@levelsio You should listen to the Acquired podcast. They have episodes on Ferrari and Rolex (among others) which are super fascinating because they give tons of insight into the historical, cultural relevance of these businesses in addition to finances. I think you’ll dig it
This might be helpful. The idea behind this wasn’t to save every single message back and forth, but rather the valuable learnings and memories that can be extracted from those conversations. The specific content that should be saved can vary depending on the project or your particular interest. The repository is open-source, so you can easily modify it if you prefer to save every single message back and forth. https://t.co/VJDFcp35Ra
If you use more than one AI tool, you're re-explaining yourself every single conversation. Your coding agent doesn't know what your mobile assistant discussed yesterday. Every session starts cold.
Engram fixes this — it's an open-source memory layer that sits between all your AI tools via MCP. Store a fact in one, search it from any other. Three search methods run in parallel, results in under a second, costs about $5/month to host.
The full breakdown covers the search pipeline, the architecture, and how to get your agents to use memory without being asked.
If you use more than one AI tool, you're re-explaining yourself every single conversation. Your coding agent doesn't know what your mobile assistant discussed yesterday. Every session starts cold.
Engram fixes this — it's an open-source memory layer that sits between all your AI tools via MCP. Store a fact in one, search it from any other. Three search methods run in parallel, results in under a second, costs about $5/month to host.
The full breakdown covers the search pipeline, the architecture, and how to get your agents to use memory without being asked.
If you use more than one AI tool, you're re-explaining yourself every single conversation. Your coding agent doesn't know what your mobile assistant discussed yesterday. Every session starts cold.
Engram fixes this — it's an open-source memory layer that sits between all your AI tools via MCP. Store a fact in one, search it from any other. Three search methods run in parallel, results in under a second, costs about $5/month to host.
The full breakdown covers the search pipeline, the architecture, and how to get your agents to use memory without being asked.
I can be if you hook it up to https://t.co/6iFeU4HdAq MCP :)
I better data set of every product on whop marketplace and I track this every every single. So you can see which products are going up or down and users in revenue and then using the MCP ask Claude to just you know build you something in that category or just copy the product as it is