🗞 Alexa+ Prime Upgrade, GPT-4.5 "Orion" Coming, and Nvidia's Record Growth

🔎 The Latest on the AI Frontier:

• Amazon Introduces Alexa+ with Advanced AI Capabilities for Prime Members

• OpenAI Likely Preparing GPT-4.5 "Orion" Model for Initial Pro Subscriber Release

• Nvidia Reports Record Revenue as CEO Dismisses DeepSeek Competition Concerns

• Perplexity Updates iOS App with Redesigned Voice Interface and Claude 3.7 Integration

• Meta Fixes Technical Error That Exposed Instagram Users to Graphic Content

• Baidu Announces Strategic Shift to Open-Source with Upcoming Ernie 4.5 AI Model

• Other news you might find interesting

🔊 Amazon launches Alexa+, a generative AI-powered upgrade to its voice assistant with enhanced capabilities

• The new system offers deeper personalization, improved conversational abilities, and proactive assistance that can remember preferences and orchestrate complex multi-step tasks across different services.

• Alexa+ integrates with platforms like OpenTable, Ticketmaster, and Uber Eats, while adding "agentic" features that allow it to navigate the web and complete actions independently.

• Available free for Amazon Prime members (otherwise $19.99/month), initially rolling out on Echo Show 8, 10, 15, and 21 devices in the U.S.

👀 OpenAI appears to be preparing GPT-4.5 (codenamed Orion) for initial release to Pro subscribers

• Android users spotted an unintentionally released tooltip stating "Pro users now have access to our newest, largest model," suggesting a tiered rollout strategy.

• The timeline remains uncertain, with rumors suggesting the release could happen within days or potentially take more than two weeks.

• This follows Sam Altman's recent communication about OpenAI's model release roadmap, which indicated GPT-4.5 would be available "soon."

📈 Nvidia CEO Jensen Huang dismisses DeepSeek concerns as company reports record $39.3 billion revenue

• Huang called DeepSeek R1 an "excellent innovation" that benefits Nvidia as reasoning models require 100x more compute power.

• Nvidia's data center sales nearly doubled to $115 billion in 2024, rising 16% from previous quarter.

• Company forecasts $43 billion revenue next quarter, citing "extraordinary" demand for its Blackwell chip designed for reasoning models.

✨ Perplexity rolls out iOS app update featuring redesigned voice mode and Claude Sonnet 3.7 integration

• The revamped voice interface includes an interactive animated sphere design with six different voice options and direct access to search functionality.

• New personalization options allow users to set stock tickers and sports teams to appear on their home screen widget.

• Claude Sonnet 3.7, Anthropic's latest model, is now available in Perplexity's model selector for both mobile and web users.

🔍 Meta resolves error that exposed Instagram users worldwide to violent and graphic videos in their Reels feeds

• The technical glitch bypassed content filtering systems, including the "sensitive content control" setting, prompting a wave of complaints across social media platforms.

• This incident comes shortly after Meta's controversial decision last month to discontinue its U.S. fact-checking program across Facebook, Instagram, and Threads—platforms with over 3 billion global users.

• The company has faced ongoing criticism for its content moderation practices, including previous incidents involving the Myanmar genocide content, eating disorder promotion to teens, and pandemic misinformation.

🔄 Baidu announces significant strategy shift with Ernie 4.5 AI model launch in mid-March, embracing open-source approach

• The upgraded model will feature enhanced reasoning capabilities and improved multimodal processing for text, images, video, and audio, with full open-source release planned for June 30.

• CEO Robin Li's pivot from closed-source development comes in response to growing competition from Chinese AI startup DeepSeek, which offers affordable models comparable to Western competitors.

• This strategic realignment follows Alibaba's recent open-source release of its Wan 2.1 video and image-generating AI model, signaling a potential industry-wide shift toward greater transparency in China's AI landscape.

🖨 More news you might find interesting:

• Elon Musk reignites feud with OpenAI CEO by reviving "Scam Altman" nickname in response to 2023 testimony video.

• Qualcomm and Nokia Bell Labs demonstrate interoperable multi-vendor AI models for wireless networks.

• Hugging Face introduces FastRTC, an open-source Python library that simplifies real-time AI voice and video application development.

• ElevenLabs introduces Scribe, a high-performance AI transcription model supporting 99 languages.

• Hume AI debuts Octave, a text-to-speech model with advanced emotional intelligence capabilities.

• 25 practical tips for using ChatGPT to boost programming productivity.

• IT roles are rapidly evolving beyond traditional hardware and software support to become integrated across all business functions.

Have any feedback? Send us an email