AI-Powered news roundup: Edition 30
Our bi-weekly news roundup is designed to keep you informed on the latest developments in AI, all in under 5 minutes.
In this edition:
China's DeepSeek challenges top AI models with significant update
Odyssey AI introduces a new 'interactive video' medium
Viral 'emotional support kangaroo' video highlights the power of AI fakes
Claude's new voice assistant takes on the competition
Perplexity Labs automates complex research and project creation
The debate over AI and job displacement intensifies
Duolingo's AI-first strategy leads to social media blackout
1. China's DeepSeek challenges top AI models with significant update
Source: Reuters
Chinese AI firm DeepSeek has released a major update to its R1 reasoning model, positioning it as a strong competitor to industry leaders like OpenAI's o3 and Google's Gemini 2.5 Pro. The new version, R1-0528, boasts a 45-50% reduction in "hallucinations" and enhanced capabilities for complex reasoning tasks. This launch follows the company's initial R1 release in January, which sent shockwaves through the tech industry by demonstrating that powerful AI could be developed with greater efficiency and lower costs.
The competitive landscape has shifted since DeepSeek's debut, with Google introducing discounted pricing for its Gemini models and OpenAI releasing the more compact o3 Mini. However, DeepSeek appears to be maintaining its edge. Independent benchmarks show that R1-0528 delivers performance on par with Claude 4 and Gemini 2.5 Pro but at a more competitive price point. The company has also released a smaller, distilled version of the model that can run on a single GPU. This progress is reflected in its rise on the Artificial Analysis Intelligence Index, where it has climbed from 60 to 68.
2. Odyssey AI introduces a new 'interactive video' medium
Source: Odyssey Announcement
Odyssey AI has unveiled what it calls an "entirely new medium of entertainment" with the launch of its interactive video experience. Controlled with standard WASD keys, the company is careful to distinguish it from a game, instead framing it as an early version of a "Holodeck." The technology is a significant step forward, generating new video frames in real-time at 25 frames per second (every 40ms) by utilizing real-time world models. This dynamic generation sets it apart from traditional video models that produce fixed clips and comes at a cost of $1-2 per user-hour.
This innovation places Odyssey AI alongside other major players exploring generative interactive environments, such as Microsoft's Muse AI - Quake, Decart's Minecraft, and Google's Genie 2. What makes Odyssey's approach unique is its learning model, which is trained on real-world video footage rather than confined to game environments. Despite the ambitious claims, the current experience is described as a "glitchy dream," offering a raw and unstable, grainy walking simulator.
3. Viral 'emotional support kangaroo' video highlights the power of AI fakes
Source: Forbes
A video that recently went viral, showing a woman attempting to board an airplane with her "emotional support kangaroo," has been revealed as a complete fabrication, generated by AI. The clip, which amassed over 58 million views on X (formerly Twitter), was created by visual effects artist "Infinite Unreality" and was convincing enough to fool a significant portion of its audience into believing it was genuine.
The incident marks a critical moment in the evolution of AI-generated content, demonstrating that it has crossed the "uncanny valley" and can now seamlessly blend into the "ambient internet noise" without detection. This development is particularly timely, as it coincides with the launch of Google's Veo 3 model, which has achieved photorealistic video generation. The ease with which convincing fake videos can now be created and distributed raises serious concerns about the potential for widespread disinformation.
4. Claude's new voice assistant takes on the competition
Source: Anthropic Announcement
Anthropic has rolled out a voice assistant for its Claude mobile app, making it a more direct competitor to other AI-powered personal assistants. The new feature, which is being made available to all users for free, includes multiple voice options and allows for hands-free interaction. The assistant is designed to handle a variety of tasks, positioning itself as a more capable alternative to existing voice assistants like Siri.
A key strength of the new Claude voice assistant is its integration capabilities. It can connect to a user's Google Drive, Gmail, and Calendar, enabling it to perform practical tasks such as reading urgent emails or summarizing upcoming meetings. While the basic functionality is free, Pro and Max subscribers will have access to the full suite of integration features, allowing for a more comprehensive and personalized assistant experience.
5. Perplexity Labs automates complex research and project creation
Source: Perplexity Announcement
Perplexity has launched Perplexity Labs, a new feature for its Pro subscribers that introduces a powerful AI agent capable of autonomously working on complex projects for over ten minutes. This new tool can handle in-depth research and web Browse to generate complete reports, spreadsheets, dashboards, and even web applications, effectively turning days of work into a matter of minutes.
The capabilities of Perplexity Labs are showcased through a series of impressive demonstrations. These include generating a detailed [suspicious link removed] complete with charts, creating [suspicious link removed] that include links to the LinkedIn profiles of candidates from top AI startups, and developing entire [suspicious link removed] with storyboards and screenplays.
More examples are available here.
6. The debate over AI and job displacement intensifies
Source: Axios
The conversation around AI's impact on the workforce has gained new urgency, with prominent figures like Anthropic CEO Dario Amodei warning of significant job displacement in the near future. Amodei predicts that AI could eliminate as much as half of all entry-level, white-collar jobs within the next one to five years, potentially causing unemployment to spike to between 10% and 20%. This disruption is expected as companies increasingly adopt AI "agents" to automate tasks in fields such as coding, finance, law, and consulting.
This transition from AI as a tool for augmentation to a driver of full automation is already underway. Major companies like Meta, Microsoft, and Walmart have announced job cuts, and many CEOs are privately considering how AI can replace human roles. The tipping point is anticipated to arrive within the next one to two years. OpenAI CEO Sam Altman echoes these concerns, stating that while there will be "scary times ahead," early and open collaboration on the development and regulation of AI can help ensure its benefits outweigh the risks.
7. Duolingo's AI-first strategy leads to social media blackout
Source: Forbes
Language-learning platform Duolingo is facing a significant backlash following its CEO's announcement of an "AI-first" strategy. The plan, which involves gradually replacing contract workers with artificial intelligence and making AI proficiency a factor in employee performance reviews, sparked outrage among users. The public's negative reaction was swift and intense across various social media platforms.
In a surprising turn of events, Duolingo's social media presence went dark in the wake of the controversy. The company's popular TikTok and Instagram accounts were wiped, leaving behind only cryptic messages such as "real eyes realize real lies."
A bizarre video titled "Exposing Duolingo" also surfaced, featuring a masked individual claiming to be an employee who stated the company "was never funny." It remains unclear whether this is a genuine employee revolt or an elaborate and risky marketing stunt.
Are you as passionate about AI as we are? Explore open positions at Siili and work on exciting, cutting-edge projects. Join us and take the next step in your career!
Explore our AI Advisory services here.