
Currently I am sharing my knowledge with the Power Platform, with PowerApps and Power Automate. With over 8 years of experience, I have been learning SharePoint and SharePoint Online
In a recent YouTube video titled "Claude Memories and Chrome Extension | How I'm Switching From ChatGPT," creator Andrew Hess - MySPQuestions walks viewers through his decision to move from ChatGPT to Claude for everyday AI work. He emphasizes experimentation over loyalty, showing how he imports memories from other models and configures features in Claude. Also, he tests a browser extension in an isolated environment to limit security risk while evaluating functionality. Overall, the video presents hands-on steps and candid tradeoffs for users considering similar moves.
Hess demonstrates how to export accumulated context from one model and import it into Claude, explaining that the process can save users time by preserving preferences, project notes, and conversational history. He shows copying an extraction prompt into the original assistant, pasting the resulting text into Claude’s memory settings, and then confirming that the data merged rather than overwrote existing entries. As a result, users can restore months of context quickly, which matters when projects span multiple sessions and tools.
However, the video highlights practical caveats: imported memories may need cleaning and editing to avoid irrelevant or duplicate details, and automatic merging can introduce inaccuracies if items conflict. Hess notes that Claude separates memories by project and offers controls to edit or delete entries, but he cautions that free and paid plans may handle persistence and automation differently. Thus, while import tools reduce setup friction, they require active curation to keep AI responses accurate and useful.
One notable section of the video focuses on the Claude Chrome Extension, which Hess installs in Google Chrome Canary to test features without exposing his primary browser profile. He explains that browser extensions can access page data, so running experimental extensions in a separate build reduces the risk that a malfunctioning add-on could leak sensitive information. In short, his isolation strategy balances the convenience of a desktop-integrated assistant with prudent security practices.
During the demo, Hess uses the extension to prototype content tasks and to have Claude create and design SharePoint artifacts, noting both successes and limitations. He treats the extension as a beta feature that speeds certain workflows but increases the attack surface if not confined, especially on work machines. Therefore, his recommendation is to test new extensions in controlled environments and monitor permissions closely before full adoption.
The video carefully explores the tradeoffs between convenience and security when moving conversational context across AI platforms. On one hand, importing memories makes an assistant immediately more productive and consistent, which reduces manual re-briefing and saves time on recurring tasks. On the other hand, transferring sensitive workplace details requires strong governance and personal discipline to prevent unintended exposure, particularly with browser extensions or third-party import scripts.
Moreover, Hess discusses portability versus lock-in: while import tools reduce the pain of switching, differences in how models interpret and use memory can lead to misaligned behavior, requiring continued adjustment. For enterprises, additional concerns include compliance, retention policies, and who can access or delete recorded memories, which complicates scaling these features across teams. Consequently, organizations must weigh productivity gains against the overhead of policy, access controls, and auditing.
Hess ends the video with practical advice that balances enthusiasm and caution, urging users to experiment but to keep tight control over what they import and enable. He recommends testing in isolated browsers or accounts, reviewing imported entries for relevance, and using edit/delete controls to avoid context drift and privacy issues. By doing so, users can enjoy the benefits of persistent memory while minimizing surprises in automated responses.
Ultimately, the video frames the move from ChatGPT to Claude as a pragmatic choice rather than a permanent allegiance, highlighting that tool selection should follow current needs, features, and security posture. As AI memory features evolve, Hess suggests a hybrid approach: adopt new conveniences for routine tasks while enforcing clear governance in sensitive workflows. This balanced stance helps readers judge whether to follow his path or to adapt it for their own environments.
Claude AI memories, Claude Chrome extension, switch from ChatGPT to Claude, Claude vs ChatGPT, Claude memory feature, Claude browser extension tutorial, AI assistant migration guide, ChatGPT alternative Claude