Welcome back!

The competition in AI image generation is heating up with the recent launch of Flux, now also integrated into Grok-2. In response, Google Imagen and Midjourney are pushing forward with their latest innovations. Imagen 3 is raising the bar for image quality and detail. Let’s explore...

Today’s Summary:

• 🎨 Google launches Imagen 3 model

• 🛠️ Midjourney new Unified Web Editor

• 🎮 Startup “Exists” unveils AI 3D Game Creator

• 🤖 Apple working on tabletop robot

• 🖼️ Elon Musk’s xAI to develop own image model

• ⚙️ Nous Research releases Hermes 3

• 🌍 Google expands AI answers to new countries

• 🛠️ 2 new tools

TOP STORY

Google releases Imagen 3 text-to-image model

The Summary: Google has released Imagen 3, its most advanced text-to-image AI model. Google claims Imagen 3 sets new benchmarks for image quality and detail, surpassing DALL-E 3 and Midjourney in internal tests.

Key details:

• Imagen 3 now available to all US users via the ImageFX platform, following its announcement in June

• Outperforms competitors like DALL-E 3 and Midjourney v6 on the GenAI benchmark

• Users praised improved textures and text generation capabilities, but criticised its heavy-handed censorship

Why it matters: Imagen 3 pushes the boundaries of AI image generation with top-tier quality and is now freely accessible in the US. However, some users have expressed frustration over its unavailability in other countries and its creative limitations compared to open-source models like Flux.1.

Try it on ImageFX

TOOLS

Midjourney releases Unified Web Editor for AI image creation

The Summary: Image generator Midjourney has launched a new web-based editor that integrates various AI image editing features into a single interface. The tool includes inpainting, outpainting, and a precise brush tool for detailed edits, making image editing more accessible and user-friendly.

Key details:

• New editor unifies inpainting, outpainting, and canvas extension

• Added precise "brush" tool for inpainting

• Available at midjourney.com/imagine for users with 10+ created images

Why it matters: By simplifying the web-based editing process, Midjourney –previously closely tied to a complex Discord interface– is making its tools more user-friendly. These new features could attract more users to the platform, as competitors continue to push their own AI image tools, further accelerating innovation in this space.

GAMING

Startup “Exists” Launches AI 3D Game Creator

The Summary: Startup “Exists” unveiled a GenAI platform that turns text prompts into 3D games. The cloud-based system uses AI models to generate game environments, characters, and mechanics.

Key details:

• Creates high-quality 3D games from text prompts

• Fully automated, cloud-based process without coding

• Neural network integrates GenAI with gaming engine

• Include AI-generated assets and multiplayer

• Closed Beta phase underway; early access sign-ups available

Why it matters: Applying AI to game creation could enable anyone to turn their ideas to life without technical barriers, while also opening new possibilities for user-generated content within established games. Though still an early shot, this offers a glimpse of how gaming could evolve in the coming years.

Signup for beta

QUICK NEWS

Quick news

• Apple reportedly working on a tabletop robot

• Elon Musk’s xAI to develop its own image generator model

• Nous Research releases Hermes 3 models

• Google expands AI answers to new countries

TOOLS

🥇 New tools

• Tome - AI powered lawyer

• Shellmate - Mac terminal productivity tool with superpowers