Paul Beck
@paulbeck1.bsky.social
Tech guy based in London looking at architecture , Azure, power platform, Pulumi, playwright, low code, automation, Microsoft Fabric.
🇬🇧🇿🇦💻🐘👨💻🧪☁️🪠
🇬🇧🇿🇦💻🐘👨💻🧪☁️🪠
Bing/M365 copilot was solid similar to ChatGPT response but missing a key point and I didn’t like the way the result was presented.
Deepseek, first time I’ve used. It wouldn’t do the calc/response. It worked on simple queries. Might be too busy due to its recent release but it failed miserably.
Deepseek, first time I’ve used. It wouldn’t do the calc/response. It worked on simple queries. Might be too busy due to its recent release but it failed miserably.
January 29, 2025 at 2:45 PM
Bing/M365 copilot was solid similar to ChatGPT response but missing a key point and I didn’t like the way the result was presented.
Deepseek, first time I’ve used. It wouldn’t do the calc/response. It worked on simple queries. Might be too busy due to its recent release but it failed miserably.
Deepseek, first time I’ve used. It wouldn’t do the calc/response. It worked on simple queries. Might be too busy due to its recent release but it failed miserably.
Thoughts:
Claude won, factually correct, nice break down, offered facts that could affect the tax calculation. First time I’ve used Claude.
ChatGPT was good, slightly unclear on a point and the follow-up was not brilliant.
Perplexity is my default and I love it but the results was disappointing.
Claude won, factually correct, nice break down, offered facts that could affect the tax calculation. First time I’ve used Claude.
ChatGPT was good, slightly unclear on a point and the follow-up was not brilliant.
Perplexity is my default and I love it but the results was disappointing.
January 29, 2025 at 2:43 PM
Thoughts:
Claude won, factually correct, nice break down, offered facts that could affect the tax calculation. First time I’ve used Claude.
ChatGPT was good, slightly unclear on a point and the follow-up was not brilliant.
Perplexity is my default and I love it but the results was disappointing.
Claude won, factually correct, nice break down, offered facts that could affect the tax calculation. First time I’ve used Claude.
ChatGPT was good, slightly unclear on a point and the follow-up was not brilliant.
Perplexity is my default and I love it but the results was disappointing.
New annotations:
AllowedValues: Specifies a list of acceptable values for a property.
DeniedValues: Defines values that are not permitted for a property.
Base64String: Validates that a string is Base64 encoded.
Super useful
AllowedValues: Specifies a list of acceptable values for a property.
DeniedValues: Defines values that are not permitted for a property.
Base64String: Validates that a string is Base64 encoded.
Super useful
December 23, 2024 at 1:14 PM
New annotations:
AllowedValues: Specifies a list of acceptable values for a property.
DeniedValues: Defines values that are not permitted for a property.
Base64String: Validates that a string is Base64 encoded.
Super useful
AllowedValues: Specifies a list of acceptable values for a property.
DeniedValues: Defines values that are not permitted for a property.
Base64String: Validates that a string is Base64 encoded.
Super useful
100% and sometimes years later, something clicks and my understanding changes
December 7, 2024 at 1:39 PM
100% and sometimes years later, something clicks and my understanding changes