Rick Dronkers (@rickdronkers) 's Twitter Profile
Rick Dronkers

@rickdronkers

Usually tweet about the intersection of technology, data, marketing, privacy & business. - 👨‍💻 Founder of @data_to_value

ID: 26812072

linkhttps://datatovalue.com calendar_today26-03-2009 18:27:07

2,2K Tweet

2,2K Followers

288 Following

Rick Dronkers (@rickdronkers) 's Twitter Profile Photo

AI tip: do a couple of personality tests. (Myers Briggs, Talent Management Assessment, Kolbe, CliftonStrenghts, PRINT, etc) - throw them all in an LLM, ask it to write a detailed personality guide about yourself in markdown for LLM consumption. Add that guide to all your agents.

Rick Dronkers (@rickdronkers) 's Twitter Profile Photo

Having LLM in between automation steps is ideal. I have a Zapier stream that gets emails I flag into a todo app, it tags the task, estimates time, creates action points, writes the task title. Etc. This is so useful. Should be native Akiflow feature - basically create an

Rick Dronkers (@rickdronkers) 's Twitter Profile Photo

Another Monday morning meeting where we pretend those marketing dashboards actually create value. Let's be honest... 👇 Here's my tip: Resist the temptation to go into reactive firefighting mode and get fixated on what's already happened based on what you see in the numbers...

Another Monday morning meeting where we pretend those marketing dashboards actually create value. Let's be honest... 👇

Here's my tip: Resist the temptation to go into reactive firefighting mode and get fixated on what's already happened based on what you see in the numbers...
Rick Dronkers (@rickdronkers) 's Twitter Profile Photo

I see you thinking: Whats the moat of ELT tools (eg FiveTran) if Model Context Protocol (MCP) works really well? Short term: Reliability, reprocessing failed exports, having the ability to point a finger when it breaks. Longer term? 🤷‍♂️

Rick Dronkers (@rickdronkers) 's Twitter Profile Photo

Solving my problems with AI - I scan paper to PDF's, auto extract the text, feed it into AI, have the AI describe things like: recommended pdf name, what its about, action items (e.g. pay bill), due dates, etc. Feed that all into my archive for backup and todolist for action. ❤️

Rick Dronkers (@rickdronkers) 's Twitter Profile Photo

I keep getting these ads for cheap handheld game emulators that play all old games. It’s very tempting. If you don’t hear for me for 3 months you know what happened.

Rick Dronkers (@rickdronkers) 's Twitter Profile Photo

Suddenly you see 2 year subscriptions in the dataset being sold for 2026-2028 and all your assumptions about how everything worked are back to square one...

Rick Dronkers (@rickdronkers) 's Twitter Profile Photo

Have seen this take more frequently and gone down this path myself💀 A lot of people are going to fuck around and find out the difference between a vibe coded MVP and a production ready solution. 😅

Rick Dronkers (@rickdronkers) 's Twitter Profile Photo

Some of the best performing performance marketing teams I’ve seen over the last ten years had heavy finance involvement or even reported to finance directly 🤔

Rick Dronkers (@rickdronkers) 's Twitter Profile Photo

Im vibe coding something that has been in my head for quite a while. Not sure if im gonna make it to the end of the debugging tunnel though 😂🤞

Rick Dronkers (@rickdronkers) 's Twitter Profile Photo

Stuck vibe coding? Ask Cursor to write a detailed prompt for a perplexity deep research query about the problem. Then feed that back into Cursor.