VIDEO REVIEW · TAIWAN VOICE
The Hardest Hire Isn't the Operator, It's the CAM Engineer: A Lesson from a Taiwanese Channel's AI CAM Test
01What this video is about
A-Hong's Workbench (阿鴻的工作桌) is a Taiwanese CAD/CAM tutorial channel, and behind the host is the technical team of SolidCAM's reseller in Taiwan — meaning this is content made by people who face Taiwanese machine-shop customers every day. In April 2026, the channel uploaded this nearly 30-minute video: "寫CAM不用人？AI秒生路徑的時代來了！" (No people needed to write CAM? The era of AI generating paths in seconds is here!), testing the AI CAM assistant embedded in SolidCAM (CAM Assist)[1].
One thing to note up front: a reseller's channel naturally has a promotional stance, but this video's value is that it defines the problem in the language of the Taiwanese shop floor and runs the whole flow on real parts — including what AI can and can't do (the video has a dedicated segment on "what kinds of parts AI can't handle and still need a senior engineer")[1].
02The opening pain point: orders stuck at CAM, not at the machine
The video's opening argument is worth every shop owner hearing once: in the machining industry, the hardest person to hire isn't the CNC operator, but the CAM engineer who takes 5 to 10 years to develop. You can buy machines and train operators, but the people who can turn a drawing into a reliable program are simply that scarce on the market[1].
The reality that follows is the title of the video's second segment: "The machine shop's real predicament: orders stuck at CAM." The machine has idle time and the order is in hand, but the program can't be scheduled out — programming has become the capacity bottleneck. This aligns exactly with the Taiwanese reality we laid out in The CNC Labor-Shortage Self-Help Guide: what's missing isn't just manual labor, but knowledge labor.
03What the video tested: quoting, demo parts, and a clamp
The video's hands-on portion is split into several segments, each with something to watch for:
- A human-vs-machine comparison at the quoting stage: the video separately shows an engineer and the AI CAM assistant working through the quoting stage — to quote, you first have to roughly estimate operations and machining time, and this segment shows how AI speeds up the "estimating" prep work[1]. For why quoting eats programming resources, see The AI Approach to Machining Quotes.
- Demo parts and a LANG clamp: from a standard demonstration part to a real workholding clamp, the video runs the full AI path-generation flow and also demonstrates CAM Assist's module requirements — letting viewers see the software prerequisites before adoption[1].
- Chatbot: the closing segment shows the SolidCAM Chatbot, handing the learning cost of "looking up a feature, asking how to operate" over to AI as well[1].
The video's defining line is: "This isn't AI vs humans, it's people who can use AI vs people who can't." — which arrives, by a different route, at the same conclusion as the hands-on tests in the international TITANS of CNC series (see The US Hands-On Case Study)[1].
04The labor shortage isn't unique to Taiwan: what international research says
Pull the camera back and Taiwan's CAM-talent crunch is a local version of a global phenomenon. The World Economic Forum's Future of Jobs Report notes that manufacturing faces a clear skills gap and that reskilling needs keep growing[2]; Deloitte and the Manufacturing Institute documented the structural gap of "the jobs are here but the people aren't" even earlier[3].
The automation literature's consistent answer is this: when a class of skilled labor becomes the bottleneck, systematizing and automating the rule-based part of that skill is the path to getting the greatest leverage out of limited talent[4]. CAM programming is a textbook-grade case — part of it is rules (features, strategies, parameter combinations), part of it is judgment (tolerances, datums, process methods). AI takes over the former, the scarce engineer focuses on the latter, and one person's output is multiplied.
05Action options for shop owners: get the team "able to use AI" first
If you agree with the "people who can use AI vs people who can't" framing, the action list is actually short:
- Inventory the bottleneck: over the past three months, how many orders had their delivery held up by "waiting for the program"? That's your baseline.
- Controlled pilot: pick one or two representative drawings and have the AI tool run the full flow of read drawing → model → generate code → simulate, and measure the labor-hour difference (for the method, see The SMB ROI Assessment Checklist).
- Train the people you already have: you don't have to wait for an engineer who takes a decade to develop — having your current technicians learn to review AI's output is often far faster than learning programming from scratch.
BestAI CAM's place on this path is to pull the starting point one step further forward: what the Taiwanese contract-manufacturing floor receives is often a 2D drawing rather than a 3D model, so our AI takes over from reading the drawing and modeling, generates G-code based on the in-house tool library and machine limits, forces cutting simulation, and has an independent AI cross-check dimensions — with people keeping watch only at the key decision points. It doesn't conflict with shops already using SolidCAM or Mastercam either — we've done integration testing with mainstream CAM, and AI is a value-add layer, not a replacement (see Does AI CAM Replace It or Add Value?).
06FAQ
Why do we say orders get stuck at CAM, not at the machine?
Machine capacity can be expanded with money and operators can be trained relatively quickly, but the CAM engineer who turns a drawing into a reliable program takes years to develop and is scarce in the market. When the programming schedule is fuller than the machine schedule, orders pile up at the "waiting for the program" gate — the video calls this the machine shop's real predicament, and it's daily life at many Taiwanese shops.
The video was made by a reseller — is its content trustworthy?
Watch it with an awareness of its stance: the product choice and evaluation naturally lean toward promotion. But the video's value lies in its problem definition (the CAM-talent crunch) and its full hands-on flow (including the part types AI can't handle and the module prerequisites), and this content can be cross-checked against other sources — international channel tests and labor-shortage research all point the same way. Treat it as a "localized statement of the problem" rather than a purchasing basis, and make real decisions by piloting with your own drawings.
What's the relationship between the AI CAM assistant and BestAI CAM?
The CAM Assist tested in the video is an AI plug-in embedded in existing CAM software (such as SolidCAM), generating machining strategies from a 3D model. BestAI CAM starts further upstream, from the 2D drawing: AI reads the drawing, builds the 3D model, generates G-code based on the in-house tool library and machine limits, and adds cutting simulation plus independent dimensional cross-checking. The two share the same philosophy (AI takes over the rule-based work, people handle judgment); they enter at different points and can coexist.
A small shop has no CAM engineer — can it just replace one with AI?
We'd advise against thinking that way. AI output still needs a person with machining judgment to review it — tolerances, datums, workholding, and on-machine decisions are all human responsibilities. The practical approach is to have your most experienced current technician learn to "review AI output," which is faster than training a full programming capability and safer than having no one keeping watch.
Get notified when new articles and video walkthroughs go live — no inbox spam, unsubscribe in one click.
FOR TAIWAN SHOPS
Short on CAM engineers? Let AI take over from drawing to G-code first
Pilot with one of your own 2D drawings: AI models it, generates code based on your tool library, simulates, and cross-checks dimensions — and measure how many programming hours it saves.
Contact an adoption consultant Training courses📋 Download the free "CNC Outsourcing Checklist" (printable) (中文)
07References
- 阿鴻的工作桌 (YouTube). 寫CAM不用人？AI秒生路徑的時代來了！ (published 2026-04-02). youtube.com/watch?v=u4xUMT-99xo
- World Economic Forum (2023). The Future of Jobs Report 2023.
- Deloitte & The Manufacturing Institute (2018). The jobs are here, but where are the people?
- Groover, M. P. (2019). Automation, Production Systems, and Computer-Integrated Manufacturing (5th ed.). Pearson.