Explanation and Access to Leading Generative AI Systems

A useful reference for new and experienced users of generative AI.

September 26, 2024

Several subscribers have asked for more explanation on accessing some of the more popular generative AI systems. So here is a brief overview. Note that this information is of the date published and will change over time.

First, some explanation of terminology. When referring to these AI systems, the term "model" describes the underlying technology, while "application" focuses on the user-facing utility, and "system" encompasses the full suite of operational components.

ChatGPT

ChatGPT, developed by OpenAI, is an AI-powered chatbot based on the GPT (Generative Pre-trained Transformer) architecture. It has been trained on vast amounts of internet text data to understand and generate human-like responses. ChatGPT predicts the next word or sentence when users input prompts or questions based on its training, producing context-aware responses. The model can answer questions, engage in conversations, assist with tasks such as writing and coding, and provide information across various topics. However, it doesn’t access live data unless integrated with external tools, and it doesn't "think" like a human; it relies on learned patterns.

ChatGPT Versions and Pricing

ChatGPT offers two main subscription tiers:

• Free Version: This version offers basic features, such as conversational AI for writing, problem-solving, and more. Limitations include slower response times and shorter session lengths.

• ChatGPT Plus: For $20/month, subscribers can access more advanced models like GPT-4, faster responses, longer sessions, and priority access to new features.

The most advanced model, GPT-4, offers text, voice, and vision improvements. For example, GPT-4 can analyze images and translate menus, offer historical insights, and provide food recommendations. It’s expected to support voice and video interactions soon, enabling real-time conversations and more natural exchanges.

ChatGPT's API uses a token-based pricing structure, where costs are based on the number of tokens (pieces of words) processed during requests and responses. The cost per token varies depending on the model you use.

How to Access ChatGPT

ChatGPT can be accessed in several ways.

1. The easiest way to access it is simply using a web browser at
    https://openai.com/chatgpt/

2. There are also mobile apps for ChatGPT available on iOS and Android devices. It looks like this in the Apple app store. Be careful; there are similar apps that charge additional money for access.

3. You can also use systems or apps that incorporate Openai through an API. To do so, you need to sign up for an account at Openai and then get an API key for the applications. I use MacGPT as an application on my desktop with a ChatGPT API.

Anthropic (Claude)

Claude is an advanced AI assistant developed by Anthropic and designed for diverse tasks and conversations. Known for its strong language understanding and generation, Claude excels at problem-solving, analysis, and creative tasks. It emphasizes ethical interactions by providing accurate information and being transparent about its limitations.

Claude is versatile, handling various tasks, from casual conversations to technical coding, creative writing, and data analysis. It also incorporates safeguards to prevent harmful content and ensure user privacy and ethical integrity.

Claude Versions and Pricing

Claude comes in several versions:

1. Claude 3 Haiku: Fastest model for everyday tasks.

2. Claude 3 Sonnet: Balanced model, offering both speed and capability.

3. Claude 3.5 Sonnet: Enhanced version of Sonnet with better capabilities.

4. Claude 3 Opus: The most powerful model, suitable for complex tasks.

I currently use the Claude 3.5 Sonnet free version.

Anthropic offers Claude through:

• API Access: Pricing varies based on usage and the model selected.

• Claude Pro: Available in some regions, with pricing subject to variation.

• Enterprise Solutions: Custom pricing based on business needs.

Anthropic has several pricing plans for Claude.

Free Version

• For individuals to get started

• Talk to Claude on the web, iOS and Android

• Ask about images and docs

• Access to Claude 3.5 Sonnet

Pro Version - $20/month per person

• For Claude power users

• Everything in Free, plus:

• More usage than Free

• Priority bandwidth and availability

• Use Claude 3 Opus and Haiku

• Early access to new features

Team Version - $25/month per person

• For fast-growing teams

• Everything in Pro, plus:

• More usage than Pro

• Central billing and administration

• Early access to collaboration features

Enterprise Version – Custom Pricing

• For businesses operating at scale

• Everything in Team, plus:

• More usage than Team

• Expanded context window

• Single sign-on (SSO) and domain capture

• Role-based access with fine-grained permissions

• System for Cross-domain Identity Management (SCIM)

• Audit logs

• Data source integrations

How to Access Claude

Claude can be accessed through:

• The website at anthropic.com

• Various third-party applications and platforms integrate Claude.

Google (Bard and Gemini)

Google offers two generative AI systems: Bard and Gemini. While both are large language models (LLMs), they differ in capabilities and training.

• Bard, released in March 2023, focuses on text-based tasks like content creation, language translation, and answering questions. It is trained on a large dataset of text and code.

• Gemini, introduced in February 2024 as a successor to Bard, is more versatile. It processes text, images, and audio and is designed for more complex, multimodal tasks, including reasoning and problem-solving.

Gemini Versions

Gemini has four variations:

1. Gemini Ultra

2. Gemini Pro

3. Gemini Flash: A faster, distilled version of Pro.

4. Gemini Nano: Two small models, the Nano-1 and the more capable Nano-2, are designed for offline use.

As of September 2024, there is no publicly available information about the pricing structure for different versions of Gemini. Most use is currently free. Google has not released specific details regarding the cost or subscription models for accessing and using Gemini. Pricing will likely depend on factors such as Gemini's specific version or capabilities, the scale of usage, and any additional features or services offered.

How to Access Bard and Gemini

Bard can be accessed via a web browser at bard.google.com, and Gemini is available at https://gemini.google.com/app.

Microsoft (Copilot)

Microsoft CoPilot is an AI-powered assistant integrated into Microsoft products. It helps users increase productivity through natural language processing and automation. CoPilot is built on OpenAI's GPT models and is deeply embedded in Microsoft Office applications, including Word, Excel, PowerPoint, and Teams.

While useful with Microsoft applications, it isn’t as helpful as generative AI applications.

Versions of Microsoft CoPilot

Microsoft CoPilot comes in different versions designed for specific applications or platforms within the Microsoft ecosystem:

Microsoft 365 CoPilot

• Integrated directly into the Microsoft 365 suite, including Word, Excel, PowerPoint, Outlook, and Teams.

• It assists with document creation, data insights, and communication management.

• Users can interact with CoPilot via natural language prompts to perform various tasks like drafting emails, generating Excel reports, or summarizing meetings in Teams.

• It is deeply embedded in the Office workflow, helping with tasks like adjusting slides in PowerPoint or formatting documents in Word.

GitHub CoPilot

• Specifically designed for coding environments, GitHub CoPilot is an AI-powered tool that assists developers by auto-completing code, suggesting snippets, and helping debug errors.

• It integrates with various code editors, such as Visual Studio Code.

• This version leverages AI to predict the next lines of code or provide suggestions, reducing the effort needed for routine coding tasks and speeding up the development process.

• It also helps explain complex code or provide recommendations for approaching a programming challenge.

CoPilot for Business

• This version is tailored to business users who need assistance with tasks like generating reports, analyzing market trends, or managing large datasets.

• Integrated with Microsoft Dynamics 365, CoPilot for Business enhances sales, marketing, and customer service by offering automated suggestions, customer insights, and data-driven forecasts.

• It also enables users to automate customer communications, personalize outreach, and predict customer behavior.

Security CoPilot

• Aimed at IT and security professionals, Security CoPilot provides AI-driven insights and automation in cybersecurity environments.

• It integrates with Microsoft’s security stack (Microsoft Defender, Sentinel, etc.) to help identify threats, provide responses, and mitigate risks more efficiently.

• It analyzes large datasets from security logs, flags vulnerabilities, and suggests remediation actions.

Key Differences Between the Versions

Each version of CoPilot is designed for a specific domain—Microsoft 365 CoPilot focuses on office productivity, GitHub CoPilot is for developers, CoPilot for Business targets sales and customer engagement, and Security CoPilot focuses on cybersecurity.

 Depending on the version, CoPilot is embedded into specific Microsoft platforms, such as Microsoft 365 apps for office productivity or GitHub for coding. The core functionality of text generation and task automation is present across all versions. Still, the specific capabilities (like coding assistance in GitHub or security threat detection in Security CoPilot) vary based on the platform.

How to access Microsoft CoPilot

You can access CoPilot via:

• Web-based CoPilot at copilot.microsoft.com

• Microsoft Edge Sidebar

• The CoPilot Mobile App for iOS and Android

Bing Chat (Now CoPilot)

Bing Chat, integrated into Microsoft's Bing search engine, uses GPT-4 to provide conversational search and interactive responses to user queries. Unlike traditional search engines, Bing Chat offers in-depth, natural language responses and can pull real-time data from the web, making it useful for up-to-date information.

Bing Chat is an AI-powered chatbot integrated into Microsoft’s Bing search engine. It leverages large language models (LLMs), specifically OpenAI’s GPT-4 technology, to provide conversational search and interactive responses to user queries. Unlike traditional search engines that return a list of links, Bing Chat offers more in-depth, natural language responses, allowing users to ask complex or follow-up questions in a conversational style.

Bing Chat and Microsoft CoPilot are complementary AI-powered tools. Both are built on similar underlying GPT-based technology from OpenAI, but they serve different purposes and are integrated into different platforms. Together, they form a cohesive ecosystem of AI-enhanced productivity tools across the Microsoft landscape.

Bing Chat has a distinct role. It is primarily designed for interactive, conversational search and content generation via Bing’s platform. It functions as a standalone chatbot embedded in Bing and can be accessed directly through the search engine or the Microsoft Edge browser sidebar.

Versions and Pricing

For the most integrated Copilot experience, download Microsoft Edge. Copilot is built into the browser’s sidebar for fast and easy access. Select the Copilot icon to open the Copilot pane using Microsoft Edge. You’ll have access to Copilot, Compose, and Insights tabs: three powerful and distinctive AI tools.

Copilot is also accessible on your smartphone or tablet. You’ll need to download the Copilot mobile app on iOS or Android to use it on the go. Here is the pricing.

Copilot (formerly Bing Chat)

• Free tier with limited features

• Copilot Pro: $20/month (priority access, enhanced features)

GitHub Copilot

• Individual: $10/month or $100/year

• Business: $19/user/month

• Enterprise: Custom pricing

How to Access to Bing Chat/CoPilot

Bing Chat can be accessed through:

• Bing’s website: Visit bing.com and select the "Chat" button.

• Microsoft Edge Sidebar: For quick access.

• The Mobile App: Downloadable for smartphones and tablets.

Apple Intelligence

Apple Intelligence is a generative artificial intelligence (AI) system developed by Apple Inc. It's designed to enhance user experiences across various Apple devices by offering intelligent assistance and capabilities.   

Key features and functionalities include:

• Writing Tools: Helps users with writing tasks, such as composing emails, messages, and documents.   

• Photos Clean Up: Suggests organizing and deleting photos based on relevance and similarity.   

• Notification Summaries: Groups similar notifications together for better management.

• Image Playground: Allows users to create and edit images using AI-powered tools.   

• Genmoji: Generates personalized emoji suggestions based on context.   

• Siri Integration: Enhances Siri's capabilities with improved understanding and responses.   

Apple Intelligence is designed to be privacy-focused, processing data locally on devices whenever possible to protect user information. It's expected to be a significant addition to Apple's ecosystem, providing users with more intelligent and personalized experiences.   

Apple’s AI ecosystem focuses on enhancing usability and personalization while maintaining privacy, setting its AI solutions apart from cloud-dependent systems.

AI in Apple Products: A Summary

• Siri: Voice assistant powered by NLP and on-device learning.

• Photos: Facial recognition, object detection, and smart search.

• Health/Fitness: Predictive insights and monitoring using AI.

• Maps: Real-time traffic and location-based suggestions.

• Face ID/Touch ID: Secure, AI-driven biometric authentication.

• Privacy-Focused: On-device intelligence for secure AI processing.

How to Access Apple Intelligence

Apple Intelligence is only currently available in beta test versions. I’ve used these, and there is little functionality yet. It will be released progressively over the next 6-9 months.

Other Systems

There are several other generative AI systems that you may want to consider trying. Some of these specialize in specific functions.

1.  Midjourney (Image Generation)

Midjourney is an AI tool for generating images based on user prompts. It is accessed primarily through Discord. There are currently four pricing tiers.

1. Basic: $10/month

2. Standard: $30/month

3. Pro: $60/month

4. Mega: $120/month

Higher tiers offer faster image generation and more usage.

How to Access Midjourney

You can access Midjourney through two main methods:

Discord: This is the original and primary method. You'll need a Discord account to join the official Midjourney server. Once you have a subscription, you can generate images using the /imagine command.

Midjourney Web: This is a newer option that offers a web-based interface. You can access it directly through the Midjourney website.

Here's a brief overview of the steps:

1. Create a Discord account (if you don't have one already).

2. Join the Midjourney Discord server.

3. Subscribe to a Midjourney plan.

4. Use the /imagine command in a designated channel to generate images.

For more detailed instructions and tips, you can refer to the Midjourney documentation

2.  Adobe Generative AI (Firefly)

Adobe's Firefly is a generative AI tool integrated into its suite of creative applications. It focuses on content generation for creative professionals and enables users to create images, enhance photos, generate text effects, and automate tasks in Adobe applications like Photoshop, Illustrator, and Premiere Pro.

Key Capabilities:

• Image Generation: Create custom visuals from text prompts.

• Text Effects: Generate artistic text styles and effects.

• Photo Enhancements: Automatically adjust and enhance images.

• Video Editing: Simplifies video production by automating effects and generating scenes.

Versions and Pricing

• Free Access: Available with limited features for experimentation.

• Creative Cloud Subscription: Firefly is included as part of Adobe's Creative Cloud, with pricing plans starting at:

• Individual Plan: $54.99/month for access to all Adobe apps.

• Photography Plan: $9.99/month for Adobe Photoshop and Lightroom (with limited Firefly access).

• Enterprise Plan: Custom pricing based on business needs.

3.  Grammarly Generative AI (GrammarlyGO)

GrammarlyGO** is Grammarly's generative AI tool for writing and editing tasks. It helps users generate content, rewrite text, and improve tone and clarity. GrammarlyGO is available within Grammarly's writing assistant, which integrates into browsers, desktop apps, and mobile devices.

Key Capabilities:

• Content Generation: Generate complete drafts from prompts.

• Text Rewriting: Improve text for tone, clarity, or style.

• Suggestions: Personalized writing suggestions based on user input and writing habits.

• Context-Aware Edits: Tailored suggestions for formal or casual writing, emails, reports, and essays.

Versions and Pricing

• Free Version: Basic grammar and spelling checks with limited AI-assisted writing features.

• Premium Plan: $12/month, offering full access to advanced AI features, including GrammarlyGO.

• Business Plan: $15/member/month, offering collaboration tools and enterprise-level AI writing support.

How to Access Firefly and GrammarlyGO

Adobe Firefly and GrammarlyGO are integrated into their respective ecosystems, enhancing user productivity through AI-driven automation and content generation.

4.  Stability AI (Stable Diffusion)

Stable Diffusion, by Stability AI, is a generative AI model that produces high-quality images from text inputs.

Versions and Pricing

• DreamStudio: Pay-as-you-go credits system with membership tiers:

  • Starter: $10/month

  • Professional: $40/month

  • Ultimate: $80/month

• API Access: Custom pricing based on usage.

How to Access Stability AI

Access is available through the web-based interface at dreamstudio.ai or via API.

5.  Replicate

Replicate offers a platform for running various AI models, using a pay-per-use pricing model that varies by model and usage.

How to Access Replicate

Replicate can be accessed through its web interface at replicate.com or via API.

6.  Runway ML

Runway ML provides AI tools for creative professionals specializing in video editing and media generation.

Versions and Pricing

• Free: Limited access.

• Standard: $15/month (billed annually).

• Pro: $35/month (billed annually).

• Unlimited: Custom pricing.

How to Access Runway ML

Runway ML is available via:

• A web-based platform at runwayml.com

• Desktop and mobile applications.

7.     Hugging Face

Hugging Face offers access to numerous open-source AI models for various tasks, including text generation, image analysis, and more.

Versions and Pricing

• Free: Access to community models.

• Pro: $9/month.

• Enterprise: Custom pricing.

How to Access Hugging Face

Hugging Face models can be run via its web interface at huggingface.co or accessed through an API.

General Access Tips

There are some general guidelines and tips for accessing most of these generative AI systems.

1. Most services require users to create an account.

2. API access typically requires obtaining API keys or tokens.

3. Many services offer both web interfaces and mobile applications.

4. Developer-focused platforms often provide SDKs or libraries for various programming languages.

Access methods may vary by region, and it's always best to check the official websites for the latest information on how to access each service.

If you don’t want to miss any of my new articles, please subscribe and receive them by email.

Be sure to check out my other articles on generative AI. To read the previous articles, click this link.