Search "phone" with CONTAINSSTRING, and you also get headphones and microphones.
Power BI semantic models can now build a full-text index used by two DAX functions that look for the correct meaning of words: TEXTCONTAINS and TEXTSIMILARITY.
How to enable it, and the traps we found:
sqlbi.com/articles/introduci…#dax#powerbi
70 years of AI research. The real heroes of the story never read a single paper, but they had excellent reflexes and a strong opinion on frame rates.
marcorusso.com/writing/thank…
On Thursday, we found out which blocks are load-bearing.
The funny thing is, I picked this picture for today over a week ago.
Is this version correct? Would you add something else?
There are no longer any excuses for a Minimum Viable Product. From now on, an MVP should be a Maximum Valuable Product, as the SQLBI Whiteboard story teaches us.
marcorusso.com/writing/maxim…
Draw two shapes, connect them with an arrow, then move a shape. The arrow stays attached.
No redrawing the diagram halfway through an explanation. 12 seconds below.
SQLBI Whiteboard, free and open source: whiteboard.sqlbi.com#powerbi#dax
Yesterday's article on optimizing DAX with AI was supposed to come with a video, and the video was missing at launch. Now it's online: the same test, human vs AI vs human plus AI, in video form.
youtu.be/F52wwF9IiKk#dax#powerbi#ai
Live today: Inside DAXBench v2, with @maxanatsko.
16:00 CET / 15:00 UTC / 10:00 ET / 07:00 PT
Maxim and I go through the results of DAX Benchmark v2 and answer your questions live.
x.com/i/broadcasts/1qKDzWkPm…
Leaderboard: daxbench(dot)com
#DAX#PowerBI
We optimized a DAX measure ourselves and asked AI to do the same. Neither version was faster on every query.
Then AI refined our version: the slowest query went from 4,396 ms to 1,538 ms.
It also wrote more equivalence tests than we ever had the patience to write.
sqlbi.com/articles/optimizin…#dax#powerbi#ai
DAX with AI: Scenarios is complete!
The video course includes ten business scenarios: currency conversion, comparable-store sales, budget vs. forecast, shared-cost allocation, inventory from sparse snapshots, financial statements, workforce analytics, and more. The training also includes seven "concept" lessons that are shared across several scenarios.
We announced it in September for November 30, and we're proud to release it four weeks early.
If you already purchased the video course, log in to your SQLBI account: the whole training is ready for you!
You don't need to watch it from start to finish (but you can if you want to): Pick the scenario closest to your next project, check the concepts it lists, and start there.
The details are in my blog post: sqlbi.com/blog/marco/2026/11…
If you haven't enrolled yet, there's a launch offer on the course and on the bundle: sqlbi.com/p/dax-with-ai-scen…
Over the last few months, I've started writing essays that aren't strictly technical; they are more about how we work, think, and build things.
I also created a website that collects everything I publish, because with AI, I no longer have an excuse to postpone it: marcorusso.com
Here are the essays so far:
- The orchestra conductor
- Code as a hobby
- No SQL, no analyst
- Function first
If you'd like new essays by email, subscribe on Substack: marcorus.substack.com
This is still the place for the conversation: your feedback is very welcome!