OpenAI unveils SearchGPT

ALSO: Meta’s Llama 3.1 and more

Howdy, Futurists!

Welcome to the latest edition of "The Foresight AI" newsletter.

The future of AI is loading...

Here’s the scoop for you today:

  • OpenAI reveals 'SearchGPT'

  • Meta’s new Llama 3.1

  • Google’s AI-powered weather model

  • Grok 3.0 gears up for release

  • How to Manage Your Data on X

  • The Job Interview Coach Prompt

  • More updates on AI and tech

Estimated Read Time: 4-5 minutes

BREAKING NEWS

OPENAI

The Quick Look:

OpenAI has unveiled SearchGPT, a search feature designed to give “timely answers” to questions, drawing from web sources.

The Closer Look:

  • SearchGPT aims to provide up-to-date information with AI-generated summaries.

  • It’s designed to enhance conversational capabilities with real-time web data.

Launch Date?

The prototype is in testing, with plans to integrate its best features into ChatGPT.

The Importance:

SearchGPT could significantly alter how we interact with information online, offering a more insightful search experience.

META

The Quick Look:

Meta has introduced Llama 3.1, a large AI model with 405 billion parameters.

The Closer Look:

  • Supports multiple platforms including Google, Nvidia, and AWS.

  • It’s positioned as a cost-efficient and open alternative to closed models.

  • Demonstrates strong capabilities in general knowledge, steerability, math, tool use, and multilingual translation.

The Importance:

Llama 3.1 represents Meta’s commitment to advancing AI technology and its potential applications.

GOOGLE

The Quick Look:

Google’s researchers have developed NeuralGCM, a new AI-powered weather prediction model that combines machine learning with traditional physics.

The Closer Look:

NeuralGCM is more accurate and efficient than traditional models, capable of running on a laptop instead of supercomputers.

The Importance:

This model could lead to more accurate and accessible weather predictions, especially for extreme weather events.

XAI

The Quick Look:

Elon Musk’s AI startup, xAI, is preparing to launch Grok 3.0, which is expected to be the most powerful AI by December 2024.

The Closer Look:

  • Grok 3.0 will be trained on an unprecedented scale, requiring a massive number of Nvidia H100 GPUs.

  • The Memphis Supercluster, powered by these GPUs, is touted as the most powerful AI training cluster in the world.

The Importance:

Grok 3.0 represents a significant advancement in AI capabilities and could solidify xAI’s position as a leader in the AI space.

AI TUTORIAL

HOW-TO GUIDE

🎥 How to Manage Your Data on X

X/Twitter has rolled out a new feature that uses your posts to train its Grok AI.

Here’s how you can manage your data:.

Step 1: Accessing the Settings

  • Navigate to the Settings page on X’s desktop version.

  • Click on Privacy and safety.

Step 2: Adjusting Grok’s Data Access

  • Select Grok from the menu.

  • Uncheck the box to prevent your data from being used for AI training.

Step 3: Clearing Your History

  • If you’ve interacted with Grok, you can delete this history by clicking Delete conversation history.

The Importance:

Your data is valuable, and how it’s used can affect your privacy. By adjusting these settings, you take an important step in controlling your digital footprint.

PROMPT CORNER

The Job Interview Coach

Practice for job interviews by having ChatGPT simulate an interview for any role, asking one question at a time and waiting for your response.

Copy and Paste this Prompt into ChatGPT:

Act as an interviewer for [insert job role]. Ask me a question and wait for my reply before asking the next question. Do not ask them all at once.

What You’ll Get:

Best times to use this prompt:

College Students:

  • Prepare for internships or first jobs.

Nervousness Management:

  • Reducing interview anxiety by familiarizing yourself with the question-and-answer format.

Post-Interview Analysis:

  • Reviewing and improving your performance after real interviews based on typical questions.

Skills Assessment: 

  • Identify areas for improvement in your responses to ensure you present your best self.

QUICK UPDATES

  • Mistral Large 2: Mistral AI’s latest model, Mistral Large 2 (ML2), allegedly competes with large models from industry leaders like OpenAI, Meta, and Anthropic, despite being a fraction of their sizes.

  • Google’s Gemini 1.5 Flash: Google has rolled out the Gemini 1.5 Flash upgrade, enhancing response times and expanding the context window by fourfold.

  • Substack introducing a new way to publish: Substack is rolling out the ability for writers to draft and publish new posts directly from their phone via its iOS app.

  • Adobe is rolling out more AI generative tools: Adobe’s new features let creative workers describe what they want with brief prompts and receive AI-generated textures or images that could otherwise take hours to create.

  • Reddit’s new policy: Reddit has implemented a policy that restricts major search engines and AI bots from accessing its recent content without payment, aiming to control data scraping and monetize its content.

  • Udio AI music model: Udio unveils version 1.5 of its AI music model, boasting improvements in sound quality, control over musical keys, plus new features for downloading stems and remixing audio.

THAT’S A WRAP!

If you have suggestions or a gem to share, we’re all about it. Just reply to this email.