Hello, AI Explorers!

This week OpenAI released their latest newest voice features on ChatGPT - its goal is to create an environment where you can have “conversations" with the model, similar to Siri and Alexa.

Could OpenAI create a new standard for conversational AI? Let’s find out…

This week’s edition:

• OpenAI enhances ChatGPT’s vocal capabilities

• Google’s Gemma 2 2B outperforms GPT-3.5

• Your 5 min AI learning - turn your prompts into music using Loudly

• 5 new AI tools

• More AI news highlights

• AI Image of the week

Read time: 5 minutes

LATEST NEWS

CHATGPT’S VOICE MODE

OpenAI enhances ChatGPT’s vocal capabilities

What you need to know: OpenAI's introduction of the voice feature in ChatGPT which improves user experience by facilitating more natural and expressive interactions, similar to those with human assistants like Siri or Alexa.

The finer details:

• There are 4 voice options to choose from with their own distinct tone and personality: Breeze, Cove, Juniper, and Ember.

• Currently only a small number of ChatGPT plus users can use this feature, but accessibility will be available for all plus users by fall 2024.

• Video and Screensharing capabilities, which was showcased in OpenAI’s demo will release at a “later date”

• OpenAI plans to share the findings and reports of user testing in early August

So what? OpenAI's voice features for ChatGPT mark a significant advancement in AI interaction, making the technology more user-friendly and versatile. This development sets a new standard for conversational AI, reshaping how users engage with digital assistants in their daily lives.

TINY AI MODEL BEATS GPT 3.5

Google’s Gemma 2 2B outperforms GPT-3.5

What you need to know: Google has unveiled Gemma 2 2B, a compact AI model that outperforms much larger models like OpenAI's GPT-3.5 and Mistal’s Mixtral 8x7B. This development challenges the assumption that larger models are inherently superior and marks a significant milestone in AI efficiency and performance.

The finer details:

• Superior Performance: On the LMSYS Chatbot Arena evaluation, Gemma 2 2B achieved a score of 1130. This performance level is on par with that of GPT-3.5-Turbo-0613, which scored 1117, as well as Mixtral-8x7b, which scored 1114 on the same benchmark - models which are 10 times larger.

• Accessibility and Efficiency: Despite its compact size, Gemma 2 2B is designed to run efficiently on laptops and smartphones, making advanced AI capabilities more accessible to a broader range of users.

• Context Length: Gemma 2 2B has a context token set of 8192, which allows it to handle more extensive inputs and maintain coherence over longer text interactions.

• Open-Source Availability: The Gemma 2 models are open-source, allowing developers and researchers to explore and build upon Google's advancements in AI technology

So what? Google's Gemma 2 2B not only outperforms GPT-3.5 but also redefines the expectations for AI model size and performance. This breakthrough could significantly impact the AI industry by promoting the development of more efficient and accessible AI models.

YOUR 5 MIN AI LEARNING

Turn your prompts into music using Loudly

What you need to know: Loudly allows you to create your own music through either picking and selecting your options, or using their innovative text-to-music feature.

Step-by-step:

1. Visit Loudly and sign up for a free account.

2. Click the text to music option to generate your desired music.

   

3. In the "Prompt" section, describe your perfect tune that you want to create.

4. Click generate to watch your tunes come to life

5. As an extra tip. Once the music has been made, you can fine tune it by asking for more specific changes in the prompt, any previously generated music will still be available to listen to.

LATEST AI TOOLS

• 🍽️ Food Mood - Combine two cuisines in to one recipe using AI.

• 🎞️ Synthesia - Create AI generated videos using digital avatars.

• 🤝 AI Notebook - Create meeting notes automatically.

• 💹 Blaze - Automate marketing with lead generation optimisation.

• 📺 Banani - A Figma plugin that edits UI designs using prompts.

AI HIGHLIGHTS

OpenAI pledges to give the U.S. AI Safety Institute early access to its next model, GPT-5.

NVIDIA and Meta CEOs state that every business will ‘have an AI’

Apple opts for Google chips in AI infrastructure, instead of tech leader Nvidia

AI development may cause tech giants to revisit their green initiative.

AI IMAGE OF THE WEEK

Copilot Designer Prompt: A cosy woodland cabin nestled in a snowy forest, with a warm fireplace glowing inside and a pair of snow-covered pine trees framing the scene. The cabin has a rustic, handcrafted appearance with a stone foundation and wooden beams, rendered in a distinctive, painterly style that blends realism with abstract elements. In the foreground, there is a frozen lake with a small dock and a lone ice fisher, the water's surface reflected in a dreamlike, impressionistic manner.

Send us your AI generated art for a chance to be featured.

FEEDBACK

• Let us know how we did - shoot us a quick email. We read every single one (no AI in sight!).

• If you enjoyed this newsletter please share with an awesome friend or two.

• If you are that awesome friend then hit the below button: