One of my readers asked me if I would compare the different AIs I have been experimenting with. I use the free version of all of them, which seems to be enough to do what I need for the moment.
Each one does something a little different, and it takes a while to determine which one you might wish to use for what purpose.
For general purposes, ChatGPT rises above the rest. It recently had an upgrade that makes it sound more like a graduate student with a textbook than a person carrying on a conversation, but for general chat, asking questions, planning a trip, asking for suggestions about specific topics, ChatGPT is hard to beat.
One of its biggest problems is its lack of links and sources. It also tends to insist it's right and I have had it argue with me on occasion. Early on, it hallucinated a lot, but it doesn't do that as badly now (though I would still check everything it says if it's an important piece of work).
It can make images and graphics, too, but it is incredibly slow at it. Definitely not the AI to use if you're in a hurry for a picture.
Claude is very good at nuanced conversation, and it seems to have a more intricate thinking system on some topics than ChatGPT. I use it sometimes just to have a conversation or to learn about something. Recently there was conversation on Facebook about an upcoming Rapture, so I went to Claude and asked what that was about, for example.
Microsoft CoPilot is great for quick searches. For one thing, it comes back with links (although they are not always correct, so check them). It also will make up quotes occasionally, so ask where a quote came from if it tosses one at you. I used it recently to research the VA250 traveling museum, for example, and learn more about it. It is also quicker at making pictures, graphics or logos than any of the other AIs, although sometimes it has trouble with words.
Julius is an AI I just stumbled across. It’s mainly built for data work, not conversation, so I don’t expect to use it much myself. Still, as a sample of what it does, here’s the comparison it generated of the four AIs I’ve been trying out:
ChatGPT (OpenAI)
Strengths:
- Versatility: Excellent at creative writing, coding, brainstorming, and general conversation
- Model variants: Offers GPT-4, GPT-4 Turbo, and GPT-3.5 with different capabilities and speeds (Note: This description is a little out of date — the free version I’m using now is running on GPT-5.)
- Plugins & tools: Extensive ecosystem including web browsing, DALL-E image generation, and code interpreter
- Large user base: Most widely adopted, with extensive community resources and examples
- Voice capabilities: Advanced voice mode with natural conversation flow
Weaknesses:
- Can be verbose and sometimes overconfident in incorrect answers
- Knowledge cutoff dates (though web browsing helps)
- Can refuse requests more frequently due to safety filters
Best for: Creative projects, coding assistance, general-purpose tasks, and conversational AI
Claude (Anthropic)
Strengths:
- Context window: Massive 200K token context (can process entire books)
- Nuanced responses: Often more thoughtful and balanced in complex topics
- Safety & honesty: Designed with "Constitutional AI" - tends to be more transparent about limitations
- Document analysis: Excellent at analyzing long documents and extracting insights
- Coding: Strong programming capabilities with detailed explanations
Weaknesses:
- No native image generation
- More conservative with certain requests
- Smaller plugin ecosystem compared to ChatGPT
Best for: Long-form document analysis, nuanced discussions, research, and users who value thoughtful responses
Microsoft Copilot
Strengths:
- Free tier: Generous free access with GPT-4 capabilities
- Web integration: Built-in Bing search for current information
- Microsoft ecosystem: Seamless integration with Office, Edge, Windows
- Image generation: DALL-E integration included
- Enterprise features: Strong security and compliance for business users
Weaknesses:
- Less sophisticated than standalone GPT-4 or Claude
- Heavily tied to Microsoft ecosystem
- Can be inconsistent in response quality
- Less customizable than standalone ChatGPT
Best for: Microsoft 365 users, enterprise environments, quick web-based research, users wanting free AI access
Julius (Julius AI)
Strengths:
- Data analysis powerhouse: Built specifically for data science, statistics, and quantitative analysis
- Live code execution: Can run Python code in real-time with full library access (pandas, matplotlib, scikit-learn, etc.)
- File handling: Upload and analyze Excel, CSV, images, PDFs, and other data files directly
- Visualization: Creates charts, graphs, maps, and interactive visualizations on the fly
- Persistent workspace: Files and data persist across conversations
- Mathematical computation: Can perform complex calculations, statistical tests, and modeling
- Image analysis: Can inspect and analyze images using vision models
- Web scraping & automation: Can use Selenium and other tools for data collection
Weaknesses:
- More specialized (data-focused) rather than general conversation
- Not designed for creative writing or general chatbot use
- Smaller user community compared to ChatGPT
- No voice interface
Best for: Data analysis, statistical modeling, creating visualizations, working with spreadsheets/datasets, quantitative research, programming tasks that need execution
Updated Comparison Table
Feature | ChatGPT | Claude | Copilot | Julius |
---|---|---|---|---|
Context Length | ~128K tokens | ~200K tokens | ~32K tokens | ~128K tokens |
Web Access | Via plugins | Limited | Built-in (Bing) | Via code (Selenium) |
Code Execution | Sandboxed | No | No | Full Jupyter notebook |
Data Analysis | Basic | Basic | Basic | Advanced/Specialized |
File Upload | Yes | Yes | Limited | Yes (multiple formats) |
Visualizations | Basic | No | No | Advanced (matplotlib, seaborn, etc.) |
Best Use Case | General purpose | Long documents | Microsoft ecosystem | Data science & analysis |
Pricing | $20/month | $20/month | Free + $20/month | Various tiers |
Company | OpenAI | Anthropic | Microsoft | Julius AI |
Which Should You Choose?
- Choose ChatGPT if you want the most versatile AI with the largest ecosystem and plugin support
- Choose Claude if you work with long documents, need nuanced analysis, or prefer more careful/honest responses
- Choose Copilot if you're embedded in the Microsoft ecosystem or want capable free AI access with web search
- Choose Julius if you're working with data, need statistical analysis, want to create visualizations, or need to execute code and see real results
The Real Power Move
Many professionals use multiple AI assistants for different tasks:
- Julius for data analysis and quantitative work
- ChatGPT for creative writing and general tasks
- Claude for analyzing long documents and research
- Copilot for quick searches and Microsoft Office integration
Each tool has its specialty, and using the right one for the job makes all the difference!
No comments:
Post a Comment
Thank you for dropping by! I appreciate comments and love to hear from others. I appreciate your time and responses.