Category: Ukulele parody songs

Thanks for InterSystems asking me to give a Keynote presentation at the Global Summit in National Harbor, MD this week. I enjoyed the conference, with various speakers from UCSF, the ONC and others discussing the effect of predictive as well as generative AI on the practice of medicine, the health of patients and the burnout of doctors, APP’s, nurses, and all clinical staff in healthcare.

The more we talk about it, the more we can shape our tools into effective interventions. I strongly believe that changing a healthcare organization is only 20% about new technologies or tools, and 80% about relationships, effective communication, and redesigned teamwork.

Performed at the University of Colorado Advanced Practice Providers Symposium, following the talk “I’m Sorry Dave, I’m Afraid I Can’t Do That: Big Data and Generative AI in Healthcare” by CT Lin

I realize Taylor Swift is all the rage for her ERAS tour and her boyfriend Travis Kelce, and yes I am a recent converted Swiftie.

Thanks to Casey Greene for collaborating with ChatGPT to develop and then iteratively improve the lyrics. I added a few human flourishes of my own.

How the sausage was made

Want find out how this particular set of lyrics came about? I’m sure you’re out there experimenting with ChatGPT and similar Generative AI tools.

Here’s a sampling of how we built these lyrics. First an actual chat between Dr. Greene and the Gen-AI, and then my layered edits.

Casey Greene (to ChatGPT): 
Summarize what this verse from Taylor Swift's song Blank Space is saying: 

Nice to meet you, where you been?

I could show you incredible things

Magic, madness, heaven, sin

Saw you there and I thought

"Oh, my God, look at that face

You look like my next mistake

Love's a game, wanna play?" Ay



GPT: 
In this verse from Taylor Swift's song "Blank Space," the narrator expresses a mix of fascination and caution upon meeting someone new. The lyrics "Nice to meet you, where you been? I could show you incredible things" suggest an interest ... (etc)



CG: 
That's a great interpretation! Now I'd like to create a version of this about AI in medicine. It'd keep the same cadence and structure as the original, but the fascination would be a doctor thinking about deploying AI & instead of non-serious as a close, perhaps it'd be nervous about getting things quite right.



GPT: 
Certainly! Here's a version of the verse inspired by "Blank Space" by Taylor Swift, adapted to reflect the theme of a doctor considering the use of AI in medicine:



Nice to meet you, what's your code?

I could show you data, endless rows.

Logic, learning, cure, and glitch,

Saw your specs and I thought,

"Oh, my tech, look at that interface,

You look like my next big leap,

Health's a maze, ready to guide?" Eh.



CG: 
This is pretty great! Two lines from that aren't quite working from me: "Saw your specs and I thought, Oh, my tech, look at that interface," -> Could we rework these?



GPT: 
Of course! Let's rework those lines to better fit the context. Here's a revised version:



Nice to meet you, what's your code?

I could show you data, endless rows.

Logic, learning, cure, and glitch,

Peered into your lines and I mused,

"Ah, this code, such a promising trace,

Seems like a venture into uncharted space,

Health's a maze, ready to guide?" Eh.



CT Lin: 
Additional revisions to include informatics experience into the first stanza. Swift's original in green, CG/GPT words in blue, CTL changes in red. 



Nice to meet you, what's your code?

Data streaming in shiny rows.

Logic, learning, sepsis risk.
Saw your specs and I thought,

You might catch just one more case

OR, you could be my next mistake

Prediction’s hard: what’s your take?

It is a delicate balance to

• keep the rhyme scheme across the stanza
• keep the syllable count per line
• maintain the rhythm and “sing-ability” of each line
• the story within the song: challenge and resolution
• insert coherent electronic health record content
• retain occasional call-backs to the original lyrics

Here’s the entire parody lyric, color coded:
Swift original;   Greene/GPT;    Lin

Nice to meet you, what's your code?

Data streaming in shiny rows.

Logic, learning, sepsis risk.
Saw your specs and I thought,

You might catch just one more case

OR, you could be my next mistake

Prediction’s hard: what’s your take?



AlgoRITHmic BIAS in the AI

Signal to Noise is tough to spy

Hidden dangers with DEI

Workflow hiccups for you and I

So-o hey, lets collaborate

I’m eager to see what we’ll create

Grab your blueprints, take my hand

We can make the old ways better in a weekend



So I test your sensitivity

And I feel like it’s gonna work out

Now I check your specificity

And I worry and I’m full of doubt

Here’s a long list of false positives

You shrug and give a smirk

'Cause you’re a black box, baby

And I’ll check your work.


So it's gonna be forever
Or it's gonna fail and fade,
You can tell me when it's over,
If the leap was worth the upgrade
Got a long list of successes,
These are the lives we’ve saved,
So I’ve got a Blank Space, baby
And I'll write your name.

CMIO’s take?
Its a new world baby, ready or not!

Recorded at the Vanderbilt Annual Otolaryngology Seminar at Vail, Colorado. An invited talk by CT Lin on AI in Healthcare, but wait, does he end with a NEW song?

Yes. Here it is: “ChatGPT” sung to the melody of “Sweet Caroline.”

THANK YOU to Dr. Eben Rosenthal at Vanderbilt University, Sarah Rosenthal at UCSF, and the entire VUMC Otolaryngology Group.

The Clinical Informatics Conference has come to a close in Chicago. I am always gratified by the community of clinical informaticists who come together to share ideas, to make each other better.

I had a chance to participate in 2 panels: Blowing up the Classroom by deconstructing training / Putting the Roadsigns on the Highway. Also: Redesigning the Inbasket, along with colleagues from UCSF, Epic and MedStar.

Here’s my contribution to the fun; an updated version of Epic Man.