Github Archives

Last Week in AI

1. Samsung’s Generative AI: Samsung Gauss

What: A new generative AI family, featuring models for language processing, code generation, and image editing.
Features:
- Language Model: Similar to ChatGPT, adept at understanding and generating human-like text.
- Code-Generating Model: Automates and streamlines coding processes.
- Image Generation and Editing Model: Creates and modifies visual content.
Availability: Currently in internal use, set for public release soon.

2. Microsoft’s AI Support for Startups

What: Free Azure AI infrastructure for startups.
Benefits:
- No-Cost Access: High-end Nvidia GPU virtual machine clusters available at no cost.
- Enhanced AI Training: Ideal for training and operating generative models.
- Scalability: Supports the growth and development of startup projects.

3. YouTube’s AI Fe atures

What: New generative AI features for premium subscribers.
Features:
- Conversational Tool: Provides AI-based answers and recommendations.
- Comment Summarization: Summarizes and highlights key points in video comments.

4. DeepMind’s Robotics Ambitions

What: Insights from Vincent Vanhoucke, head of robotics at Google DeepMind.
Focus Areas:
- General-Purpose Robots: Envisioning versatile, multi-functional robotic solutions.
- Generative AI Integration: Enhancing robot capabilities with advanced AI technology.

5. K ai-Fu Lee’s AI Model: 01.AI

What: The unveiling of the Yi-34B model by Kai-Fu Lee’s AI startup.
Highlights:
- Open Source: Accessible and modifiable by the developer community.
- Innovative Design: Developed within seven months of the startup’s founding.

6. GitHub’s Customizable Copilot Plan

What: A new enterprise subscription tier for Copilot.
Customization:
- Codebase-Specific Tailoring: Adapts to the unique coding practices of different companies.
- Enhanced Programming Support: Assists in code generation and problem-solving.

7. Hugging Face’s Two-Person Model Team: H4

What: A small, efficient team developing advanced AI tools.
Key Offerings:
- Data Science Hosting: Provides hosting services for AI and machine learning projects.
- Development Tools: Offers a suite of tools for AI development and application.

8. Mozilla’s AI Ch atbot: Fakespot Chat

What: An AI-powered shopping assistant developed by Mozilla.
Capabilities:
- Product Query Assistance: Answers questions about online products.
- Shopping Suggestions: Offers recommendations based on user queries and preferences.

9. ChatGPT Custom GPTs

“ChatGPT custom GPTs is a significant advancement that deserves its own detailed section. For an in-depth look at this exciting development, check it out here.

That wraps up our update. If you didn’t catch the previous ‘Last Week in AI,’ you can find it here. See you next week, and in the meantime, keep engaging, stay curious, and enjoy the journey through the ever-evolving world of AI.

Last Week in AI Read More »

GitHub’s Big Move: Copilot Enterprise

Copilot Enterprise: What Does It Do?

Copilot Enterprise is more than just a coding tool. It can explain and give advice on internal source code, making it easier for developers to understand and improve their work. Think of it as a smart assistant for coding in the workplace.

The Price Tag: $39 per Person per Month

It’s not cheap, though. This new tier comes at $39 per person each month. That’s over double what the business offering costs. But for what it does, many companies might find it worth the expense.

Microsoft’s Strategy: Boosting Cloud Business

Why is Microsoft doing this? It’s all about boosting their cloud business segment. By using OpenAI’s tech more fully, Microsoft is aiming to make a bigger splash in the cloud services market.

A Million Paid Users and Growing

Just last month, Microsoft CEO Satya Nadella shared some impressive numbers. GitHub Copilot has already got a million paid users. That’s a lot of developers finding value in this AI-driven coding assistant.

Caution: Check the Code

But it’s not perfect. There have been some errors, and security pros are keeping a close eye on it. That’s why GitHub advises developers to test and review any code Copilot suggests, just to be safe.

A Billion-Dollar Future?

Looking ahead, analysts at Piper Sandler are making some bold predictions. They estimate GitHub Copilot could bring in $3 billion annually by 2026, assuming 16% of GitHub’s 100 million users sign up. That’s a huge potential market.

So, what’s next? GitHub’s Copilot Enterprise could be a game-changer for developers, offering smart, AI-powered coding assistance in the workplace. For companies that rely heavily on software development, this could be the tool that takes them to the next level.

Check out the GitHub Universe 2023 opening keynote.

GitHub’s Big Move: Copilot Enterprise Read More »

Github