Google Home is the little speaker/virtual assistant thing put out last year by Google. It's a similar idea to Amazon's Echo and Apple's new HomePod. You can use it to play music, make phone calls, answer questions, and control different things in your house (lighting, thermostat, tv) so long as those things are "smart" (so it won't work on your wife, right guys? hahahahahaha, that dumb old bitch!).
I have to be honest, when reader L.B. wrote in to suggest maybe utilizing the Google Home as part of a trick, I thought it might be mildly amusing but not really all that great. I thought that both about the concept of using it in a magic trick, and about the Google Home unit itself. But, after having a unit for a few days and playing around with it, I'm actually surprisingly happy both with the unit itself and the responses I've received when using it for a magic trick.
When it comes to technology and the Product Adoption Curve, I'd put myself as a "late early adopter." So somewhere near the right side of that green area. I'm not a super tech-savvy guy, but I usually have a pretty good eye for what technology is going to stick around and what's not. (I remember when my friend was buying his movies on LaserDisc in the late 90s and I was like, "You really think this is a technology with a shelf-life greater than 6 months? You don't see a slight issue with this as a storage medium?")
Google Home (as well as the other similar products on the market) didn't seem like something I needed to have, but at the same time I knew it was probably just a matter of time before I got one and I'd end up finding a number of uses for it.
So, when L.B. suggested the idea I went and bought one. The Google Home Minis are $50. So it's not a super-big investment.
Let's start with the basic, somewhat obvious, idea of how to use this as part of a magic.
You can configure the Google Home to respond in a certain way to whatever you say. So if you say, "Hey Google, what card did she pick." You can have it respond to that with, "She picked the three of spades," or whatever. So then you would force the three of spades, ask google what she picked, and that would be the trick.
That's fine, and I'm sure it would go over as well as any standard card revelation. But I wanted to build on the idea a little and add a few more elements to embellish the presentation and also hide the idea that there is something pre-set.
Here's how it looks.
My friend Sara comes over, notices the Google Home and asks me if I like it and what I'm using it for.
"I didn't think I was going to be such a big fan of it, but I'm actually really glad I bought it. I might get another one for my bedroom." Then I demonstrate some of its features, like how it can control my lights and music.
"It's a little weird though to think that's it's always listening. And they say it's not constantly recording, but who the fuck really knows. And it just does some... strange things sometimes. Like it will play a song I had stuck in my head when I ask it to play music, even though I don't specifically name that song out loud. And sometimes it starts answering questions before I've completely verbalized them, like it knows what I'm about to say."
"I know it sounds like horseshit, but I'm not kidding. It kind of makes sense. It picks up audio waves, so why not brain waves? That seems like it could be possible, at least in some rudimentary form. I'll show you. Think of something."
Sara settles on something in her mind.
"Hey Google, what is she thinking of?" I ask.
I can't tell. There are too many thoughts coming in at once.
"Okay. Yeah, it helps to have one particular idea that exists outside your mind to concentrate on. So... uhm... grab a couple books from the bookshelf. Two that have a similar number of pages."
I now run through the process of the Hoy Book Test to have her settle on a word to think of. You could also do some kind of peek of a word they wrote down. Or, less ideally, have them freely choose a card and figure out what it is in some manner. My preferred usage is with the Hoy Book Test. Nothing is written down, it's not playing cards, it feels very random because any two books could have been chosen.
So my friend is now thinking of a word. I ask again, "Hey Google, what is she thinking of?"
And again the reply comes: I can't tell. There are too many thoughts coming in at once.
"Dammit. I swear this works. I'm not crazy," I say. "Wait, I know."
I go to my kitchen and come back a few seconds later with a large square of tin foil and start shaping it around my head. "This will help block out my thoughts so she can home in on yours."
[There is very little in my magic performances that is "scripted" in the traditional sense. But saying "I'm not crazy," right before going and putting on a tinfoil hat is one of those beats I always intend to hit.]
Once I have my tinfoil hat on and I'm looking like a complete dork I say a final time, "Hey Google, what is she thinking of?"
This time Google replies, "Okay. It's coming through clearer now. I think there is an S sound in there somewhere. No. Wait. I know what it is. She's thinking of the word: history."
She is and she freaks out.
"It knows everything!" I say. "Pick something up off the table." She picks up a remote control.
"Hey Google, what is she holding?" I ask.
"She's holding a remote control," Google Home says.
Sara tosses the remote aside like it's a cursed object.
"This thing is too scary," I say, and I chuck it out the window.
[No, I haven't thrown it out the window. But for $50, it might be worth it. What I really do is I unplug it like I'm a little freaked, and thus putting an end to asking the thing more questions.]
I'm not going to dwell too much on the technical details other than to say it's very easy to set this up.
So, we want to Google Home to reply a particular way when we say something. And then we want to be able to change that reply to something else in just a few seconds. This is how we're going to add a little shade to the method. Instead of just being a card force and a little robot that always replies the same way to a certain question. We're going to have it respond different ways to the same question.
Getting Google Home to Respond What You Want It To
Method 1 (Don't Use) - In the Google Home app there's something called "shortcuts" where you supposedly can get your GH to reply whatever you want to a particular input. I tried it and it didn't work. So I don't recommend it.
Method 2 IFTTT - IFTTT stands for If This Then That and it's just an app that connects two apps/programs/smart objects in your house so that when one thing happens it triggers something else. I'm not even going to get in the potential uses for this, but I have a feeling there are all sorts of magic related ones that you could find.
So you get the IFTTT app, which is free, and it connects to your google home. And you create an "applet" which is just a simple conditional statement. "If X happens, then make Y happen." Again, if it sounds complicated or techy in any way, it's not. The image to the left shows you the extent of the "programming" required. And this video walks you through the whole process.
So that "applet" gets saved and now whenever you say "What is she thinking?" you will get the "I can't tell..." response.
Now what you do is go in and edit that applet so it has the following line in the What do you want the Assistant to say in response? field
Okay. It's coming through clearer now. I think there is an S sound in there somewhere. No. Wait. I know what it is. She's thinking of the word
So this is everything but the actual word itself. Don't worry about the potential non-hit in the middle. If there is an S sound in the word, it's a minor hit before the actual reveal. And the "No. Wait," suggests "No, I'm not just getting a letter, I'm getting the full word." And if there's not an S sound then the "No. Wait," makes it seem like it's correcting itself. Either way the statement will make sense. And it makes the revelation bigger so that it's not just the word itself.
Keep in mind: you don't save this new message yet. You just turn off your phone at this point and set it somewhere in your kitchen near the tin foil.
This is all pretty straightforward from here.
You try to have it read her mind in a general sense, and you get the "I can't tell" response.
You say, "Ah, we need you focusing on a single thought." Then you go through whatever process (Hoy book test, peek, card selection) to get them to narrow it down to a single thought. You ask Google Home again what your friend is thinking but you still get the "can't tell" message.
You realize you need to block your own thoughts. Go to the kitchen, turn on your phone and type in the one word at the end of the phrase you already entered then hit save. Put the phone back and grab a piece of tin-foil you've already removed from the roll. This shouldn't take long. Maybe 10 seconds, 15 at most, to type one word. That doesn't seem like an unusually long period of time to get some foil.
Return. Make yourself a hat. Ask again what she's thinking. React. But remember, you're not supposed to know the word so get her to verify it first.
But what about that bit at the end where I tell her to pick up something off the table and the GH tells her what it is? This is simply a matter of only have a few items on the table and then having a key phrase for each one. "What did she pick up?" "What is she holding?" "What's in her hand?" Etc.
You might wonder why that bit is even there. In a theatrical magic presentation you would never follow up the revelation of seemingly any word in the English language with a 1 in 5 revelation. Certainly not within the same broad "effect." You might feel like the correct structure would be, "Let's see if Google Home can guess which object you picked up." And then follow that up with, "Okay, let's up the ante and have you think of any word in the English language." But the reason I prefer to do it in the opposite order is because it feels less like I'm showing concern for dramatic structure. I don't want it to feel like a two phase "routine." I want the "what item is she holding" bit to come off kind of as a throwaway afterthought. I get into it before the reaction from the first effect has died down. What I really want to do is confuse the issue. Since there is a bit of a build-up before the first reveal, they might start formulating some hypotheses about what happened and when it happened. By slapping on another similar effect—but with a different method—at the end, I think that helps obscure the method used in the main part of the trick.
1. I don't know if there's a limit to how many IFTTT statements you can create, but assuming there's not, you could theoretically create an "if this" statement for every card in the deck. Then you could have a card freely chosen (say from a stacked and/or marked deck), cue it to Google Home in your question and have it name the correct card in a very fair way. You could use a crib. Or you could do one of those things where the first word of the sentence cues the value and the last word cues the suit. Then you only have to remember 17 things (suits and values) rather than 52 different phrases for each card.
2. Most laypeople don't realize how few items there are that people draw when asked to draw something. You could have cues for the top 25 most drawn items. Then you could peek a drawing, or openly just watch the person draw whatever they want and then cue the correct drawing with your question. "Hey Google. What did he just draw?" He drew a car. "Isn't that crazy?" you say. "I found out how it works. It actually listens to the marker strokes while you make the drawing. That's how crazy sensitive this thing is. What will they think of next"
3. You can, of course, just use these same tools to do dumb stuff like this.