Holy shit things are moving fast.
We’re living in a world where technological advances are far outpacing society’s ability to acclimate to them. The mind-blowing capability of artificial intelligence is here; however, the safety protocols and regulation of this technology is lagging behind.
The discussion of AI (and ultimately AGI – artificial general intelligence) is front and center in the world today. That discussion seems to be producing two extreme opinions of how it will impact society.
On the one side: AI is a truly breakthrough technology that has the potential to cure cancer, combat climate change, and overall enhance human life.
On the other side: AI advances are being made by a select few powerful organizations at record speed without any regard for safety or regulation, and could ultimately lead to human extinction if AGI – being orders or magnitude smarter than humans – decides that humans are no longer necessary.
I’m not here to talk about all that though. I’m only interested in a very narrow application… how AI can (or cannot) harmonize with woodworking, and any other craft.
I’ll make arguments for what areas AI can help to advance, and where it can’t replace the hands and the mind of a human craftsman. Ok bear with me as I speculate my ass off…
Where AI Can Advance Woodworking
AI has a huge allure on the design side of woodworking. Creating and iterating on designs, making optimized cut lists, creating build plans, and documenting the build process to share with others are all high value areas that could greatly benefit from an “AI assistant”. AI could also help with building out your workshop by helping with layout and tool setup.
Design Iterations
They key with having AI assist in designing furniture is that it is in fact “assisting”. Its not making the decisions, its simply doing what is asked of you and doing it quicker and more efficient than the maker can.
Let’s say you’re working in a 3D design software coming up with a design for a dining table for a client. You develop what you think is a good first draft; if you’re lucky they might love it and want no design changes. But when they inevitably ask “what if the legs were a little more tapered?” or “can you modify the apron to be a little more low-profile?”, you are back to the design drawing board tediously making changes to be able to help the client visualize their piece and give the OK.
In walks AI… you start with a base design and you give it some detailed prompts such as “Please create 4 other versions of this dining room table design. One with X. One with Y. One with X and Y. One with X, Y and Z.” (The letters being specific design changes). The speed at which you could produce alterations to the design and relay those back to the client for their feedback would increase 10-fold. If you’re a maker who is eager to settle on a final design so you can finally take that cathartic trip to the lumberyard for the materials, this could be a win for you.
Build Plans and Cut Lists
If you’re like most, making chicken scratch sketches with hand written dimensions and notes to take into the shop is the standard MO.
Creating high quality build plans that are clear and easy to follow can be incredibly time consuming. Several well-known woodworkers I follow develop and market their own furniture build plans and they are worth every penny because of the time and effort that goes into producing them and the value they provide.
In walks AI… You have your final 3D design completed and you provide an AI bot with parameters such as who the audience of the plans are (kids or adults), the order of the build, special considerations with certain steps such as complicated joinery or glue ups. The AI bot could learn your preferences after one set of plans and continue to learn what you like in terms of visuals, explanation style, level of detail, etc.
Picture a young, ambitious graphic design/copywrite intern who’s eager to learn, works whenever you want, doesn’t need health benefits or compensation. I’d hire them.
You could then either use the plans to build your own furniture or sell then plans for others to utilize.
Shop Layout and Tool Setup
You have a new space to call your shop and you’re ready to outfit it with tools, storage, dust collection and wood storage… But you don’t have the first clue on how to layout your shop to optimize for the type of woodworking you want to do. (This sounds all too familiar)
I have first hand struggled with this for years. As new tools come into my shop, and as my work focus changes, I am always playing with the layout of my shop to try to find the “optimal” setup. The only problem is that most tools require specific power and dust collection, so picking a tool up and moving it somewhere else doesn’t always work without the accompanying rework to power, ducting etc. It’s a nightmare.
In walks AI… You provide it with the type of woodworking you want to do (turning pens and making conference tables don’t require the same setup), the specific tools you have, the storage you’d like, and any other needs.
AI then works to provide an optimal layout for your shop. The tools are located and spaced such that infeed and outfeed requirements are met, you have the optimal workflow for your builds, you have appropriate power and dust collection at every tool. And oh by the way.. those other tools that you hope to acquire in the future (the ones your wife doesn’t know about) – those are factored in and you get specific instructions on how you will want to revise the layout to accommodate those new tools.
Maybe you can even ask for an AI generated pitch to the Wifey on why buying those tools benefits your long-term relationship quality…
What AI Cannot Replace
While AI has some promising applications on the design side of woodworking, there are plenty of areas on the craft side where I feel that AI has absolutely no place. These are the aspects of woodworking that infuse purpose and meaning both into the finished piece and into the maker. Furniture built by machine doesn’t have a story, it doesn’t make mistakes, it can’t build relationships, and it can’t make on-the-fly judgement calls – all of which are uniquely human qualities.
Void of Story
The definition of “Art” – The expression of human creative skill and imagination.
The definition of “Craft” – Activity involving skill in making things by hand.
Without the creativity of the human mind, and the finesse of the human hand, there is no art or craft. There are only prompts and algorithms, controllers and machines.
The opposing argument might be that human thought and creativity are simply biological algorithms, and AI can generate algorithms much more powerful than humans can. AI machines can perform operations and manipulate objects with infinitely more precision and repeatability. So why wouldn’t we just opt for smarter, faster, and more precise?
Because art or craft produced by machine is void of feeling, it’s deprived of story. There exists no human struggle or emotion that can be channeled through an object and felt by the observer or recipient when they interact with it. It’s just dead.
Humans are meaning architects. They take an item that inherently has no meaning or value, and they give it a story, they associate it with something of value, they imbue it with significance, and they share it.
Take a moment and think of some of the most cherished things you own. Ask yourself why those things are so valuable to you. Do they have objective worth? Or did you assign worth based on what they mean to you and the story behind them?
We can’t let the story and the emotion that underpins craft be stripped. If we prioritize human-created art and craft, meaning can continue to be perpetuated and expression of the human condition can continue to be shared.
Human Judgement
AI may be able to rival us in specific areas, but our holistic ability as humans to use our judgement and past experiences to make the right decision seems hard to surpass.
Feeling grain direction, assessing the perfect tightness of a joint, knowing when and how to avoid tear-out on cuts, smelling when the wood is starting to burn on a cut, dealing with minor imperfections in the material… the list goes on.
There are so many aspects of the build process that require intuition and experience. There are simply too many variables to trust with a machine, even if it can self-teach. Let me walk through an example to help visualize…
You starting a new project. It’s going to be a mid-century modern coffee table for a client. You pick up the wood for the project, white oak. You try to pick the boards all from the same lot and thus the same tree so that the color and grain prominence is consistent (AI would fail).
You bring it home and carefully select the area of each board to be used for the legs and the table-top so that the grain direction provides enough strength for the part (AI would fail).
You mill the wood down to its final size over several days and multiple sessions so the wood can acclimate to the shop and come to equilibrium. This will ensure the tabletop stays flat over time and over seasonal humidity changes (AI will fail).
Because of the shape of the table base and the design of the joinery, you must cut joinery for the legs after they’ve been glued up into subassemblies. This requires crafting a table saw jig specific to this one operation (AI might deem this inefficient and choose a sub-optimal joint).
Fast forward – you fasten the table base to the table-top with special fasteners to allow for seasonal expansion of the large table-top. To do this you need to understand what the movement is going to be based on the wood species and where the table will live. (OK maybe AI could do this)
You apply a hard wax oil finish to the table. The instructions say only one coat is required but from experience you know that two coats produces a better sheen and more protection with no downsides, aside from a little extra cost (AI would fail).
With AI at the helm, the finished table might have:
- Inconsistent grain color and direction
- Mediocre joinery due to efficiency of manufacture
- Less than optimal finish
- Fasteners that aren’t suited for the table’s new climate
- A vague feeling of automation and sameness.
Conclusion
The benefits that craft provides the creator cannot be overstated. The work works on us just as much as we work on it – a mutually beneficial contract with nature.
Humans need challenges, problem solving, skill development, critical thinking, and bringing something to life with our minds and our hands gives us exactly that.
AI has immense promise to support craft if it is adopted with careful consideration to help assist the maker, not decide for them. It should simply be another tool, used appropriately for the job at hand, and directed by the craftsman – not the other way around.
Thanks for reading! For all the makers and artists out there, I’d love to hear your perspective on this topic. Are you bullish or bearish on what AI can offer? What would you use it for?