The $15,000 A.I. From 1983
“Hello!” Hello. I thought this was gonna be awesome. But this proto-Siri-Alexa virtual assistant nearly destroyed me. I was flipping through the March 1987 issue of Popular Science and saw THIS image of a cartoon butler genie-bursting from a small black box, with octopus arms juggling all sorts of different tech.
It’s the Mastervoice “Butler in a Box,” and it was billed as “The world’s first artificial intelligent environmental control system for the home.” It can also call up to 16 different people, provide home security, and do it all in 4 different languages. It’s supposed to make everything around your home simpler, but it ends up generating more questions than answers. How does it work without wifi, bluetooth,
or the internet? If this was such a breakthrough technology, why did it fade into total obscurity? Who invented this – and why is he a professional magician who owned a 7-11 at the age of 21 and had a world champion Frisbee-catching dog and also made a cannon to feed fish from a submarine – and why does his YouTube channel have videos of his involvement in a MURDER TRIAL? You’ll get a glimpse of its one-of-a-kind inventor, Gus Searcy – and you’ll also see if I can get this thing to work. That was my first challenge… is it even possible to acquire a Butler in a Box that functions today? Amazingly I found one on eBay that appeared to be in mint condition – including the one critical part that’s missing almost all the time and totally bricks the device. I cracked open the decades-old thick styrofoam shell and found a nearly-indecipherable 134-page Owner’s Guide, the 3” x 9” x 11” Butler in a Box – which is roughly the size of a PS4 – something called an “appliance module,” a lamp minder, a training cassette and a standard RJ-45 phone jack and cord. I don’t have a landline… and I don’t even have a cassette player. At least everyone had both of those in the 1980s – but the dream of technology that magically responds to your voice commands goes back centuries. Ali Baba triggered the door-opening mechanism of the 40 Thieves’ den just by saying, “Open Sesame” – that folk tale is at least 300 years, and probably a lot more.
As the industrial world worked to apply Newtonian physics to technology, the dreamers still dreamed. In 1911, Hugo Gernsback detailed the “Luminor” system in his science fiction novel “Ralph 124C 41+” – Luminor used voice commands to turn on and adjust the intensity of lighting, just like Alexa-controlled WiFi bulbs do now. By the 1960s, the Jetsons had a robot maid that interacted with the whole family, Captain Kirk interfaced seamlessly with the USS Enterprise, and a Heuristically programmed ALgorithmic computer terrorized Dr. David Bowman – HAL-9000.
But even into the 80s, the prospect of a voice-operated assistant still seemed like magic – and when magician Gus Searcy kept dealing with people asking why he could pull a rabbit out of a hat but couldn’t turn on the lights without getting up, he decided to invent a device that could. Or… that’s his official story, which also includes citing the 1977 sci-fi horror film “Demon Seed,” in which a fully-automated home is controlled by an artificial intelligence that forcefully impregnates its owner, as a direct inspiration for Butler in a Box. If that seems unconventional… so is Gus. His backstory is wild: he claims that at age 16, his parents signed over power of attorney to him so that he could play the stock market, and he was earning six figures. At 19 he owned a condo in California, and at 21 he became the
youngest owner of a 7-11 franchise in the world. Gus is still alive – he invented the ShowerMi$er, a water-saving solution for RVs, wrote a motivational book, and worked for a legal services subscription company – that was where he became deeply involved in a murder trial. So what’s real? What’s not? I don’t know? But the Butler in a Box does have a listing on the Smithsonian website. So I got a tape player – and then I had to search “cell phone landline adapter” to find this thing that acts as a bluetooth go-between for your smartphone and an old-style home phone connection. It’s easy to connect stuff now – but how did a voice-activated home automation system work a decade before the world wide web went public, 14 years before WiFi, and 16 years before Bluetooth was even available? Carrier current and the X10 protocol for electronic communication.
Basically, a microprocessor in the Butler sends a series of pulses through the electrical wiring of the house, which are then read by modules that lamps and appliances plug into, and those are each assigned a specific house code and unit code. I’ve got a lamp here, and this is my goal for the Butler – I want to be able to turn this lamp on and off just with my voice. So I plug the lamp into a lamp module, plug that into the socket, manually turn the lamp on and leave it on, then set the house code dial to something like A and the unit code dial to, say, 1. Then you log the information that device 1 is a lamp assigned to A1 in the setup form in the back of the manual as a reminder before you spend the required 25 minutes training the Butler to recognize your voice, with the help of the included training cassette tape, which literally plays the Beverly Hills Cop theme song “Axel F” in the background.
Which I SERIOUSLY doubt Mastervoice licensed. And then you designate a specific phrase like “LAMP” so the Butler knows you’re referring to Device 1 which is set to lamp module A1. That’s the easy part. I plugged in my Butler In A Box and turned it on, which was a huge win – there was no guarantee that an obscure 40 year old piece of tech was going to power up. There’s a 16-digit alphanumeric vacuum fluorescent display, and that works great. So then I just had to program it. And to do that, you need that critical bit of information that for almost all of the surviving Butlers in a Box has been lost: the four digit alphanumeric pin. It’s meant
to be a security measure – it’s located on the inside cover of the owner’s guide and you cannot change it. If you lose it? No pin, no butler – the box is locked forever. CelGenStudios on Youtube has been trying to reverse engineer the pin to his Butler In A Box for 8 years. But I’ve got my pin: #S93K. So how do I input it? If you guessed “talk to it,” because this is a cutting edge voice-activated device, you guessed wrong. It turns out there’s a hidden pressure-sensitive keyboard flap that folds out of the front bottom of the Butler. But it’s actually laid out like an old alpha-numeric keypad, not a QWERTY keyboard – and here are the verbatim instructions on how to input that pin to unlock the device: “To enter information requiring letters of the alphabet, you would use the Alpha Select mode of operation… The last three keys, top row at right, are Alpha 1, Alpha 2, and Alpha 3 (see figure 6-2). These three keys have two functions each. First,
in the Alpha Select mode of operation, these keys will permit you to choose which letter on the number keys 1 through 0 will be entered into your unit. The three small squares above the thick line correspond to the sequence of alphabet letters above the thick line on the number keys. Example: Alpha Select Key 1, with the solid square in the first of three possible positions, allows you to select the letter A, the first letter selection on Key 2. Alpha Select Key 3, with the solid square in the third of three possible positions, allows you to select the letter C, the third letter selection on Key 2, and so on.”
AND SO ON, INDEED. I’m pretty used to translating poorly-written instructions with cheap Amazon electronics, but how was anyone but the most hardcore tech enthusiast expected to navigate 134 pages of this decades before average people were even used to using home computers? “And that in the future, most people will have a computer system right next to their TV.” It’s kafkaesque, and the whole thing seriously tested my patience before I even discovered what was broken on this thing – but before we do that – you need to know what this cost.
Gus Searcy started talking in public about the Butler in 1983, and he said the ‘commercial applications of the Butler would be staggering’ – he also recognized that a voice-activated home control device could be seriously useful for the elderly or someone with a disability. He was right about that part, but he was wrong about the price tag. In 1983, his target price was about $300 – half of the initial price of the Commodore 64 and twice the price of the first NES a couple years later. In 1984, he said $500 – this 1985 ad listed it for $1,195. By the time Jim Schefter tested the Butler in 1987 and wrote about it in Popular Science, it was $1,495.
We’re not done. A 1990 episode of The Home Show announced it was $3,000, which became $4,000 by the year 2000 – adjusted for inflation, it started out as a $940.94 device that became a $7,268.86 device… and that’s including only two modules, one for a lamp and one for an appliance.
Every single device in your home required its own module, so if you maxed out your whole house with 62 more modules at $36 each – for your coffee makers, TVs, radios, VCRs, alarms, electronic curtains, garage door, sprinklers, thermostat, fish tanks and pool pumps, whatever – you’re nearing $10,000 in today’s dollars… and you also needed microphones and speakers in every single room, and an electrician to install those and whatever in-wall modules. For this Butler to work on more than just a lamp and a TV in your living room, it was… $15,000? And one of the most amazing things to me as I was researching Butler In A Box is that for more than 20 years it was announced over and over again as a magical, futuristic, life-changing technology culminating with a feature on Extreme Home Makeovers in 2005. Alexander? Did you call? laughter HOW? How did it stay futuristic for two decades while pretty much every other piece of tech got a lot better and a lot cheaper!? Well, speech and voice recognition was actually a really hard nut to crack – in 1952 Bell Laboratories’ “AUDREY” was able to recognize each of the digits 0 through 9, but with only 90% accuracy – and the only guy who could use it was its inventor. By the 1970s, the Defense Advanced
Research Projects Agency – DARPA – was funding research to improve on the limitations of the two basic recognition models at the time: trying to match spoken words to pre-recorded waveforms, or using a series of algorithms to create a ‘score’ that matched with a pretty good guess at the word. When Gus Searcy was touting his miracle Butler, the rest of the industry was working with Hidden Markov Models, like the earliest Dragon Dictate software – it was another 20 years before DARPA’s research broke enough ground for Siri to translate your voice into a code, which was then matched with patterns and keywords, all in different tones and accents, so that it can determine and carry out a specific action, and continually collect and process information to get better at all of it over time. Okay. Wait. So if devices at the time struggled to process ONE language, how was the Butler in a
Box able to understand four separate languages? Well, it didn’t. I mean, it did, but it didn’t. The Butler could handle voice commands from four separate users, but every user had to program it individually. So if I’ve programmed it for me, my wife can’t say, “Turn Lamp On” and have it work unless she has programmed it for her voice, too, with exactly the same phrases that I used. Except for naming the Butler. She would have to name it something else. So if I named “Jeeves” she would have to name it “Godfrey” or something else, “Beyonce” it didn’t matter. Because the Butler in the Box isn’t Siri or Alexa. It doesn’t know anything, and it can’t really learn, either.
But it can respond to four distinctly-programmed sets of sounds. That can be Japanese, or Klingon, or The Black Speech of Mordor. You could even make up your own language just to use with the Butler. It can’t tell the difference, since it doesn’t know anything – that technically means it can know everything. Or at least 4 of everything.
Alright, let me tell you what’s broken. So I plugged it in and got this awful crackling sound – that was supposed to be the Butler saying “Hello.” So alright, the speaker is broken – I could bypass that by plugging in an external set of old computer speakers. And I got… THE SAME BROKEN CRACKLE SOUND. Whatever. I moved on with the setup, did the microphone test, and got NO response at all and realized that Jeeves here may be dead. I took it down to my Dad’s shop, he knows audio electronics and circuitry, so I thought
maybe he could take it apart and solder some loose connections or something. This was the something. He turned it over and showed me the volume slider for the speaker, which was turned all the way to 0. So we raised the volume, plugged it in, and heard “HELLO.” They call that “USER ERROR.” I’m user. There was a microphone sensitivity slider set to 0, too – so I assumed that would fix my mic problem. The Butler has a visual indicator that shows asterisks to display how well your mic is working. And my mic wasn’t working. I got zero asterisks. The internal microphone WAS broken. I ordered a karaoke mic with a jack
adapter and that kinda worked – then stopped working. At that point we brought in the big guns. My dad called his electronics wizard friend, Neal, who took it apart, cleaned up the mic jack and sliding potentiometers, fixed a cracked board and saved the day. SIDENOTE: While Neal had it in his shop, he tested its power usage and was blown away. It turns out this draws a constant 23 watts in standby. To
put that in perspective it’s up to 11 ½ times an Alexa or the equivalent of 110 hours of Playstation gaming per month. Just sitting there. Racking up your electric bill. Now. I read in a few places that setting up the voice control was going to take SEVERAL hours… but I actually trained my Butler to do the basics in 25 minutes. I taught
it to recognize my voice by speaking a series of 10 specific words, like Air Conditioner, Mood Lighting, and Airplane. You can then program up to 10 commands, 18 commands to call 16 different people, set 4 alarm words, and train 32 device words. But I just did LAMP.
And like I said you can the Butler anything you want – that name is the “Entry Word” to wake up the Butler. So, to get it to work, it’s your Butler’s name, then tying a specific device to the command you’re about to give – like, Jeeves, Lamp, On. Then you just have to test whether it or not all that actually works. If it does, you press YES. If your voice commands don’t work, you press NO, and then re-train it until they do.
And multiply most of that whole process by THIRTY-TWO to max out the Butler’s devices. And make that 128 times to set up all 4 different users. And then it can all just… disappear. Because I haven’t told you the worst part, the Butler In A Box kryptonite… What happens if the power goes out is the most depressing, defeating aspect of this entire device – really worse than the PIN. If there’s a power outage, the backup batteries will preserve the Butler’s memory – all the device programming, all the voice programming – for 3 hours. After that? It’s gone. It’s all gone.
It was possible to change the battery clip cable to accept a 9-volt battery that lasted for 10 hours, and it was possible to back up the programming memory – NONE of the voice training – to a home computer using an RS-232 port, which was even more complicated than any of the setup. The reality is if a storm knocked out power for more than a few hours, your Butler – Alfred, Jeeves, Godfrey, Lurch, Geoffrey, whatever you decided to name him – was dead. In that 1987 article, Jim Schefter concluded that “the Butler in a Box is for the gadget freak, the computer buff, and the handicapped,” noting that Gus Searcy himself compares the device to “a slightly-deaf 80 year old man.” Jim suggested that it wasn’t uncommon to have to repeat yourself 2 or 3 times. So… did the Butler in a Box work? Yeah. yeah it really did. “You again? I’m home. Alright wait a minute.”
Y’know we’re used to driving in comfort on the highway, and the Butler in a Box is a horse and buggy – but it’ll get you there. Early technology is just usually really messy – sometimes it’s not good to be first. In 1937, Jan Romein called the phenomenon “Wet van de remmende voorsprong” – economists call it the “First Mover Disadvantage.” – because a promising product or idea tends to become a canary in the coalmine for competitors, who then solve those problems and achieve success. They’re free riders as you take all the damage from blazing the trail. It’s why you might have an Amazon Alexa in your room, or Apple’s Siri in your pocket – and why you definitely don’t have a Mastervoice Butler in a Box in your living room. You’ve never even heard of this. Not only does it not have a wikipedia page,
it’s not a footnote on the pages for Home Automation or Virtual Assistant. This isn’t a tech video. It’s a ghost story. There’s a reason why so many of the people who tend to do things first are eccentric – they’re the ones who are willing to take the risks, or sometimes they just don’t care about the consequences. The more rational people wait
to improve on a product or to address the concerns of its early adopters. But we needed gas street lamps to get to cheaper, safer, better electric street lamps. We needed an RS-232 port to get to USB-C. And we needed Butler in a Box to get to the intuitive voice assistants that, honestly, don’t even impress us anymore – we’ve just come to expect them.
Because we need the eccentrics, and we need their technology, even if it turns out to be flawed. Mr. Belvedere? May I help you? Lamp. Okay. On. As you wish! See you in the future.
2024-03-09 03:31
