What's new in TECH?

Hold onto your hard drives, tech enthusiasts, because this week has been a whirlwind! From developer tools to cutting-edge AI, major players delivered updates that promise to shape the future. Let's buckle up and explore the highlights:

For developers:

• Flutter 3.19: Get ready for lightning-fast development with 50% faster hot reload! This update also unlocks stable iOS 17 support and even lays the groundwork for building stunning macOS apps using Flutter. Whether you're a seasoned mobile developer or just a starter, Flutter 3.19 makes your workflow smoother and your app's reach wider than ever.
• Gemini 1.5: Power up your AI applications with the latest iteration of Gemini's large language model. Brace yourself for 2x faster inference, meaning your AI can think and respond even quicker. The model also sheds memory weight, making it perfect for resource-constrained environments. Plus, Gemini 1.5 expands its skillset, tackling diverse tasks and domains with newfound accuracy and fluency.

For data scientists and AI enthusiasts:

• SORA: Get ready to chat with AI in multiple languages! SORA, Google's multilingual language model, now understands and responds in English, Japanese, and French, with more languages on the horizon. It delves deeper into complex topics, grasps emotions and opinions, and even flexes its creative muscles by generating poems, code, scripts, and more. Prepare for truly conversational AI experiences and versatile content creation tools.
• AWS LLRT (Logarithmic Lambda Runtime): This experimental runtime might just be the answer to your serverless woes. LLRT cuts memory usage by up to 90% and boasts faster startup times, translating to significant cost savings and improved performance for memory-hungry Lambda functions. Embrace the green side of tech with LLRT's reduced environmental impact too.

Gemini 1.5

Google recently announced the release of Gemini 1.5, the next-generation AI model, following the rollout of Gemini 1.0 Ultra. Gemini 1.5 offers significant advancements in performance, leveraging a new Mixture-of-Experts (MoE) architecture and enhancing long-context understanding capabilities. It can process up to 1 million tokens, enabling the analysis of vast amounts of information across various modalities such as text, code, images, audio, and video. Gemini 1.5 Pro, the initial model available for testing, maintains high performance even with extended context windows, outperforming its predecessor in most benchmarks. The model undergoes extensive ethics and safety testing, aligning with Google's AI Principles. A limited preview of Gemini 1.5 Pro is now accessible to developers and enterprise customers via AI Studio and Vertex AI, with plans for wider release and pricing tiers based on context window size.

SORA

OpenAI is developing Sora, a text-to-video model, to teach AI about the physical world in motion. Sora can create videos up to a minute long, aiming to assist people in solving real-world problems. It's being made available to red teamers and creative professionals for assessment and feedback. While Sora shows promise in generating complex scenes and understanding language, it still has weaknesses in accurately simulating physics and understanding cause and effect. Safety measures, including adversarial testing and content detection tools, are being implemented before deployment. Sora is a diffusion model, leveraging a transformer architecture like GPT models, to generate videos gradually while maintaining subject consistency. It builds upon past research in DALL·E and GPT models, aiming to be a foundation for understanding and simulating the real world, a step towards achieving AGI.

Flutter 3.19

Flutter 3.19 introduces several new features and improvements. It includes a new Dart SDK for Gemini, a widget for controlling widget animations, updates to Impeller for rendering boost, deep link implementation tools, Windows Arm64 support, and more. The release also highlights community contributions, with 1429 pull requests merged by 168 community members, including 43 who made their first contributions. Notable additions include AI integration with the Google AI Dart SDK beta release, scrolling improvements, a new AnimationStyle widget, enhancements to the TableView widget, and various engine improvements such as Impeller progress and GPU tracing.

AWS LLRT

Amazon Web Services (AWS) launched Limited Liability Right Transfer (LLRT) on February 15, 2024. LLRT is a new feature that allows customers to grant limited access to their AWS resources to third-party vendors or partners. This can be beneficial for tasks such as granting a managed service provider (MSP) access to manage your AWS infrastructure. Enabling a software vendor to access specific resources needed to run their application on your AWS account. LLRT provides granular control over the level of access granted, ensuring security and compliance.