I asked ChatGPT to generate SWOT analysis on Claude.
Then asked ChatGPT to generate a SWOT on itself.
I then asked Claude for a SWOT on ChatGPT.
Then a SWOT on itself.
I then copied and pasted all the responses back into Copilot in a new message and asked for succinct summaries.
I refined all the responses in Word.
Then, created 3 images in ppt.
Time 15 minutes.
I LEARNED ALOT, FAST !!
Below are comparisons between ChatGPT and Claude regarding their knowledge cut-off dates and capabilities:

ChatGPT:
- Last training update was January 2022
- Claims it can access the web to look up current information
- Offers to search for up-to-date information
Claude 3.5 Sonnet:
- Last updated in April 2024
- Explicitly states it cannot access the internet or look up real-time information
- Is transparent about its limitations regarding information after its April 2024 cutoff date
- Recommends consulting authoritative sources for the most current information
The key difference is that while ChatGPT claims to have web access capabilities, Claude clearly states it cannot access the internet or look up current information, demonstrating different approaches to transparency about their capabilities and limitations.


Image Generation comparison
I asked Google Gemini to explain the benefits and challenges of generating images with Claude AI, Google Gemini Imagen, ChatGPT, Copilot in Edge summmarise below.
Claude AI is best for tasks that require strong natural language understanding and creative text generation. So NOT images!
Google Gemini Imagen is the top choice for generating high-quality, photorealistic images.
ChatGPT is a versatile tool that can perform a wide range of tasks, including generating creative text formats and translating languages.
Copilot in Edge is best for coding tasks, such as generating code suggestions and completing code snippets.
I also did a comparison using the same prompt, and other than the Claude crap image, the other three are similar in high-quality so have all improved since their knowledge cut-off training input.


Leave a reply to Simon Hudson Cancel reply