singularity, we’re going to need someone like John Connor to save us from our artificial intelligence overlords.  

You might be wondering why, all of a sudden, we’re jazzed about Project Strawberry. Especially when it’s still a bit vague. Hush-hush even.

Fun fact:

When we asked ChatGPT on August 14 whether OpenAI was behind Project Strawberry, we got a blatant denial.

You can see for yourself:

Oooh

There’s plenty of buzz on social media, though. Especially on X, formerly known as Twitter.

Check this out:

On August 7, Sam Altman posted an image (not AI-generated) of strawberries growing in a garden, with the seemingly innocent comment “i love summer in the garden.” 

Well, don’t we all?

But since Sam Altman isn’t exactly your average horticulturist, the AI community went bonkers. As of August 14, this post/tweet has an eye-watering 6.5 million views, 12K likes, 1.2K reposts, and 1.2K replies. 

It goes without saying that this massive activity wasn’t all about the quality of Sam’s produce.

That was just the beginning, though.

It wasn’t long before an account named @iruletheworldmo and since dubbed Mr. Strawberry, which supposedly joined Twitter in August 2022 (we have our doubts), started churning out Project Strawberry-related posts faster than you can type your own name.

And guess who joined the party before too long? Yep, Sam Altman himself.

So, what can we make of all this? Honestly, at this point, your guess is as good as ours.

Our best bets right now are:

That being said, Mr. Strawberry has been predicting certain OpenAI releases and getting them right with a suspiciously high degree of accuracy. Hmmm…

So, could we possibly have a genuine insider’s view?

Stay tuned. Our gut feeling is that we’ll find out soon enough. 

Whatever’s going on, we suspect the release of Big Strawberry as a groundbreaking update to ChatGPT is just around the corner. 

But OpenAI is far from the only big player in the game coming out with new releases right now.

Here’s the scoop.

Remember the halcyon days of early 2023 when everybody and their brother were fighting tooth and nail to be the first to release a state-of-the-art AI model?

We know we do.

And guess what? Those days are back. Times 10, in fact!

While Project Strawberry has been all the rage in the last couple of weeks, X/Twitter, Google, and others have just dropped updates of their own solutions. 

Here’s what they’ve had to say so far.

Grok-2 Beta Release

Grok-2, released on August 13, is the latest advanced language model from the Grok family, featuring state-of-the-art reasoning capabilities. Available on the X platform, Grok-2 and its smaller version, Grok-2 mini, offer enhanced performance in chat, coding, and reasoning tasks. 

Tested as "sus-column-r," Grok-2 is said to have outperformed key competitors. We have a hunch X might have deliberately leaked this bit at first, only confirming it was official once they received positive feedback. Either way, both models are currently in beta and will be available via an enterprise API later this month, with improved security and global access.

Made by Google 2024

At the ninth Made by Google event, held on August 13,  the company showcased how its AI advancements are integrated across the Gemini platform, Android, and the latest Pixel devices, including the Pixel 9 phones, Pixel 9 Pro Fold, Pixel Watch 3, and Pixel Buds Pro 2. 

Google emphasized AI’s role in enhancing productivity, learning, and self-expression through its devices, supported by custom silicon chips. New AI-powered features were also introduced to improve accessibility and user experience across the Pixel and Android ecosystems.

Speaking of developments in AI… A quick ArtSmart tutorial to turn those low-res images to high-quality masterpieces!

Ever wished that a pixelated photo from your phone could magically turn into a crisp, high-resolution masterpiece? Well, dream no more! With ArtSmart’s Upscale feature, you can take your low-resolution images and boost them up to stunning 2K quality. 

But why should you care? Upscaling is like giving your images a makeover, perfect for breathing new life into old, blurry photos or enhancing your latest digital art. It’s all about making those precious pixels work harder for you!

Here’s how to use the Upscale Feature:

Once the process is done, you’ll notice that both the Detected and Estimated Sizes have changed, making your image sharper and more detailed.

Modes Explained:

Here’s the before–after results!

Curious about the magic? Take a peek at the before and after—watch as your image goes from pixelated to polished with ArtSmart’s Upscale feature!

Where were we? Ah right fun and Strawberries …

Let’s face it:

The identity of Mr. Strawberry might remain a mystery for some time to come. But over on Manifold, user Daniel Scott set up betting odds on whether the X account was actually AGI unleashed. 

Yup, people will bet on anything these days.

The question was resolved as “No” on August 13. But is this definitive? Well, you might want to hedge your bets for the time being.

With cream on OpenAI’s ChatGPT while waiting for Project Strawberry to drop? 

With your fingers on X’s Grok 2.0?

Or perhaps while gazing at the stars with Google’s Gemini?

Hit the reply button and let us know your thoughts on the escalating AI arms race.

We might even feature your input in an upcoming episode as we continue to follow the latest developments.

As always, we’ll see you next week!