This week’s discussion dives into Apple's latest AI features for iOS and macOS, promising intriguing updates for Siri users. Luma’s Dream Machine faces a traffic surge after its launch, raising questions about AI video creation. Meanwhile, advancements from Runway's Gen-3 Alpha and Leonardo's new video mode are game changers. The potential shift of OpenAI to a regular profit model and Huawei's competition with NVIDIA add layers to the evolving AI landscape. Plus, explore emerging language models and the drive to reduce hallucinations in AI.
Apple introduces new AI features for iPhone and Mac.
Luma's Dream Machine overwhelmed by traffic upon release.
OpenAI may transition to a for-profit model under Sam Altman.
Adept AI in talks with Microsoft for a potential deal.
Deep dives
Stable AI releases Stable Diffusion 3 Medium under new licenses
Stable AI has announced the open release of Stable Diffusion 3 Medium in collaboration with NVIDIA, optimized for TensorRT. The model is available under Stability Non-commercial Research Community License and a Creator License for commercial use.
Meta introduces new AI models for audio, text, and watermarking
Meta has unveiled a new AI model called JASCO for music generation, Audio Seal for watermarking voice, and the Multi-Model Chameleon for vision and text tasks. These models, ranging from 7 billion to 34 billion parameters, will be released under different licenses for research purposes.
Astrocade announces open release of Stable Diffusion 3 Medium
Astrocade has released Stable Diffusion 3 Medium, a text-to-image model, under non-commercial and commercial licenses. The model offers high-resolution rendering and prompt faithfulness, continuing the legacy of Stable Diffusion in the text-to-image generation space.
AI-Generated Content on Website and Attribution
Buzzy AI search engine Perplexity is using content from news outlets for its Proplexity Pages feature, replicating segments from sources like Forbes and The New York Times. Although there is some attribution with links to the original sources, the practice raises concerns about the originality and reliance on external content.
Huawei's Concern Over Lack of Advanced Chips
A Huawei executive voices worry over China's inability to access 3.5 nanometer chips and advanced chipmaking technologies due to US sanctions. The challenge lies in obtaining next-generation chip manufacturing equipment, such as extreme ultraviolet lithography, essential for future semiconductor advancements.
Adobe's Terms of Service Change for Training on User Data
Adobe revises its terms of service after facing backlash for allowing training on customers' data uploaded to their cloud-based tools. The updated terms aim to address concerns raised by the creative community by specifying how user data may be used for AI training purposes.
AI Chop Shops and AI-Generated News Content
An investigation reveals AI chop shops creating AI-generated news websites that replicate and republish content from reputable sources like Forbes and The New York Times with minimal attribution. This trend raises ethical questions regarding the source and validity of AI-generated news content.
Meta's Paused AI Models in Europe
Meta pauses AI models in Europe following advocacy group backlash and Irish Data Protection Commission scrutiny. This action reflects the ongoing challenges tech companies face regarding AI regulations and data privacy compliance in different regions.
