As MCP adoption grows, teams are moving quickly to secure how LLMs connect to tools and data. We put together 7 best practices to help you:
- Lock down supply chains
- Enforce least privilege
- Add human oversight
Currently google has access to data about where we navigate via Chrome. But something like Chat or Claude could have pretty amazing data about where we go via mcp and mcp-ui
@liadyosef Thats a true point. I thinking that context will be very interesting esp. if convo and data is moving between different services. The context of where and what you've been doing will most certainly lead to better experiences.
@liadyosef Point 2 is defo spot on. The fact that the de facto in the situation is language and conversational. Without any context or visual shifting. A flowing dialog.
In the 1940s, engineers built machines for war.
In the 1950s, those same minds returned and applied their skills to civilian manufacturing building machines that could produce at scale.
The result? More products than ever before.
AI code generators are lowering the cost of software and increasing the volume we can produce. Just like post-WWII machines scaled manufacturing, we’re now scaling software.
Next wave?
Smart product minds.
Storytellers who can sell.
#AI#SoftwareEconomy#FutureOfWork
@mustafasuleyman Adaptive UIs are exciting—users could have journeys truly tailored to their needs. But how do we ensure accuracy with non-deterministic tech? What kind of testing ensures relevance? Curious what others think. Is this a real concern or just growing pains?