Models & Labs
Google Allows Removal of Visible Watermarks in AI Outputs
TechCrunch AIhigh confidence
Why it matters
- →Allows for greater creative flexibility in professional use of AI-generated media.
- →Maintains transparency with invisible watermarks and metadata for content identification.
- →Reflects a broader industry trend towards balancing creative freedom with regulatory compliance.
Google has announced that users will now have the option to remove visible watermarks from AI-generated content, including images, videos, and songs. This feature will be available for the Nano Banana, Omni, and Lyria models, and can be toggled in Gemini and Google's video editor, Flow. Despite this change, invisible SynthID watermarks and C2PA metadata will remain to ensure transparency. The update is part of Google's effort to balance creative control with the need to identify AI-generated content. The feature will be available in the coming days.
Read originalMore from TechCrunch AI
© TechCrunch AIMarket & Regulationother
Lawsuit Filed Against xAI Over Grok's Alleged Misuse
A legal case has been initiated against Elon Musk's xAI, now integrated with SpaceX, over allegations that its chatbot Grok was used to transform a childhood photo into explicit images. Jane Doe 4 claims her stepfather exploited Grok to produce over 7,000 explicit images of her, leading to his suicide following a police raid. The lawsuit, supported by three Tennessee teenagers, accuses xAI of neglecting to implement necessary safeguards to prevent Grok's misuse in generating explicit content involving minors. This situation brings to light significant concerns about the ethical deployment of AI tools and the pressing need for robust protective measures.
TechCrunch AI
© TechCrunch AIMarket & Regulationwriting
Anthropic Details Claude's New Text Watermarking
Anthropic is introducing a watermarking system for its chatbot Claude to comply with the EU AI Act’s Transparency Code. This watermarking, using the SynthID-Text approach, embeds patterns in text that are invisible to readers but detectable with a key. While light editing may not remove the watermark, a complete rewrite could. The watermarking aims to ensure AI-generated content is identifiable without affecting the quality of the output. This move aligns with other major AI developers who are also adopting similar practices.
TechCrunch AI
© TechCrunch AIMarket & Regulationbusiness
SpaceX Completes Acquisition of AI Startup Cursor
SpaceX has finalized its acquisition of AI coding startup Cursor, marking a pivotal development in the tech industry. This acquisition grants Cursor access to SpaceX's extensive computing infrastructure, including what is claimed to be the largest fleet of GPUs globally. Initially announced in April, the deal reflects SpaceX's ambition to enhance its technological capabilities and scale AI intelligence. With this acquisition, Cursor is poised to leverage SpaceX's resources to advance its AI development, potentially transforming the application of AI across various sectors.
TechCrunch AI
More in Models & Labs
Models & Labsmodels
llama.cpp adds MiniMax model support
The latest llama.cpp update expands its functionality by integrating the MiniMax-Text-01 and MiniMaxM1ForCausalLM models, enhancing its role in causal language modeling. This release focuses on refining the MiniMax-Text-01 model by eliminating state transpose operations and implementing a logits mask to manage zero-valued embeddings. These adjustments aim to streamline the token sampling process and boost model efficiency. While no new model architectures are introduced, the update significantly refines existing processes, making llama.cpp more robust and efficient for developers working with these specific models.
llama.cpp Releases
Models & Labsmodels
llama.cpp b10441 Release Updates Load Mode
The latest release of llama.cpp, version b10441, introduces a significant change by replacing deprecated flags with a unified --load-mode argument. This update simplifies the configuration process across scripts, examples, and documentation, making it easier for developers to manage memory mapping and loading options. The release also includes updates to internal warning messages and environment variable documentation, ensuring clarity and consistency. While this update doesn't introduce new features, it streamlines the user experience and reduces potential confusion for developers working with llama.cpp.
llama.cpp Releases
Models & Labsmodels
b10442 Release Enhances Vulkan Support for Intel Xe
The b10442 release of llama.cpp brings notable improvements to Vulkan support, specifically targeting Intel Xe platforms. By adding SHMEM_STRIDE_PAD and APPLY_SLM_A_RESHAPE for cooperative matrix operations, this update aims to optimize performance on Intel hardware. Additionally, it addresses a critical out-of-bounds read issue in kvalues_mxfp4 initialization, enhancing stability. While these changes are technical, they signify a focused effort to refine performance and compatibility for developers working with Intel's Vulkan drivers. This release doesn't introduce new models but strengthens the existing infrastructure for better efficiency.
llama.cpp Releases