The 5 Levels of OpenAI and Project Strawberry.

Today’s Motive:

• 💡OpenAI’s five steps to AGI.

• 🍓 Inside leaks reveal OpenAI’s Project “Strawberry.”

• 📱 Best unknown AI features for your Pixel.

AI News

💡OpenAI’s five steps to AGI raising some ?s

OpenAI’s pyramid of 5 levels to envision an AI public framework. The company says they are approaching level 2 and hints that AI is still in its early stages. What are these 5 steps and what does it mean?

1. Conversational AI: The step we are currently on at the moment. We are able to have engaging conversations, and AI is capable of creativity with text prompts.

2. Reasoners: An AI system capable of solving complex problems with human-level efficiency and reasoning. At this level, an AI should be able to problem-solve at a Ph.D. level, a significant leap toward human-level behavior. OpenAI feels close to this point.

3. Agents: At this stage, AI systems will operate autonomously without human input. These models can go on for days, acting on the user's behalf, taking on complex tasks, solving problems, and adapting to changing situations.

4. Innovators:At this point, AI models will be capable of generating new concepts and fresh ideas, breaking free from human input. AI systems will have surpassed obeying human commands and begin to contribute at or beyond our level.

5. Organizations: This is the point, AI system will take over all oversight of organizations, like a corporation. These systems will thrive as their own entities with capabilities to achieve goals and problem-solve beyond our level.

While this roadmap is theoretical, why would OpenAI come up with this now?
The company has stated they are currently at level one, and there is a ton of ground to cover before reaching AGI. Could this be a move to divert attention from current conflict and controversy within their framework or an investment ploy? OpenAI insinuates they are at the forefront of early AGI stages, so join now before it’s too late.

AI Leaks

🍓New Leaks within OpenAI reveal Project Strawberry.

OpenAI is looking to create an AI model capable of human-level reasoning and deep research under the secret code name "Strawberry."
For further reading, article posted on reuters.com by Anna Tong and Katie Paul, here. It’s really well done, and wraps up AI advancements quite nicely.

• At this point, it remains unclear how advanced Strawberry is and if it has human-level reasoning capabilities. OpenAI has hinted before that human-level AI thinking is their next big step.

• OpenAI did provide an employee-only hands-on meeting concerning a model with human-like reasoning earlier this week but gave no information in correlation to the program.

• The primary function is for the AI model to comb the net to prepare for complex tasks, a pretraining approach.

• Strawberry could be focused on Deep Research rather than other aspects. At this point, the project remains a work in progress, but the leak does reveal some interesting points.

Right now, most of the AI LLM hype is on new training methods. Just the other week Meta released their multi-token training approach that could revolutionize how models take in information on hugginface.co.

OpenAI has been a leader in the AI field - or the public figure - if they want to stay that way, they’re going to have make some big AI splashes soon.

AI Tools

🙋Discover 6 AI tricks with your Pixel.

This one is a little lengthy, but here are six cool AI tools you can use on your Pixel. Some of these are best used on line-up 8.

Full article written by Jo Snyder at Google Blog.

Curious if you snore or cough at night?

Your Pixel 8, Pixel 8a, or Pixel 8 Pro can track it for you.

How to use it:

• Go to Settings > Digital Wellbeing and parental controls.

• Tap Bedtime mode.

• Allow access and enable cough and snore detection.

• View cough and snore info under your weekly summary.

The feature processes sound frequencies into data for AI analysis without storing raw audio. Pair it with your FitBit to get comprehensive insights into your sleep habits.

Use your camera to create and scan PDFs.

Quickly digitize paper documents like receipts or contracts for quick storage, sharing, or filing. This is especially useful for work expenses, tax records, and any paperwork lost in that neglected pile in the corner.

How to use it:

• Open Google Drive on your Pixel 6 or later.

• Tap +New, then "Scan," or use the scan document icon.

• Use manual mode for precise framing or auto-capture for quick scans.

• Add, preview, delete, or clean up scanned pages.

• Press "Done," rename the file, and save it to your Drive.

• From Drive, you can send, share, or download your scan.

No more embarrassing text errors.

If you often send texts with typos, Gboard powered by AI is your solution. Proofread with Gboard fixes typos, grammar, and punctuation errors with a single tap before you send your message.

How to use it:

• Have Gboard running on your Pixel.

• Draft your text message.

• If errors are detected, the “Fix it” chip will appear on the suggestion bar.

• Tap the chip to correct the text.

• You can manually trigger Proofread by clicking the checkmark icon under the letter A on Gboard.

Unlock the valuable information hidden within images.

Need more info while watching YouTube Shorts or reading an article on your Pixel? Now you can get details without leaving your screen.

How to use it:

• Hold down the navigation bar on your Pixel.

• Circle, highlight, or scribble on the video, image, or text you want more info.

• Add context in the search bar at the bottom of the screen, like "What is this?", "What kind of car is the background?" for refined results.

Use Audio Magic Eraser to clean up those horrible video sounds.

Recording a birthday video, but background noise is ruining it? Audio Magic Eraser on Pixel can help by reducing or cleaning up unwanted sounds.

How to use it:

• Record a video on your Pixel 8, 8a, or 8 Pro.

• Open Google Photos, select the video, and click "Edit."

• Tap "Audio" to access Audio Magic Eraser.

• The feature identifies sound categories like talking, wind, or nature sounds.

• Choose “Auto” for automatic improvements or manually adjust specific sounds.

• When satisfied, tap “Done” to save your changes.

Use Gemini to summarize webpages.

Want a quick summary of a webpage? Use Gemini to save time and get the main points instantly.

How to use it:

• Set up Gemini as your mobile assistant.

• Open a webpage in Google Chrome.

• Long press the power button to bring up Gemini.

• Ask Gemini to summarize the article when the overlay appears.

• You can also voice-activate by saying, “Hey Google.”