Picard algorithm achieves State-of-the-Art performance for Text-to-SQL

Text to SQL has been a problem many people have tried to address. I'm highly skeptical of a solution that is devoid of a human explaining the schema of a database, or intent, or even metadata.

Don't get me wrong, I'm sure Picard is making good progress, but I'd be surprised if "State-of-the-Art" is actually "useful"

SELECT * FROM Lights WHERE Qty = 4;

ENGAGE