The Big Story
New York City May Require Landlords to Disclose AI Use in Listings
In a move that could have far-reaching implications for the real estate industry, New York City Mayor Mamdani has announced plans to require landlords and realtors to disclose the use of AI in property listings.
The decision comes as AI-powered image generation tools have become increasingly popular among real estate professionals, allowing them to generate high-quality images of properties without the need for human photographers. While this technology has many benefits, such as reducing costs and increasing efficiency, it also raises important questions about transparency and fairness in the market.
Under the proposed rules, landlords and realtors would be required to clearly indicate whether AI-generated images are being used in property listings, allowing potential buyers to make informed decisions about the properties they are considering. The move is seen as a major step forward in promoting transparency and accountability in the industry.
The decision is also expected to have significant implications for the broader technology industry, as it highlights the need for greater regulation and oversight of AI-powered tools. As AI becomes increasingly integrated into various aspects of our lives, it is essential that we establish clear guidelines and standards for its use, particularly in areas where it has the potential to impact people's daily lives.
What Shipped
NVIDIA released DeepStream 9.1, introducing 13 agentic skills that enable coding agents like Claude Code and Codex to build multi-camera video analytics pipelines from natural-language prompts. This significant update brings agentic AI capabilities to vision AI, opening up new possibilities for developers to create more advanced video analytics applications.
Perplexity AI released WANDR, an open benchmark evaluation harness with 500 evidence-heavy tasks. WANDR tests whether research agents can discover many qualifying entities and back each one with cited, re-vealed references. This milestone marks a significant step forward in the development of more intelligent and accountable AI models.
OpenAI reduced Codex Model Context Size from 372k to 272k, a crucial update that improves the efficiency and scalability of this popular AI model. With this change, developers can now fine-tune Codex using smaller context sizes, making it easier to integrate into various applications.
Fine-Tuning Qwen3 with LoRA Using NVIDIA NeMo AutoModel: A Complete Single-GPU Google Colab Workflow Tutorial was published, providing a comprehensive guide on how to build an end-to-end workflow in Google Colab using a single GPU. This tutorial showcases the power of NVIDIA's NeMo AutoModel and its ability to streamline the fine-tuning process for Qwen3 with LoRA.
Kimi K3 vs DeepSeek V4 Pro vs GLM-5.2: Open Trillion-Scale MoE Models Compared on Benchmarks, License, and Serving Cost was released, highlighting the performance differences between three open trillion-scale MoE models. This comparison provides valuable insights for developers looking to choose the right model for their specific use cases.
From the Labs
NVIDIA released DeepStream 9.1, introducing 13 agentic skills that enable coding agents like Claude Code and Codex to build multi-camera video analytics pipelines from natural-language prompts. This significant update brings agentic AI capabilities to vision AI, opening up new possibilities for developers to create more advanced video analytics applications.
Perplexity AI released WANDR, an open benchmark evaluation harness with 500 evidence-heavy tasks. WANDR tests whether research agents can discover many qualifying entities and back each one with cited, re-vealed references. This milestone marks a significant step forward in the development of more intelligent and accountable AI models.
Other Notable News
AI Mania Is Eviscerating Global Decision-Making: According to an article on Ludic Mataroa, AI is increasingly playing a significant role in global decision-making processes, raising concerns about accountability and transparency.
Could Your AI Systems Already Be High-Risk Under the EU AI Act?: The latest guidance from the European Union's AI Act has sparked questions about the risk level of various AI systems. An on-demand webinar aims to provide clarity for organizations looking to understand their AI governance program.
Fine-Tuning Qwen3 with LoRA Using NVIDIA NeMo AutoModel: A Complete Single-GPU Google Colab Workflow Tutorial was published, providing a comprehensive guide on how to build an end-to-end workflow in Google Colab using a single GPU. This tutorial showcases the power of NVIDIA's NeMo AutoModel and its ability to streamline the fine-tuning process for Qwen3 with LoRA.
Perforce charges $500 for training training videos.. and it's AI narrated: The popular training platform Perforce has announced that they are now offering AI-narrated training videos, with a price tag of $500 per video. This innovative approach aims to enhance the learning experience.
The Take
In this week's news, AI has taken center stage in various domains, from real estate to decision-making and even video analytics. The most striking development is perhaps the New York City Mayor's proposal to require landlords and realtors to disclose the use of AI in listings. This move could have significant implications for the property market, as it highlights the increasing importance of transparency in AI-driven applications.
Meanwhile, Perforce has sparked controversy by charging $500 for training videos that are narrated by AI. While this may seem like a minor issue at first glance, it raises questions about the role of AI in education and whether we're heading towards a future where humans are no longer necessary as instructors.
In other news, NVIDIA has released DeepStream 9.1, which introduces 13 agentic skills that enable agents to build video analytics pipelines from natural-language prompts. This development could have far-reaching implications for industries such as healthcare and finance, where AI-driven insights can be used to improve decision-making processes.
It's also worth noting the release of WANDR, an open benchmark and evaluation harness designed to test research agents' ability to search wide and deep. As AI continues to evolve and become more integrated into our daily lives, it's crucial that we develop standardized tools for evaluating its performance and potential biases.
Finally, OpenAI has reduced the context size of its Codex model from 372k to 272k, which could have significant implications for the development of AI-powered language models. This move highlights the ongoing efforts to improve the efficiency and effectiveness of AI systems, as well as the need for continued research in this area.
Read more about the latest developments in AI and machine learning.