In partnership with

Good morning. It’s Monday, August 5th.

Did you know: Today is Neil Armstrong’s birthday?

In today’s email:

• OpenAI’s Secret AI Text Detector

• Figure AI Robot Preview

• New Midjourney Rival?

• 5 New AI Tools

• Latest AI Research Papers

You read. We listen. Let us know what you think by replying to this email.

_{In partnership with}

The fastest way to build AI apps

Writer is the full-stack generative AI platform for enterprises. Build and deploy AI apps quickly with Writer AI Studio, a suite of developer tools fully integrated with our LLMs, graph-based RAG, AI guardrails, and more.

Use Writer Framework to build Python AI apps with drag-and-drop UI creation, our API and SDKs to integrate AI into your existing codebase, or intuitive no-code tools for business users.

Start building with AI Studio

OpenAI Says It’s Taking a ‘Deliberate Approach’ to Releasing Tools That Can Detect Writing From ChatGPT

OpenAI is cautiously considering the release of a new tool designed to detect text generated by ChatGPT, as reported by The Wall Street Journal. This tool employs text watermarking techniques, subtly altering the way ChatGPT selects words to create a detectable watermark.

OpenAI acknowledges the method's technical promise but cites concerns about its complexity and potential unintended consequences, such as its effectiveness against tampering and its impact on non-English speakers. The company is also evaluating alternative solutions.

Following the Journal's report, OpenAI updated its May blog post to reflect advancements and limitations in text watermarking, highlighting its accuracy in detecting localized tampering but noting challenges with globalized manipulations. Read more.

OpenAI-backed startup Figure teases new humanoid robot

Figure, an OpenAI-backed startup, is generating buzz with its new humanoid robot, Figure 02. Set to be introduced on August 6, 2024, the robot's trailer highlights its advanced design, including robotic joints and flexible mesh elements. Notably, Figure 02 boasts torque ratings of up to 150Nm and a range of motion of 195 degrees.

Founded in 2022 by Brett Adcock, Figure AI has already made strides with its initial model, Figure 01, aimed at alleviating labor shortages across various sectors. The company has received significant backing from major tech investors like Nvidia, Microsoft, and Intel Capital. Figure's robots are designed to enhance workplace safety and efficiency, focusing on roles in manufacturing and logistics. The new model is expected to integrate OpenAI’s latest AI technologies, further intensifying competition in the humanoid robotics field. Read more.

Former Stable Diffusion developers launch Flux, a new image model challenging Midjourney

Former Stability.ai developers have launched Black Forest Labs, a startup focused on developing advanced generative AI models. The company has introduced its first product, the Flux.1 suite, which sets new benchmarks in text-to-image generation.

The suite includes three versions: Flux.1 [pro] for commercial use, Flux.1 [dev] for non-commercial applications, and Flux.1 [schnell], a fast model available under an Apache 2.0 license for personal use. Black Forest Labs asserts that Flux.1 models exceed competitors such as Midjourney v6.0 and DALL-E 3 (HD) in terms of visual quality, prompt adherence, and flexibility.

Based in Freiburg, Germany, the startup secured $31 million in seed funding led by Andreessen Horowitz, with participation from other prominent investors. The company plans to expand into text-to-video models and is recruiting AI and backend engineers to grow its team. Read more.

Etcetera: Stories you may have missed

Nvidia faces two DOJ antitrust probes over market dominance

Nvidia uses Apple Vision Pro to record humans controlling robots for more realistic training data

Adobe updates free AI education tool for K-12 classrooms

Neuralink implanted second trial patient with brain chip, Musk says

Nvidia reportedly delays its next AI chip due to a design flaw  

Wordware is an IDE that enables building AI agents with features like API deployment, conditional logic, custom code execution, and more.

Felo is a multilingual search engine using AI to deliver authoritative answers from global sources, transcending traditional language barriers.

Moonglow enables running local Jupyter notebooks on cloud GPUs, allowing machine learning researchers to scale experiments without complex setup processes.

Fleso offers HIPAA-compliant AI-driven workflow automation for healthcare, simplifying administrative tasks with a no-code chat interface and custom integrations.

Minduck replaces chat interfaces with structured visual tools, turning thoughts into creative outputs through AI-driven mind-mapping, eliminating technical barriers.

📄 SF3D: Stable Fast 3D Mesh Reconstruction with UV-unwrapping and Illumination Disentanglement

📄 Medical SAM 2: Segment medical images as video via Segment Anything Model 2 ​​

📄 MM-Vet v2: A Challenging Benchmark to Evaluate Large Multimodal Models for Integrated Capabilities

📄 POA: Pre-training Once for Models of All Sizes

📄 SAM 2: Segment Anything in Images and Videos

Thank you for reading today’s edition.

Your feedback is valuable. Respond to this email and tell us how you think we could add more value to this newsletter.

Interested in reaching smart readers like you? To become an AI Breakfast sponsor, reply to this email!