WakeTheAI
Posts
🏛️ OpenAI launches ChatGPT Gov

🏛️ OpenAI launches ChatGPT Gov

PLUS: Qwen unveils AI models that control devices

Jaynit Makwana & Dhaval Makwana
January 29, 2025

Sign up | Sponsor

Hola, AI fam 🤖

In today’s WakeTheAI edition:

OpenAI launches ChatGPT Gov
Qwen unveils AI models that can control devices
Prompt: Apply Eisenhower Matrix To Tasks
Convergence AI launched ‘Proxy’ globally today
Pika Labs launched Pika 2.1 revolutionizing AI video
DeepSeek dethrones ChatGPT as the No. 1 app on iOS
OpenAI suspects China’s DeepSeek used its models for training
Reid Hoffman’s Manas AI raises $24.6M for AI-driven drug discovery

OPENAI
🏛️ OpenAI launches ChatGPT Gov

Lazy Bits: Yesterday, OpenAI announced ChatGPT Gov, a secure AI model for U.S. government agencies, designed to enhance efficiency, security, and AI adoption. Since 2024, 90,000+ users from 3,500+ agencies have exchanged 18M+ messages using ChatGPT.

In-Depth Details:

Secure Deployment: Self-hosted on Microsoft Azure, ChatGPT Gov meets IL5, CJIS, ITAR, and FedRAMP High standards for handling sensitive government data.
AI-Powered Services: Agencies can use GPT-4o for text analysis, summarization, coding, and image interpretation, improving productivity across government tasks.
Customizable Features: Employees can save and share conversations, upload files, and build Custom GPTs, with IT teams managing access via an admin console.
Real-World Use: The Air Force Research Lab improves resource access, Los Alamos National Lab applies AI in biosciences, and Minnesota enhances translation services.
Productivity Gains: A Pennsylvania pilot found ChatGPT Enterprise saved employees 105 minutes daily, reducing routine workload.
Future Expansion: OpenAI is pursuing FedRAMP certifications and exploring Azure classified cloud integration for high-security applications.

Lazy Conclusion: OpenAI’s ChatGPT Gov cements AI’s role in U.S. government operations, boosting efficiency and security. As global AI competition intensifies, this move signals the U.S.’s push to stay ahead in AI-driven governance.

TOGETHER WITH NEWSBANG
📰 NewsBang: Smarter News, Zero Noise

Lazy Bits: NewsBang redefines the way you consume news by delivering AI-driven insights, interactive podcasts, and smart summaries—all in one place.

In-Depth Details:

Swipe for Insights – Go beyond the surface with AI-generated deep dives into the news that matters.
AI Podcasts – Listen to real-time audio updates and ask AI-powered questions while on the go.
Ask AI – Get instant answers and verified sources for any news topic, eliminating misinformation.
Smart Summaries – Scan bite-sized takeaways from diverse sources on a single screen for quick understanding.

Ready to experience the future of AI-powered news?

Try NewsBang now (Click here)

ALIBABA
🧠 Qwen unveils AI models that control devices

Lazy Bits: Yesterday, Alibaba’s Qwen announced Qwen2.5-VL, its most advanced vision-language model with enhanced image recognition, long-video comprehension, agent-like reasoning, structured document parsing, and improved efficiency.

In-Depth Details:

Advanced Visual Understanding: Qwen2.5-VL can analyze objects, text, and layouts in images, localize objects with bounding boxes, and generate structured JSON outputs.
Long Video Comprehension: It processes videos over an hour long, identifies key moments, and provides accurate segment-based summaries.
Agent-Like Capabilities: The model functions as a visual agent, enabling reasoning and tool interaction across computer and mobile interfaces.
Document & Text Recognition: With upgraded OCR and its QwenVL HTML format, it enhances text recognition, document parsing, and information extraction.
Improved Model Efficiency: Window Attention in ViT optimizes computational load, improving performance while maintaining high-resolution accuracy.
Competitive Performance: The 7B model beats GPT-4o-mini in multiple tasks, while the 3B version surpasses Qwen2-VL’s previous 7B model.

Lazy Conclusion: Qwen2.5-VL is a major leap in vision-language AI, rivaling OpenAI’s GPT-4o in visual reasoning, long-video analysis, and structured outputs. As China accelerates AI innovation, models like Qwen2.5-VL highlight the growing competition between US and Chinese AI labs.

Apply Eisenhower Matrix To Tasks

You can ask ChatGPT to act as your task management strategist, helping you implement the Eisenhower Matrix to prioritize tasks efficiently.

ChatGPT will categorize tasks based on urgency and importance, provide actionable recommendations for each quadrant, and suggest strategies for effective execution.

The output will include a structured markdown table and clear insights, ensuring practical and adaptable task management.

Prompt:

Act as a task management strategist specializing in the Eisenhower Matrix. Your task is to create a structured framework to categorize and prioritize tasks based on urgency and importance. The output should address the following key areas:

Key Focus Areas:

1. Evaluating Tasks for Prioritization – Analyze each task based on its level of urgency and importance.
2. Classifying Tasks into Quadrants – Assign each task to the appropriate category within the Eisenhower Matrix.
3. Providing Task-Specific Recommendations – Offer guidance on handling tasks in each quadrant for optimal productivity.
4. Enhancing Task Execution – Suggest actionable strategies to manage workload efficiently.

Key Information About Me:

Task List: [List your current tasks]
Context: [Describe your situation, deadlines, or workload challenges]
Priorities: [Outline your main objectives and productivity goals]

Output Requirements:

• Present the Eisenhower Matrix in a markdown table format with four columns:

Urgent & Important – Tasks that need immediate attention and should be done first.
Important but Not Urgent – Tasks that contribute to long-term goals but don’t require immediate action.
Urgent but Not Important – Tasks that should be delegated to free up time for high-priority work.
Neither Urgent nor Important – Tasks that provide little value and should be eliminated or minimized.

• Include brief explanations or recommendations below the table to guide execution.
• Use bullet points and structured insights to ensure clarity and actionable takeaways.
• Ensure the system is practical, adaptable, and applicable across different work scenarios.

Result:

OpenAI suspects China’s DeepSeek used its models for training, violating API terms. Microsoft detected data exfiltration in 2024, and an investigation is underway.
Convergence AI launched ‘Proxy’ globally, Europe's answer to OpenAI's Operator, offering free basic access and a $20/month unlimited session plan.
Pika Labs launched Pika 2.1 revolutionizing AI video creation with 1080p resolution, advanced motion control, extended video length, and realistic physics, offering pro-level tools for creators.
DeepSeek dethrones ChatGPT as the No. 1 app on iOS, surging to 2.6M downloads. Its AI models, trained cheaply, shook the industry, even hitting Nvidia’s stock.
Reid Hoffman's Manas AI raises $24.6M for AI-driven drug discovery, focusing on cancer.