Five OpenAI’s Steps to AGI, Stable Assistant Update, and Robots Using Gemini

PLUS HOT AI Tools & Tutorials

Jul 12, 2024

Hello and welcome to our weekly roundup!

Yeah, yeah, yeah, I know. Who even cares about AI now that Shrek 5 has been confirmed? But still. We've got a significant leak from OpenAI, a new model for diagnosing diseases, and a robot controlled by a chatbot.

So maybe I can get you interested too. Let's get started!

Where does astrology and AI intersect? Find out here:
AI and Astrology Weird Intersection
Ekaterina
·
Jul 9
Read full story

This Creators’ AI Edition:

Featured Materials 🎟️
News of the week 🌍
Useful tools ⚒️
Weekly Guides 📕
AI Meme of the Week 🤡
AI Tweet of the Week 🐦
(Bonus) Materials 🎁

Featured Material 🎟️

OpenAI Defines Five Steps From AI to AGI

According to Bloomberg, OpenAI has developed a way to track its progress toward building artificial general intelligence (AGI), which is AI that can outperform humans. It's a five-level classification system that has just been presented to the company's employees and will later appear in the public domain.

https://news.artnet.com/app/news-upload/2023/05/GettyImages-1240407936-scaled.jpg

And it's a pretty fair and conservative system. Let's see where we are now and where Sam Altman's startup aims.

Level 1

OpenAI executives told employees that the company is at the first level, which means AI is capable of interacting with people in dialog mode.

That said, Level 2 is very close—OpenAI believes it is on its doorstep.

Level 2 ("Reasoners")

The second level is called "Reasoners." The report says such models can solve problems as well as a person with a Ph. D.-level education.

Level 3 ("Agents")

OpenAI defines Level 3 ("Agents") as systems that can act on behalf of a user for days on end. That is, in effect, pretending to be human and not giving themselves away.

Level 4

The description of Level 4 (which does not yet have a verbal designation) sounds simple but vivid - it is AI capable of developing innovations.

Level 5 (“Organizations”)

The last, fifth level, is "Organizations." As the name implies, it implies the emergence of an AI that can perform all the functions of a full-fledged company—yes, without human employees. This model corresponds to what is commonly called artificial general intelligence (AGI).

Keep your mailbox updated with key knowledge & news from the AI industry

The company noted that the classification system is under development and may change as OpenAI receives feedback. The exact way the AI levels will be adjusted remains to be seen.

At what speed will OpenAI break barriers and transition from level to level? It's hard to say now, but Altman said last October that we're "five years, give or take," away from reaching AGI.

Incidentally, that's roughly in line with the prediction of Jensen Huang, founder and CEO of Nvidia. In March this year, he said that if we define AGI as AI that can pass any test put in front of it, it could be there within five years.

News Of The Week 🌍

Stability AI Releases Stable Assistant Features

Stability introduced its image and music generation chatbot less than two months ago, and it's already getting a major update. Here's what the new version of Stable Assistant has to offer:

Search & Replace: With it, you can specify objects in your own uploaded images and easily replace them with other objects. The developers say the tool will be helpful for creatives who want to change existing visuals without disrupting the overall composition.
Stable Audio: The feature for translating text to audio has also gotten better. Users can generate tracks up to three minutes long using licensed datasets from the AudioSparx music library. These tracks are generated at 44.1 kHz stereo.

You can try the new features now. Stability AI offers a free 3-day version for all users.

By the way, we have a newsletter about Stable Diffusion 3. If you want to know how to use it, you are welcome: https://thecreatorsai.com/p/stable-diffusion-3-is-free-for-everyone

Microsoft Leaves OpenAI's Board (And Apple won't come in)

Microsoft said it is giving up its observer seat on OpenAI's board of directors. Company officials said it has seen "significant progress," is confident in the startup's direction, and no longer sees a need for the seat. The decision is effective immediately.

Over the past eight months, we have witnessed significant progress from the newly formed board and are confident in the company's direction.

Around the same time, the FT reported that Apple had also abandoned plans for a seat on the OpenAI board.

While Microsoft claims significant progress as the reason, many speculate that the key reason lies in fear of regulatory scrutiny. If giants like Microsoft and Apple continue to grow their influence in the AI industry, they could run into antitrust problems.

Google Showes a Robot That Moves Around Office Using Gemini

Yes, chatbots are helpful not only in the virtual world and not only for online tasks. DeepMind researchers published a paper titled "Mobility VLA: Multimodal Instruction Navigation with Long-Context VLMs and Topological Graphs", showing what Google's Gemini 1.5 Pro model can do. Using this AI, the developers taught the robot to respond to commands and navigate around the office.

The interaction process was similar to how we work with voice assistants.

This news illustrates how AI is smoothly transitioning from virtual spaces to the real world. And that's fine.
Let's just be happy for Google and wish John Connor good luck.

Sharing is caring! Refer someone who recently started a learning journey in AI. Make them more productive and earn rewards!

Refer a friend

Samsung Releases Galaxy Z Fold 6 and Z Flip 6 with AI Features

The other day, Samsung held its annual Galaxy Unpacked, unveiling new smartphones and talking a lot about AI. In particular, the company promised to implement the Galaxy AI platform on 200 million devices worldwide.

The list of supported devices currently includes the Galaxy S24 lineup and the just-announced Z Fold 6 and Z Flip 6 (plus a few wearable gadgets).

Galaxy AI | Mobile AI and AI Features on Devices | Samsung Philippines

The main features of Galaxy AI are Portrait Studio, which creates portraits in various styles, and Sketch to Image, which converts drawings into images. With the first tool, users can create different portrait styles for their pictures, such as 3D cartoon or watercolor. The second comes in handy if you want to augment your images by drawing a sketch.

It's pretty hard to gauge the usefulness of the new AI without having the new Galaxy on hand. That said, I think Samsung has finally reached a point where foldable smartphones don't seem gimmicky. And given the advent of AI features, things are looking up for the Galaxy Z Flip 6 to be my new smartphone.

Bioptimus Launches the Largest AI Model for Disease Diagnosis

Let's talk about health briefly because we have some big news here. French startup Bioptimus has launched the world's largest open-source AI foundation model for pathology. With 1.1 billion parameters, H-optimus-0 is trained on a proprietary dataset of several hundreds of millions of images extracted from over 500,000 histopathology slides across 4,000 clinical practices, the company said.

This AI is expected to greatly simplify critical diagnostic tasks, from identifying cancer cells to detecting genetic abnormalities in a tumor.

I don't understand much about such highly specialized algorithms, so I'll just suggest you check out the features of H-optimus-0 on the developer's website.

Useful Tools ⚒️

SuperGPT – Use ChatGPT on Google Sheets

Almeta ML – Predict customer behavior on your site with ML

Rubbrband – Generate images in your own style, repeatably

Krut AI – AI co-pilot for e-commerce brands

Readsss – Listen to your X feed like an audio newsletter

Readsss is an app that turns written content from your feed into audiobooks or podcasts. You choose the accounts you follow, and Readsss compiles new posts for a daily audio feed. While multitasking, this lets you catch up on news, articles, or social media. Adjusting the playback speed lets you listen at a different pace and maximize your time.