GenAI in 2025: Where Models, Media, and Machines Converge.
Date
April 9th, 2025
Reading Time
10 mins
Tags cloud
What's news
1. Amazon says its AI video model can now generate minutes-long clips.
April 7, 2025
Source: Amazon
Amazon has upgraded its AI video model, Nova Reel, to version 1.1, enabling it to generate videos up to two minutes long with a consistent style across multiple shots. Announced in December 2024, Nova Reel is Amazon’s entry into the competitive generative video space.
The update introduces a “Multishot Manual” mode, which lets users upload an image and a prompt (up to 512 characters) to guide video composition, generating up to 20 shots from a 1280x720 image. Standard prompts can be up to 4,000 characters.
Nova Reel is accessible only through AWS platforms like Bedrock and requires a request for access. However, Amazon hasn’t disclosed how it trained the model or offered an opt-out for creators whose content might have been used. While this raises ethical and legal concerns, Amazon says it will indemnify AWS customers against copyright claims stemming from content generated using Nova Reel.
2. Meta exec denies the company artificially boosted Llama 4’s benchmark scores.
April 7, 2025
Source: TechCrunch
Meta’s VP of generative AI, Ahmad Al-Dahle, denied rumors that the company trained its new Llama 4 Maverick and Scout models on benchmark test sets to artificially boost performance. These rumors, which began circulating on X and Reddit, stemmed from an anonymous post on Chinese social media, allegedly by a former Meta employee.
Concerns were heightened by reports of inconsistent performance and Meta’s use of an experimental Maverick version to score higher on the LM Arena benchmark. Al-Dahle admitted users are seeing “mixed quality” depending on the hosting cloud provider but emphasized that the models were released early and are still being fine-tuned.
Meta maintains that benchmark manipulation claims are false, and that ongoing updates and bug fixes are in progress.
3. Google’s AI Mode now lets users ask complex questions about images.
April 7, 2025
Source: Perplexity
Google is expanding AI Mode, its experimental Google Search feature, by adding multimodal search capabilities. Now, users can ask questions about images they upload or take with their camera. This feature is powered by Google Lens and enables detailed understanding of an entire scene, including object relationships, materials, and layout.
Using a technique called “query fan-out,” AI Mode can generate multiple related queries based on the image and user prompt, providing richer responses. For instance, snapping a photo of your bookshelf and asking for book recommendations based on it will return personalized suggestions. Users can also ask follow-up questions to refine results, like filtering by book length.
Previously limited to Google One AI Premium subscribers, AI Mode is now rolling out to millions of Labs users, marking Google’s push to compete with services like ChatGPT Search and Perplexity.
4. Krea raises $83M to be the one-stop shop for GenAI creatives.
April 7, 2025
Source: TechCrunch
Krea, a San Francisco-based startup, has raised $83 million (latest Series B: $47M) to simplify how designers and creatives use generative AI. Valued at $500 million, Krea offers a unified platform that integrates multiple AI models through an intuitive, customizable interface, making it easier for users to generate and edit images and videos without needing technical expertise.
Founded by Victor Perez and Diego Rodriguez—former creatives and AI researchers—Krea was built to help users focus on creativity, not prompts. It automatically selects the best AI models for a user’s request and allows for direct, painter-like editing of generated images. Its clients include creators at Pixar, Samsung, LEGO, and Perplexity AI.
With backing from Bain Capital Ventures, Andreessen Horowitz, and Abstract Ventures, Krea plans to expand into audio and music generation, as well as enterprise-level features. The startup aims to empower, not replace, creative workflows—helping users stay creative while leveraging the latest AI advancements.
5. OpenAI reportedly mulls buying Jony Ive and Sam Altman’s AI hardware startup.
April 7, 2025
Source: The Verge
OpenAI is reportedly in talks to acquire or partner with io Products, a secretive AI hardware startup co-founded by former Apple design chief Jony Ive and OpenAI CEO Sam Altman. The potential acquisition could be worth around $500 million, according to The Information.
io Products has backing from investors including Laurene Powell Jobs’ Emerson Collective, and its team includes ex-Apple designers Tang Tan and Evans Hankey. The company is exploring AI-enabled hardware concepts, such as smart home devices, with a focus on creating tech that is less socially disruptive than smartphones.
While acquisition is on the table, a partnership with OpenAI is also being considered.
Tags cloud
Newsletter
DISCOVER MORE
![[Recap] UPP Global Technology JSC Establishing Anniversary](/homepage/news-section/new-4.webp)
LET’S TALK...
Content delivered to your inbox
ENTER YOUR EMAIL
YOU WANT TO...
Hanoi, Vietnam
Web3 Tower, No. 15, Alley 4, Duy Tan, Cau Giay, Hanoi, Vietnam