Gemini Handles Nearly One Million Tokens

Today’s Menu 🥠

• Google Debuts Gemini 1.5

• OpenAI’s Valuation Keeps Going Up

• 5 Trending AI-Powered Tools

• A Possum With A Crazy Weekend

Read Time: 2.5 minutes

Fast Snacks 🥡

Atlanta Tech Unicorn Buys Conversational AI
Salesloft acquires AI company Drift to create the first AI-powered revenue orchestration platform that serves the entire buyer journey…

Masayoshi Son Seeks To Build $100B Chip Venture
Softbank founder is seeking a $100B to bankroll a chip venture to compete with NVIDIA, and supply semiconductors essential for AI…

Extract Vital Insights From Audio (partner)
Deepgram has expanded their Audio Intelligence lineup to include high-level natural language models that analyze sentiment, recognize intent, and detect topics…

A New Update
Google Debuts Gemini 1.5

Google just debuted Gemini 1.5,  an update to Gemini that aims to give you more control with significantly larger processing capabilities.

Imagine your brain on a good day, after downing a triple espresso and acing a trivia night. That's basically Gemini 1.5.

This new AI model can handle up to a million tokens of information at once compared to the previous constraints of 32,000, which, as the scientists put it… is a sh*t load. 

Now, what exactly are these tokens? Think of them like Lego bricks for information. A single word can be a token, but so can an image, a video snippet, or even a line of code. So, a million tokens is like having a million Lego bricks to build with, allowing Gemini 1.5 to understand, reason, and process massive amounts of information at once.

But how did they pull this off?

The secret lies in a clever architecture called "Mixture-of-Experts" (MOE). Instead of relying on a single, overloaded system, MOE distributes tasks among a team of specialized experts, similar to how different professionals collaborate in their respective fields.

This not only makes Gemini 1.5 faster and more efficient but also empowers it to learn complex things more effectively.

What does this mean for you?

While still in testing, Gemini 1.5 has the potential to change how we interact with AI in the future. Here are some potential benefits:

• Smarter assistants: Imagine an AI assistant that can understand your entire request, even if it's complex or involves multiple steps.

• Enhanced search: Search engines powered by Gemini 1.5 could provide more relevant and nuanced results, taking into account the broader context of your query.

• Personalized experiences: AI-powered applications could tailor themselves to your individual needs and preferences based on a deeper understanding of your past interactions.

Right now, only approved developers and businesses have access to the full 1 million-token version, but expect that to change in the future.

Together With Scribe:
Automatically Create Step-By-Step Guides

Scribe just bagged $25M in Series B funding to give you a break from answering people’s questions all day. Instead, just create step-by-step guides (automatically, thanks to AI) to share with your team.

• Capture any process using the Chrome extension;

• Easily customize steps and redact sensitive info;

• Share with colleagues – and get back to your work

Test Your Skill
Snack Quiz: Choose The Real Image 🔍

Higher & Higher
OpenAI Valued At $80M

OpenAI just completed a deal that valued the business at $80 billion [or more] citing three unnamed people with knowledge of the deal.

The San Francisco-based company would sell shares in a tender offer — a "public bid for stockholders to sell their stock" — led by venture capital firm Thrive Capital.

The deal allows OpenAI employees to cash out their shares in the company.

The company was involved in another tender offer deal just last year, with venture capital firms Thrive Capital, Sequoia Capital, Andreessen Horowitz, and K2 Global agreeing to buy shares in a deal that valued OpenAI at about $29 billion — meaning it has roughly tripled its valuation.

OpenAI also recently unveiled its latest AI tool, "Sora," allowing users to create videos based on text.

"Sora can generate videos up to a minute long while maintaining visual quality and adherence to the user's prompt," the company says on its website.

Previews of the product, which is now set to be tested by red teamers "to assess critical areas for harms or risks," show highly realistic footage of an art gallery tour and a woman walking along a rainy Tokyo street.

More Tools For The Toolbox
5 Trending AI-Powered Tools

Snack Prompt: Offers a unique and multifaceted approach to prompt development and curation.

Just Prompts: This website focuses on optimizing user-generated prompts for AI models.

FlowGPT: This website provides a platform for sharing and discussing prompts for GPT models.

Deepgram: Extract vital insights from conversational audio at scale with Deepgram's AI-powered audio intelligence.

Magic: An optimization tool that upgrades your AI workflow with integrations with Bard, Claude, Midjourney, & more.

Straight Mogging
What A Crazy Weekend

Snack Quiz Results
Correct Answer ✔️