5/36
Powered by Foleon

Create the content your audience craves.

Find out more
  • Pages
01 Cover
02 MSI
03 Table of Contents | May-June 2023
04 Quantra
05 Spall
06 Capsule | May-June 2023
07 Advertising Index | May-June 2023
08 Cosentino | Pietra Kode by Dekton
09 News Highlights 1 | Prodim
10 News Highlights 2
11 Cambria
12 Installation Crews | Water Treatment Solutions
13 Installation Crews - How Many?
14 Challenges of Installation
15 Xiamen Show
16 Fabricator Focus | Hard-Surface Report
17 Fabricator Focus - AMC Countertops
18 BB Industries | SRG Preview May-June 2023
19 Hard-Surfaces Cinema | May-June 2023
20 Natural Stone Institute | compare
21 The Message | Keeping Your Employees Safe
22 Cersaie 2023
23 Arpi on Tile
24 Marmomac 2023
25 Adventures in The Trade
26 Stonebiz on the Beach
27 SFA Intro | May-June 2023
28 SFA 1 | Client Contract Terms
29 SFA 2 | Quartzite Translucence and Backing
30 SFA 3 | Contract Labor for Installs
31 ProductTalk | May-June 2023
32 Agenda | May-June 2023
33 Workshelf | May-June 2023
34 Subscriptions
35 The Directory | May-June 2023
36 Contact Info

Spall

Automatic Imitation

Computers generate; but create?

Photo by Vlada Karpovich / Pexels

Listen to this article

Hear it Here!

By Emerson Schwartzkopf

Inevitably, writers will turn to the old trope on artificial intelligence (AI) imitating themselves, so I'll do it now: Am I writing this, or is a computer? However, I’m not going to keep you in suspense until the end. Even though it may be in a terse style of almost robotic consistency, it's still me writing this column. However, the recent hubbub about Chat GPT – a computer portal that can generate hundreds of words on a topic within a matter of seconds – can make anyone start to question the origin of what they’re reading. And, once we let computers take over a creative task like writing, then what about other jobs requiring skill … like, say, stone fabrication? We’ve been using some computerized assistance (or, if you’re not a fan of AI, manipulation) for several years at Stone Update Magazine with our Workshelf feature. Starting with the September/October 2020 issue, we began offering audio versions of selected features and columns, using text-to-voice software. We do this as a convenience to readers who, for whatever reason, don’t have time to stop and read the articles. Listening while driving, waiting, or just filling idle time is a service we can provide as a digital publication. However, a recent problem led me to consider another option for AI. Earlier this month, the deadline loomed for the latest edition of our Radio Stone Update podcast with our producer/narrator, K. Schipper, suffering from a deep chest cough. She’d written the script, but her voice sounded like a four-pack-a-day smoker occasionally interrupted by a Canadian goose. I ended up processing K.’s script through our text-to-speech software, with the artificial reader proclaiming herself as Sydney Hoffman. You can sample the result here and, while it’s a bit stiff in delivery at times, the AI voice does the job. However, the presentation also brought up the debate as to whether we were fooling the audience by giving the AI voice a name. I opted for the alias to help explain why K. wasn’t at the microphone, and also made the broadcast sound as regular as possible. K. researched the news and wrote the script, and I didn’t see Sydney Hoffman as different from any other substitute voice delivering the podcast. We use a lot of technology to produce everything here at Stone Update – this magazine, Hard-Surface Report, and the two editions each of the Slab & Sheet newsletter and Radio Stone Update podcast we crank out every month. We end up using at least 10 software programs (apart from general tools, like email and web browsers) to keep things running. I can remember that, even 20 years ago, doing this kind of production would’ve required a decent-sized suite of offices and a staff of at least 10 to complete all the specialized tasks. We do it today with far less, and that’s due to the progression of AI capabilities. The same thing’s been happening in hard-surface backshops for a long, long time. Each progression of mechanization of cutting, shaping, and polishing reduces the amount of time and effort required to get the job done. It’s conceivable that AI could progress to the point of eliminating human involvement to maintaining machines and doing the installs (at least until robots do a better job of mastering stairs). Do I think that’s going to happen? Nope. AI lacks in traits like creativity and adaptability, which you probably use a lot more than you realize in day-to-day work. The wonders of AI grow by tapping recorded knowledge, but that isn’t the same as experience and problem-solving. AI could pull together the particulars of a slab and pictures of a jobsite to render up a neat and efficient nest of cuts and milling. It won’t have the eye to change the angle of a part or move a seam or one of a hundred other adjustments to solve an on-the-spot problem or a customer’s concern. Meanwhile, you aren’t likely to see AI take over complete production of Stone Update Magazine, or anything else we do. I did give ChatGPT a chance, though, by asking it to write a column about artificial intelligence in the style of Emerson Schwartzkopf. And, within 10 seconds, the program began knocking out “The Transcendental Promise of Artificial Intelligence.” ChatGPT confused me with the 19th century philosopher Ralph Waldo Emerson, so I have 600 words or so in old Ralph W.’s style on “the symbiotic partnership between human ingenuity and the infinite possibilities of the digital realm.” For the time being, I’ll stick with my words.