Computer Numerically Controlled Tool Programmers
51-9162.00Median wage $68,120/yr28,500 employed (US)Rank #163 of 923 scored · top 18% by substitution
Develop programs to control machining or processing of materials by automatic machine tools, equipment, or systems. May also set up, operate, or maintain equipment.
Sub-scores
0–100 · band = confidence interval from rater disagreement
Substitution41
Exposure39
Augmentation62
Substitution — the headline: capability discounted by cost, barriers and adoption.
Exposure — technical capability alone, regardless of whether anyone deploys it.
Augmentation — how much AI assists without replacing. High here + moderate substitution = a changing job, not a disappearing one.
Tasks on the substitution scale
16 rated tasks, binned by substitution score.
Position among all scored occupations
Distribution of 923 occupation scores; the marker is this occupation.
Tasks with substitution ≥ 70
0%
Run 1.0.0-draft.1 · computed 2026-08-05 · rater panel: claude-sonnet-5, claude-haiku-4-5-20251001 · intervals span rater disagreement.
Why this score
The five weighted dimensions of the composite, averaged across this occupation's tasks (importance-weighted, panel mean). Exact weights and formulas: /api/v1/methodology.
Task automatabilityw 35%42
panel mean rating 2.7/5 → substitution pressure 42/100
Technical feasibility todayw 20%35
panel mean rating 2.4/5 → substitution pressure 35/100
Cost vs. human wagew 15%36
panel mean rating 2.4/5 → substitution pressure 36/100
Adoption barriersw 20%inverted — strong barriers lower the score53
panel mean rating 2.9/5 (barrier strength) → substitution pressure 53/100
Sector adoption velocityw 10%29
panel mean rating 2.1/5 → substitution pressure 29/100
Task breakdown (16 tasks)
Substitution pressure per task, weighted by O*NET importance in the composite. Expand a task for the full rater audit trail — every rating, every model, every rationale.
Determine reference points, machine cutting paths, or hole locations, and compute angular and linear dimensions, radii, and curvatures.
62CI 46–79 · exposure 62 · augmentation 88 · importance 4.3/5 · click for rater detail
Determine reference points, machine cutting paths, or hole locations, and compute angular and linear dimensions, radii, and curvatures.
62| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Advanced manufacturing and aerospace sectors show rapid CAM automation adoption with production AI-assisted tool path systems; smaller job shops lag, but the trend toward lights-out and increasingly autonomous CAM in larger operations is well-established. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Manufacturing has adopted CAM/CAD automation for decades, but full AI-driven programming (beyond parametric CAM) is still pilot-stage in most shops, reflecting middling adoption typical of manufacturing sectors. |
| Augmentation potential | claude-haiku-4-5-20251001 | 5/5 | AI-powered CAM systems substantially augment programmer productivity by auto-generating initial paths, validating geometries, and suggesting optimizations; programmers remain in the loop for oversight, constraint handling, and refinement, creating a highly productive human–AI partnership. |
| Augmentation potential | claude-sonnet-5 | 4/5 | CAM and AI-assisted design tools significantly speed up dimensional calculations, path generation, and geometric checks, letting programmers focus on optimization and quality control while the software handles heavy computation. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | CNC tool path generation and dimension calculation are largely rule-based and mathematical tasks that can be automated end-to-end from CAD models using established algorithms; current CAM software (Fusion 360, Mastercam plugins with AI features) can handle significant portions autonomously, achieving >50% time savings for standard geometries. |
| Task automatability | claude-sonnet-5 | 3/5 | CAM software already automates much of toolpath generation and geometric computation from CAD models, but determining optimal reference points and cutting strategies for complex parts still requires human judgment integrated with the software.CAM automation is partial and requires setup/verification. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Safety and liability concerns are real (improper tool paths can damage equipment or cause injury), and many organizations require experienced programmer sign-off on critical parts; however, no legal licensing of the programmer role itself exists in most jurisdictions, limiting hard barriers. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement exists, but errors in toolpath computation can cause costly scrap, tool breakage, or safety issues, creating moderate liability-driven caution before removing human oversight. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 5/5 | CAM software licensing and compute costs are modest relative to a programmer's loaded wage ($60–80k+ annually); a single seat can handle dozens of programs per day, delivering an order-of-magnitude cost advantage. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | CAM software licenses and skilled operators remain costly, and while software speeds up computation, a trained CNC programmer is still needed to run and verify the software, keeping costs closer to comparable rather than order-of-magnitude cheaper. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Mature CAM/CAD-integrated tools with automated tool path generation are deployed in production across manufacturing; however, complex part geometry, non-standard requirements, or safety-critical tooling still benefit from human verification, keeping this below fully autonomous maturity. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | CAM systems (Mastercam, Fusion 360, etc.) are mature production tools that compute toolpaths, angles, and curvatures reliably for many part types, but they are assistive tools requiring a programmer to set parameters and validate results, not fully autonomous end-to-end solutions. |
Sort shop orders into groups to maximize materials utilization and minimize machine setup time.
59CI 48–70 · exposure 58 · augmentation 75 · importance 3.8/5 · click for rater detail
Sort shop orders into groups to maximize materials utilization and minimize machine setup time.
59| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Manufacturing, especially CNC-heavy sectors, has been actively adopting scheduling and optimization software for two decades; modern AI-based variants are now entering production in mid-sized and larger job shops seeking margin improvement on material costs and setup time. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing, especially small-to-mid CNC shops, tends to be a slower-adopting sector for advanced optimization software compared to information/professional services, with many shops still relying on manual or spreadsheet-based sorting. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can provide ranked order groupings with predicted material waste and setup time for each scenario, allowing programmers to validate assumptions and override decisions for customer constraints or long-tail rules the algorithm missed, substantially accelerating the planning cycle. |
| Augmentation potential | claude-sonnet-5 | 4/5 | Nesting and scheduling software substantially assists programmers by automatically suggesting groupings and optimized layouts, letting the human focus on exceptions and final validation. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | This task involves logical optimization of order grouping based on quantifiable constraints (material utilization, setup time). Modern optimization algorithms and constraint solvers can handle this end-to-end with demonstrable time savings over manual sorting, though some domain-specific knowledge about material properties and machine configurations may require light customization. |
| Task automatability | claude-sonnet-5 | 3/5 | Batch sorting/nesting to optimize material use and setup time is a well-defined optimization problem that scheduling/nesting software can partially automate, but integration with shop-floor variability and priorities still requires human judgment for full end-to-end handling. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Implementation requires integration with shop floor systems, customer approval of grouping changes that may affect delivery schedules, and human verification of material waste predictions. Liability concerns around delayed orders or scrap creation create modest organizational friction, though no legal licensing requirement prevents automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal requirement mandates human sign-off for this scheduling task, though shops may have some organizational inertia and preference for programmer judgment in complex material/setup tradeoffs. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Automated scheduling software has high upfront licensing and integration costs but minimal per-task inference cost compared to a skilled programmer's hourly labor. At scale across multiple batches and jobs, the AI cost becomes a small fraction of the human wage equivalent. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Software licensing and integration costs are moderate, and while automated nesting can reduce time spent versus manual sorting, the need for oversight and setup keeps costs roughly comparable to a skilled programmer's time for many shops. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Specialized optimization software and heuristic scheduling tools exist in production CNC environments, but they typically require significant manual input for problem setup and validation of results against shop floor realities. General-purpose AI agents struggle with real-world shop complexity and material waste prediction without domain integration. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Nesting and production scheduling software (e.g., CAM nesting tools, MES systems) are deployed in industry today, but they typically require human review and adjustment for exceptions, rush orders, and machine-specific constraints. |
Observe machines on trial runs or conduct computer simulations to ensure that programs and machinery will function properly and produce items that meet specifications.
57CI 30–84 · exposure 58 · augmentation 63 · importance 4.4/5 · click for rater detail
Observe machines on trial runs or conduct computer simulations to ensure that programs and machinery will function properly and produce items that meet specifications.
57| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 4/5 | Manufacturing and advanced industrial sectors are rapidly adopting automated QA, machine monitoring, and simulation-based verification. Deployment is accelerating across automotive, aerospace, and electronics production. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing is a moderate-to-low digitization sector; CNC programming and quality verification adoption of AI remains in pilot stages rather than widespread production deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI can assist programmers by continuously monitoring trial runs, flagging anomalies in real-time, and suggesting parameter adjustments based on simulation results, significantly accelerating the debug-and-verify cycle while the programmer stays informed. |
| Augmentation potential | claude-sonnet-5 | 3/5 | Simulation software and AI-assisted toolpath verification meaningfully help programmers catch errors before physical trials, improving productivity while humans remain responsible for final verification. |
| Task automatability | claude-haiku-4-5-20251001 | 5/5 | AI systems can observe machines, conduct simulations, compare outputs against specifications, and flag discrepancies automatically with >50% time savings. Computer vision, sensor data analysis, and simulation software can fully replace the observation and analysis steps with proper sensor integration. |
| Task automatability | claude-sonnet-5 | 2/5 | Requires physical observation of machine trial runs and interpretation of real-world mechanical behavior, which current AI cannot fully perform without human presence and judgment, though simulation analysis portions could be partially assisted. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | No legal requirement for a human to perform this observation; manufacturers can fully automate. Some organizational friction exists around adoption and maintenance of AI systems, but these are not hard barriers to substitution. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement specifically, but liability for defective parts, safety around running machinery, and quality assurance protocols create organizational friction against fully removing human oversight. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | Once vision and simulation systems are integrated (often as part of manufacturing execution systems), the per-task inference cost is minimal relative to operator time. A human programmer hour costs $30–50 loaded; automated checking costs pennies. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Simulation software has some cost efficiency, but physical trial run observation still requires skilled human oversight, keeping all-in costs comparable to human labor rather than dramatically cheaper. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 4/5 | Deployed machine vision systems, simulation software, and anomaly detection tools exist in production for CNC monitoring and quality verification. Some manual oversight remains common, but the core technical capabilities are mature and widely available in manufacturing environments. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | CAM software includes simulation tools, but AI-driven autonomous verification of trial runs and specification compliance in production CNC environments is not a mature deployed capability; humans still perform physical monitoring. |
Modify existing programs to enhance efficiency.
47CI 30–65 · exposure 45 · augmentation 75 · importance 4.1/5 · click for rater detail
Modify existing programs to enhance efficiency.
47| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Manufacturing and CNC programming remain relatively traditional sectors with slower digital adoption; while large aerospace/automotive suppliers are piloting AI-assisted tools, small and mid-sized shops lag significantly in AI adoption. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing is a lower-digitization sector with slower AI adoption; CAM/CNC programming still relies heavily on specialized software and human expertise rather than general AI agents. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | AI assistants can draft optimizations, suggest alternative toolpaths, and highlight inefficiency patterns, substantially raising programmer productivity when the human retains review and final decision-making authority. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-assisted CAM tools, simulation software, and code-suggestion assistants can meaningfully speed up identifying optimization opportunities and drafting revised code for human review. |
| Task automatability | claude-haiku-4-5-20251001 | 4/5 | AI can analyze existing code, identify inefficiency bottlenecks, and generate optimized CNC program modifications with substantial time savings. Most of the optimization logic is automatable, though edge cases and domain-specific manufacturing constraints may require human review. |
| Task automatability | claude-sonnet-5 | 2/5 | AI can suggest code/toolpath optimizations from CAM software or NC code text, but validating on the actual machine, tolerances, and material behavior requires physical testing and expert judgment that current AI cannot fully replace. |
| Adoption barriers | claude-haiku-4-5-20251001 | 2/5 | Manufacturing roles value domain expertise and accountability, but no strict licensing or legal requirement mandates human sign-off on program modifications; organizational friction and preference for human verification provide some friction but are not hard regulatory barriers. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No strict licensing requirement, but liability for scrapped parts, machine damage, or safety incidents from faulty program changes creates meaningful organizational caution before trusting AI-only edits. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 4/5 | AI inference and code analysis are low-cost; integration overhead is modest for teams already using development tools. The cost per optimized program is likely 3–10× cheaper than a human programmer's hourly rate for equivalent work. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI assistance can reduce some analysis time cheaply, but human verification, machine trial runs, and safety checks remain mandatory, keeping overall cost comparable to skilled programmer time. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 3/5 | Code analysis and generation tools (GitHub Copilot, specialized CNC software plugins) exist and perform well on standard optimization patterns, but reliability varies with program complexity and specialized domain knowledge; production systems are emerging but not yet universally proven at scale. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some CAM software includes optimization algorithms and simulation, and AI coding assistants can suggest edits to G-code, but no mature deployed product autonomously modifies and validates CNC programs for efficiency at scale in production. |
Enter computer commands to store or retrieve parts patterns, graphic displays, or programs that transfer data to other media.
45CI 25–65 · exposure 45 · augmentation 63 · importance 4.2/5 · click for rater detail
Enter computer commands to store or retrieve parts patterns, graphic displays, or programs that transfer data to other media.
45| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Manufacturing remains a relatively slow adopter of AI automation, with legacy CNC systems still prevalent and conservative risk postures around fully automated command entry. Most adoption remains in pilot phases or limited to non-critical data retrieval. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing and machining sectors adopt automation more slowly than pure information sectors, with digitization of legacy CNC systems still uneven across firms. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can meaningfully assist by auto-generating command templates, suggesting syntax, and helping format data for transfer, thereby raising programmer productivity without replacing human judgment on machine-specific safety and accuracy requirements. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-assisted file management, version control, and command generation tools can substantially speed up a programmer's data handling and reduce manual errors. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist with code generation and data transfer workflows, this task requires understanding specific CNC machine contexts, legacy system compatibility, and precision in data transfer operations. Current AI cannot reliably end-to-end execute these specialized commands without significant human oversight and domain knowledge. |
| Task automatability | claude-sonnet-5 | 4/5 | Entering commands to store/retrieve patterns or transfer data between media is a well-defined, repetitive digital operation that scripting or AI-driven CAM/PLM integration can largely automate.this is a routine data-management action, not requiring judgment., |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Manufacturing environments often have strict regulatory requirements (ISO, tool safety standards), liability concerns around faulty part patterns, and organizational preferences for human sign-off on machine commands. Many CNC systems require certified operators to authorize program execution. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing or legal barriers; the main friction is integration with proprietary CNC/CAM software and organizational IT policies rather than regulatory or human-judgment requirements. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI tools for code assistance have modest costs, but the need for specialized domain expertise and human validation to ensure correctness means total cost remains comparable to or higher than skilled CNC programmer labor for critical tasks. |
| Cost vs. human wage | claude-sonnet-5 | 4/5 | Automated scripts or lightweight AI agents can execute these data transfer/storage commands at a fraction of the cost of a human programmer's time once the system is set up. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably handles end-to-end CNC command entry and data transfer without human intervention. AI code assistants can generate snippets, but they lack the domain specificity for manufacturing systems and cannot validate machine compatibility or safety-critical data integrity. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | CAM software and file management systems already automate much of this via macros, APIs, and integrations, but full AI-driven autonomous handling across diverse legacy systems and formats is still inconsistent in production. |
Order tooling for jobs.
41CI 30–52 · exposure 38 · augmentation 63 · importance 3.6/5 · click for rater detail
Order tooling for jobs.
41| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | CNC shops and manufacturing environments show slower adoption of advanced AI agents compared to information-intensive sectors. While e-procurement and inventory systems are common, generative AI automation of tooling selection and ordering remains in pilot phases rather than production deployment in most manufacturing organizations. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing and machining shops are generally slower adopters of AI-driven procurement compared to information/finance sectors, though e-procurement tools are gradually spreading. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by suggesting appropriate tools from catalogs based on job parameters, cross-checking inventory levels, identifying cost-effective alternatives, and drafting purchase orders—all useful aids to a human CNC programmer reviewing and approving tooling decisions. The human retains final judgment on tool selection and supplier choice. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-assisted procurement and inventory systems can meaningfully speed up identifying, comparing, and ordering correct tooling while the programmer remains responsible for final specification decisions. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Ordering tooling requires understanding job specifications, tooling catalogs, supplier relationships, and budget constraints. While AI could assist in identifying standard tooling options from specifications, the task involves judgment calls, supplier negotiations, inventory management, and approval workflows that typically require human oversight and decision-making today. |
| Task automatability | claude-sonnet-5 | 3/5 | Ordering tooling involves checking specs, comparing suppliers, and placing orders, which can be substantially automated via procurement software and AI agents, but selecting correct tooling for a specific job often requires domain judgment tied to CNC program details. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | Tooling orders often require authorization sign-off from supervisors or cost-center owners, and supplier relationships and contractual terms create some organizational friction. However, there are no strict licensing or legal requirements preventing automation of the order placement itself, and digital procurement systems are already established. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for ordering tooling, though budget approval workflows and vendor relationships create some organizational friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI systems (chatbots, document processors) would require significant integration with ERP/procurement systems, human oversight for tool selection accuracy, and manual approval workflows. The all-in cost (setup, integration, error correction, oversight) likely exceeds the wage cost of a CNC programmer spending a portion of their time on this task. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Automated procurement tools reduce clerical time significantly, but oversight is still needed to verify correct tooling specs, keeping costs comparable rather than order-of-magnitude cheaper in most small-shop contexts. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No mature production systems reliably handle end-to-end tooling procurement across diverse CNC job requirements and supplier ecosystems. While e-procurement platforms and inventory systems exist, they require significant human input on tool selection, approval, and vendor management; standalone AI automation is research-stage. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Procurement automation and e-commerce ordering systems are mature and widely deployed, but integration with tooling selection based on specific machining requirements is narrower and less standardized across shops. |
Determine the sequence of machine operations, and select the proper cutting tools needed to machine workpieces into the desired shapes.
39CI 30–49 · exposure 38 · augmentation 75 · importance 4.6/5 · click for rater detail
Determine the sequence of machine operations, and select the proper cutting tools needed to machine workpieces into the desired shapes.
39| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | CNC shops and manufacturers are digitizing slowly relative to software sectors; most small-to-medium job shops still rely on manual programming or basic parametric templates. Adoption of advanced AI-driven CAM is concentrated in large aerospace and automotive suppliers, not widespread across the sector. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing is a moderately digitized but physically-grounded sector; CAM automation adoption is steady but slow-moving compared to information-sector AI adoption, with many shops still relying heavily on experienced programmers. |
| Augmentation potential | claude-haiku-4-5-20251001 | 4/5 | Modern CAM systems and AI-assisted code generation substantially speed up the initial sequence draft and tool recommendation, allowing programmers to focus on optimization and validation rather than tedious baseline generation. This augmentation is actively deployed and materially raises programmer productivity on this task. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-enhanced CAM tools significantly speed up toolpath generation, simulation, and tool selection suggestions, substantially boosting programmer productivity while the programmer still makes final decisions and validates output. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist in tool selection and generate candidate operation sequences, the task requires domain expertise in machinability, tool geometry, workpiece material properties, and machine constraints that vary significantly per job. Current AI struggles with the full end-to-end optimization and verification needed to confidently machine complex shapes without substantial human oversight and iteration. |
| Task automatability | claude-sonnet-5 | 3/5 | CAM software with AI-assisted toolpath optimization can generate operation sequences and suggest tooling for many standard parts, but complex geometries, tolerances, and material-specific tool selection still require skilled human judgment and validation. |
| Adoption barriers | claude-haiku-4-5-20251001 | 3/5 | No legal licensing barrier exists, but strong organizational friction arises from liability concerns (tool breakage, scrap, machine damage from poor sequences), quality assurance requirements, and customer expectations that a certified programmer sign off on the code. Error costs are high and asymmetric. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement mandates a human for this task, but liability for tool crashes, workpiece damage, and machine safety creates practical friction that keeps humans in the verification loop. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | AI-assisted CAM reduces programming time but does not yet eliminate the need for a skilled CNC programmer to review, adjust, and validate; the cost of the AI system, integration, and required human oversight remains comparable to or higher than the labor it partially displaces for most job shops. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | CAM software licenses plus skilled oversight to verify tool paths and sequences remain costly relative to the labor saved; full automation without human review risks costly tool breakage or scrap, keeping all-in cost comparable to or higher than a programmer's time on complex jobs. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | CAM software with parametric and AI-assisted features exists but typically requires skilled programmers to validate and refine outputs; fully autonomous sequence generation and tool selection without human review is not standard in production. Most deployed systems serve as drafting aids rather than autonomous decision-makers. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | Commercial CAM systems (Mastercam, Fusion 360, etc.) with automated feature recognition and tool libraries are deployed in production, but they require significant human setup, verification, and post-processing to be reliable for varied workpieces. |
Compare encoded tapes or computer printouts with original part specifications and blueprints to verify accuracy of instructions.
39CI 25–52 · exposure 38 · augmentation 63 · importance 3.8/5 · click for rater detail
Compare encoded tapes or computer printouts with original part specifications and blueprints to verify accuracy of instructions.
39| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Manufacturing remains moderately digitized with slower AI adoption compared to information/finance sectors; while some facilities experiment with automated inspection, CNC program verification still relies heavily on traditional manual review due to cost of errors and workforce familiarity with legacy processes. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing sectors adopt automation but at a slower, more capital-intensive pace than digital/professional services, with CAM verification tools adopted unevenly across shops. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-assisted comparison tools (highlighting discrepancies, flagging tolerance violations) can meaningfully speed a human technician's review process, though the task fundamentally requires human judgment to validate against complex specifications and approve for production. |
| Augmentation potential | claude-sonnet-5 | 4/5 | CAM simulation and verification software significantly speeds up and improves the accuracy of a programmer's manual comparison process, serving as a strong assistive tool while the human still reviews and confirms. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Comparing encoded tapes/printouts to blueprints requires visual pattern matching and technical domain understanding that current AI can partially support (OCR, image comparison), but the task involves spatial reasoning, tolerance verification, and high-consequence error detection where AI cannot reliably achieve the 50% time-saving threshold at equal quality without substantial human oversight. |
| Task automatability | claude-sonnet-5 | 3/5 | AI vision/CAM verification tools can compare G-code or toolpaths against CAD specs and flag discrepancies, but full verification against physical blueprints and tolerances often still needs human judgment and setup integration.of CAM systems.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Manufacturing quality assurance, particularly for CNC work, often carries regulatory requirements (ISO, industry standards) and liability for errors; human sign-off is frequently mandated by process control regulations and customer contracts, creating hard barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No formal licensing requirement for this specific verification task, but quality/safety liability in manufacturing creates some incentive for human sign-off before production runs. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current vision AI and document processing tools have modest per-task costs, but integration, setup, and mandatory human verification add overhead that approaches or exceeds the cost of a skilled technician performing the task directly, especially given error-consequence sensitivity. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | Verification software licenses and integration costs are moderate; while faster than manual checking, it still requires skilled oversight, keeping cost savings moderate rather than order-of-magnitude. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While vision models can identify and extract text from technical documents, deployed systems lack reliable integration for verifying CNC instruction accuracy against complex engineering specifications; benchmarks show promise but production systems in manufacturing remain limited and require heavy human validation. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | CAM software includes simulation and verification modules that check toolpaths against part geometry, and these are used in production, but full automated blueprint-to-code cross-checking with error catching is not universally deployed. |
Write instruction sheets and cutter lists for a machine's controller to guide setup and encode numerical control tapes.
39CI 25–52 · exposure 38 · augmentation 63 · importance 3.5/5 · click for rater detail
Write instruction sheets and cutter lists for a machine's controller to guide setup and encode numerical control tapes.
39| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Manufacturing sectors adopting CAM have done so over decades; AI augmentation of CNC programming is early-stage with pilots in larger shops only. Most small-to-medium shops still rely on experienced programmers, and adoption of AI-driven code generation in production remains limited and cautious. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing/machining is a moderately digitized but physically-oriented sector with slower AI adoption compared to information services, though CAM software adoption is mature. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | CAM tools substantially assist programmers by auto-generating G-code from models, reducing manual syntax work. However, augmentation is limited by the need for human expertise in tool selection, feeds-and-speeds optimization, and verification—AI today supports drafting but does not transform the core problem-solving work. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-enhanced CAM tools significantly speed up toolpath generation and documentation drafting, letting programmers focus on verification and edge cases rather than manual coding. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Modern CAM software can automatically generate G-code and cutter lists from 3D models, but creating complete instruction sheets for setup still requires human judgment about specific machine configurations, material properties, and production context. End-to-end automation with 50% time savings would require AI to understand machine-specific constraints that today's general systems cannot reliably infer. |
| Task automatability | claude-sonnet-5 | 3/5 | AI/CAM software can generate G-code and cutter lists from CAD models with human oversight, but writing accurate instruction sheets tailored to specific machine setups still requires domain verification and physical context understanding.: |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Significant barriers exist: machine-specific expertise is highly specialized, liability for setup errors causing tool damage or part defects is substantial, and manufacturers require sign-off from qualified personnel. Regulatory compliance in some precision manufacturing and aerospace contexts further restricts automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement mandates a human to write these sheets, but errors in numerical control tapes can cause costly machine damage or scrap, creating a real but not legal barrier to full automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | CAM software licensing and AI inference costs are modest, but the integration, customization, and human validation required for each job substantially increase total cost. For simple operations, cost approaches parity with skilled programmer wages; for complex setups, AI assistance costs remain high relative to human labor. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | CAM/CAD software licenses plus computation are cheaper than programmer time for repetitive tasks, but complex or novel setups still require expensive skilled labor to verify and correct output. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | CAM software exists that generates numerical control code, but these are specialized tools requiring expert human input and validation rather than end-to-end AI systems. No deployed AI product autonomously writes setup instruction sheets and generates production-ready cutter lists without human oversight. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | CAM software with AI-assisted toolpath generation is widely deployed, but fully autonomous generation of complete instruction sheets and cutter lists without skilled programmer review is not standard practice. |
Write programs in the language of a machine's controller and store programs on media, such as punch tapes, magnetic tapes, or disks.
38CI 29–47 · exposure 33 · augmentation 63 · importance 4.4/5 · click for rater detail
Write programs in the language of a machine's controller and store programs on media, such as punch tapes, magnetic tapes, or disks.
38| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Manufacturing remains a laggard sector in AI adoption relative to information and finance. CNC programming is performed by specialized technicians in often-traditional shops; digital transformation and AI integration in this domain are slow and fragmented. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing is a comparatively slow-adopting sector for AI compared to information/finance; CAM automation has existed for decades but AI-driven programming remains in early pilot stages in most shops. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can usefully assist CNC programmers by suggesting code patterns, automating routine syntax generation, or checking for common errors, raising efficiency on code drafting and review tasks. However, the human programmer must remain in full control of machine behavior and safety logic. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI and advanced CAM/post-processor tools significantly speed up code drafting, error-checking, and optimization, letting programmers focus on complex fixturing and verification while iterating faster. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Modern AI can generate or assist with CNC code snippets and syntax, but end-to-end program generation from specifications remains unreliable without domain expertise verification. The task requires understanding machine capabilities, tool geometry, material properties, and safety constraints—knowledge that AI systems struggle to encode reliably without human oversight. |
| Task automatability | claude-sonnet-5 | 3/5 | CAM software already generates much G-code automatically, and AI can assist in drafting or converting code, but final programs still require verification against specific machine controllers, tooling, and materials, limiting full end-to-end automation., especially for storing to niche media. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | CNC programming is heavily regulated in safety-critical manufacturing contexts; incorrect code can damage expensive tooling, parts, or injure operators. Liability and the requirement for human verification and sign-off before machine execution create substantial legal and organizational barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement for CNC programming itself, but safety-critical nature of machining (tool crashes, part scrap, injury risk) creates strong incentive for human verification, and legacy storage media (punch tape) adds practical friction. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 3/5 | AI inference costs for code generation are minimal, but integration into CNC workflows, domain-specific training, and mandatory human review/testing overhead make the all-in cost roughly comparable to an experienced CNC programmer's wage for new program development. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | CAM software licenses and skilled programmer time are moderate costs; AI-assisted generation could reduce programming time but still requires human oversight and verification, keeping costs roughly comparable rather than order-of-magnitude cheaper. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | While code generation tools exist (e.g., GitHub Copilot, ChatGPT), they are not deployed in production CNC programming workflows as primary generators. Existing CNC software is vendor-specific and highly specialized; general-purpose AI tools lack the manufacturing domain knowledge and testing rigor required for production use. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | CAM software with automated toolpath generation is mature, but AI-specific code generation for CNC controllers is not yet a widely deployed production practice; most shops rely on CAM post-processors rather than generative AI models. |
Prepare geometric layouts from graphic displays, using computer-assisted drafting software or drafting instruments and graph paper.
38CI 21–55 · exposure 38 · augmentation 63 · importance 3.8/5 · click for rater detail
Prepare geometric layouts from graphic displays, using computer-assisted drafting software or drafting instruments and graph paper.
38| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | CNC programming remains a specialized, skill-dependent domain in manufacturing. Adoption of AI for autonomous layout generation is minimal; most shops still rely on human programmers using CAD tools, with only early pilots of AI-assisted design emerging in leading firms. |
| Sector adoption velocity | claude-sonnet-5 | 3/5 | Manufacturing and machining sectors show moderate digitization with CAD/CAM adoption common, but full automation of layout tasks remains at the pilot/assisted stage rather than widespread autonomous deployment. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI-powered CAD tools and generative design systems can assist programmers by suggesting layouts, automating routine geometry tasks, and checking design rules, meaningfully raising productivity. However, human judgment remains central to validating designs against machining constraints and customer requirements. |
| Augmentation potential | claude-sonnet-5 | 4/5 | CAD software dramatically speeds up geometric layout creation, offering templates, parametric design, and automated dimensioning that substantially boost a programmer's productivity while they retain design control. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While CAD software can assist in layout creation, the task requires interpreting graphic displays, applying domain knowledge about CNC machining constraints, and making design decisions that current AI cannot reliably do end-to-end. AI can help generate candidate layouts but cannot autonomously produce production-ready geometric layouts meeting manufacturing specifications. |
| Task automatability | claude-sonnet-5 | 3/5 | CAD/CAM software can generate geometric layouts from parametric or graphic input with significant automation, but complex part geometry and machining constraints still require skilled human setup and verification, so only partial time savings are achievable off-the-shelf. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | CNC programming has high error-cost asymmetry—mistakes can damage expensive tooling and parts—creating liability and quality assurance requirements. Industry practice and customer specifications typically require a licensed programmer to own and sign off on layouts, creating organizational and contractual barriers to full automation. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement mandates a human perform this specific drafting subtask, though quality/safety concerns in manufacturing create some institutional caution before removing human review entirely. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | A skilled CNC programmer's loaded wage is substantial, and the cost of CAD software licenses plus integration, plus the requirement for human oversight to ensure correctness, makes AI solutions currently more expensive than the human alternative for reliable output. |
| Cost vs. human wage | claude-sonnet-5 | 3/5 | CAD/CAM software licenses and compute costs are moderate, but skilled programmer oversight is still required to verify layouts, keeping costs roughly comparable to a human programmer's loaded wage for equivalent quality. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | CAD software exists and is widely deployed, but it is a tool requiring skilled human operators, not an autonomous system. Current AI for layout generation exists only in research/early-stage forms; no production AI systems reliably generate CNC-ready geometric layouts without human review and correction. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | CAD/CAM tools with automated layout and toolpath generation features are widely deployed in manufacturing, but reliable end-to-end geometric layout creation from graphic displays without human drafting input remains narrow and error-prone for complex parts. |
Analyze job orders, drawings, blueprints, specifications, printed circuit board pattern films, and design data to calculate dimensions, tool selection, machine speeds, and feed rates.
37CI 25–49 · exposure 38 · augmentation 63 · importance 4.5/5 · click for rater detail
Analyze job orders, drawings, blueprints, specifications, printed circuit board pattern films, and design data to calculate dimensions, tool selection, machine speeds, and feed rates.
37| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Manufacturing and CNC shops are traditionally slower adopters of advanced automation; digitization and AI integration are emerging but not mainstream. While some large aerospace/defense shops pilot AI-assisted CAM, the broader CNC programming sector remains largely manual or reliant on decades-old software. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing is a physical, moderately digitized sector where AI-driven CAM adoption is growing but still slower and shallower than in software/finance/professional services. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can assist by auto-extracting dimensional data from drawings, suggesting feed rates based on material and tool type, and flagging inconsistencies in specifications, thereby reducing manual calculation work. However, the human programmer remains essential for validating spatial constraints, resolving design conflicts, and signing off on safety-critical parameters. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI-enhanced CAM software significantly speeds up tool selection, feed/speed calculations, and toolpath generation, meaningfully boosting programmer productivity while the human remains responsible for validation and judgment calls. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can extract and parse technical documents (drawings, blueprints, specifications) and perform calculations for dimensions and feed rates, the task requires integrated spatial reasoning, tool selection judgment under constraint, and validation against domain expertise. Current AI systems struggle reliably with end-to-end conversion from complex engineering drawings to production-ready programs without significant human review, falling short of the 50% time-saving threshold. |
| Task automatability | claude-sonnet-5 | 3/5 | CAM software with AI-assisted feature recognition can partially automate dimension calculation, tool selection, and speed/feed determination, but interpreting varied blueprints, specs, and PCB films still requires human validation for many non-standard jobs.rating reflects partial but not full automation. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | CNC programming has liability and safety implications—faulty tool selection or feed rates can damage machines, harm workers, or produce defective parts. Manufacturing environments have organizational risk aversion, quality certification requirements, and strong human accountability expectations, all of which create friction against full automation without a licensed programmer's sign-off. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement mandates a human specifically, but liability for tooling errors, machine damage, and costly production mistakes creates strong organizational incentive for human verification. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI document parsing and calculation tools require human oversight, integration into workflows, and validation, making the all-in cost per task comparable to or exceeding the loaded wage of a skilled CNC programmer who performs these tasks as part of their routine. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Software licenses and setup costs are significant, and the task still requires a skilled programmer to verify and adapt outputs, so cost savings versus a human programmer are moderate rather than dramatic. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably performs the full analysis pipeline (reading heterogeneous documents, calculating tool selection, determining optimal speeds and feeds) in production CNC shops. Document understanding tools exist, but they require heavy manual verification and rework; the task remains largely manual or supported by domain-specific (often proprietary) legacy software. |
| Technical feasibility today | claude-sonnet-5 | 3/5 | CAM/CAD-integrated tools (e.g., Fusion 360, Mastercam with adaptive toolpath and feeds/speeds calculators) are in production use, but they handle only structured/standardized inputs reliably; interpreting varied blueprints and PCB films still needs skilled human oversight. |
Revise programs or tapes to eliminate errors, and retest programs to check that problems have been solved.
32CI 25–39 · exposure 33 · augmentation 63 · importance 4.2/5 · click for rater detail
Revise programs or tapes to eliminate errors, and retest programs to check that problems have been solved.
32| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | CNC shops remain relatively traditional and fragmented; digital maturity and AI adoption vary widely. While larger manufacturers explore automation, most small and mid-sized shops rely on manual programming and testing, making sector-wide adoption slow and incomplete. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing is a lower-digitization sector with slower AI adoption; CAM/CNC software incorporates incremental automation but production-floor AI-driven programming remains uncommon. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI can provide useful assistance through automated error detection, suggesting fixes for common syntax issues, and flagging potentially problematic code patterns, allowing programmers to focus on semantic validation and machine-specific testing rather than line-by-line manual review. |
| Augmentation potential | claude-sonnet-5 | 4/5 | AI coding assistants and simulation tools can meaningfully speed up error detection and code revision, letting programmers focus verification effort on physical test runs. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | While AI can assist in identifying syntax errors and suggesting code fixes, the task requires understanding intent and context to eliminate semantic errors that affect machining outcomes. Current AI tools lack sufficient domain knowledge of CNC operations and machine-specific constraints to autonomously revise and retest programs to production quality without human oversight. |
| Task automatability | claude-sonnet-5 | 3/5 | AI can assist in debugging CNC G-code and identifying syntax or logic errors, but validating physical machining results, tool paths, and real-world test cuts still requires hands-on verification and machine access.' |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | CNC programming carries high error-cost asymmetry—mistakes can damage expensive tooling, waste material, or create safety hazards. Organizational practice and customer liability expectations typically require a licensed programmer to review and sign off on revised programs before production use. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement, but safety-critical consequences of program errors (tool crashes, scrapped parts, injury risk) create strong organizational incentives for human verification before production runs. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 2/5 | Current AI-assisted debugging tools have significant licensing and integration costs, and the need for expert human oversight to validate revisions negates most cost advantage over a skilled CNC programmer performing direct revision and testing. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | AI-assisted code review is cheap, but the retesting phase requires physical machine time, material, and human oversight, keeping overall costs comparable to skilled programmer labor. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | Code linters and static analysis tools exist for CNC programs, and AI-assisted code review is emerging, but no mature deployed product reliably revises CNC programs end-to-end and validates them through retesting against actual machine behavior. Most solutions remain narrow, research-stage, or require heavy human verification. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | Some CAM software includes simulation and error-checking tools, and LLMs can review code for syntax issues, but no widely deployed product autonomously revises and retests CNC programs against real machine outcomes. |
Enter coordinates of hole locations into program memories by depressing pedals or buttons of programmers.
31CI 19–44 · exposure 33 · augmentation 38 · importance 4.0/5 · click for rater detail
Enter coordinates of hole locations into program memories by depressing pedals or buttons of programmers.
31| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | CNC programming is a legacy, low-digitization domain with aging equipment; the sector has not adopted AI agents for this task, and the workforce is small and traditional, making adoption slow. |
| Sector adoption velocity | claude-sonnet-5 | 2/5 | Manufacturing and machining sectors adopt AI more slowly than information/professional services, with automation focused on CAM software rather than direct physical interface replacement. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist by generating or verifying coordinate lists externally, but the actual pedal/button data entry is mechanical and routine, offering limited opportunity for meaningful augmentation that would transform productivity while keeping the human in the loop. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-assisted CAM tools can pre-calculate and suggest coordinate values, reducing manual entry errors and speeding up the programmer's task, though the physical entry step itself still requires human action. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | The task involves manual coordinate entry via pedals or buttons, which is hardware-specific and operator-dependent. While AI could theoretically generate coordinate data, the physical pedal/button interaction and integration with legacy programmer hardware creates significant friction that prevents a clear 50% time-saving path end-to-end. |
| Task automatability | claude-sonnet-5 | 3/5 | Entering coordinates into program memory via pedals/buttons is a rote data-entry action that CAM software and modern CNC controllers already largely automate through direct programming, but the specific manual pedal/button entry described is a legacy physical interface not easily replaced end-to-end by generative AI alone.</br>Given equipment constraints, only partial time savings are realistic without hardware/software integration. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strong barriers exist: CNC programming often requires operator certification and sign-off on safety/tolerance, and the physical hardware specificity means only authorized personnel can modify program memory. Equipment liability and the need for human verification of critical coordinate data add regulatory friction. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement, but organizational and hardware inertia (proprietary controller interfaces, machine-specific programming pedals) creates moderate friction to automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The cost of building custom hardware interface software and integrating it with legacy programmer systems would exceed the wage cost of a programmer entering coordinates manually, especially given the low volume and specialized nature of the work. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Replacing this narrow physical interaction requires custom integration or machine retrofitting, which is costly relative to a human operator simply pressing buttons as part of a broader workflow. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No mainstream AI product today reliably interfaces with CNC programmer hardware to automate pedal/button input sequences. The task involves legacy equipment and hardware-specific protocols that current general-purpose AI systems cannot operate without custom engineering. |
| Technical feasibility today | claude-sonnet-5 | 2/5 | While CAM/CAD systems can generate toolpaths and coordinates automatically, products that directly replace the physical pedal/button data entry step on older or specialized machines are not widely deployed; this remains a machine-specific, semi-manual interface. |
Align and secure pattern film on reference tables of optical programmers, and observe enlarger scope views of printed circuit boards.
21CI 14–28 · exposure 16 · augmentation 25 · importance 3.3/5 · click for rater detail
Align and secure pattern film on reference tables of optical programmers, and observe enlarger scope views of printed circuit boards.
21| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 1/5 | CNC tool programming is a specialized, declining domain with limited digitization momentum; manufacturers in this niche are typically small, conservative, and rely on legacy equipment with low incentive to automate micro-level operator tasks. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Manufacturing tasks involving physical setup of legacy optical programming equipment are in a low-digitization niche with minimal reported AI adoption or displacement. |
| Augmentation potential | claude-haiku-4-5-20251001 | 2/5 | AI could assist by analyzing enlarger scope images or providing alignment guidance overlays, but the core task of manual film positioning offers limited scope for meaningful productivity augmentation via current AI tools. |
| Augmentation potential | claude-sonnet-5 | 2/5 | AI-based machine vision could assist in verifying alignment or detecting defects via the scope view, offering some incremental support, but current tools don't materially transform this specific manual task. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Only parts of this task could be automated—observing enlarger scopes via computer vision is partially feasible, but physically aligning and securing pattern film on optical equipment requires dexterous manipulation and precise spatial positioning that current AI systems cannot reliably perform end-to-end. |
| Task automatability | claude-sonnet-5 | 2/5 | This involves physical manipulation of film and alignment on hardware plus visual inspection through an optical scope, which requires manual dexterity and physical presence that current AI cannot perform end-to-end.atibility with software alone.The task is inherently physical-manual, not just cognitive. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | This task requires direct physical manipulation of precision optical equipment and immediate tactile feedback; it is inherently human-contact dependent and involves high error costs (misalignment ruins PCBs), creating strong organizational and technical barriers to substitution. |
| Adoption barriers | claude-sonnet-5 | 2/5 | No licensing requirement is typical, but the task requires physical dexterity and equipment-specific setup, creating practical friction to automation even though no strict regulatory barrier exists. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | The hardware and integration costs for an optical manipulation system capable of aligning and securing film would exceed the loaded wage of the skilled technician performing this specialized, low-volume task. |
| Cost vs. human wage | claude-sonnet-5 | 2/5 | Automating this would require custom robotics/vision integration with machine vision and precision actuators, which is costly relative to a technician performing this narrow manual task, making AI not clearly cheaper. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 1/5 | No deployed AI product reliably performs the combined mechanical assembly and visual inspection workflow required here; this remains a hands-on, equipment-specific task lacking production-grade automation solutions. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs this specific physical alignment and optical observation task for optical programmers in PCB manufacturing; this is a legacy manual process with no robotic vision-actuation product in production for it. |
Perform preventative maintenance or minor repairs on machines.
16CI 5–26 · exposure 13 · augmentation 50 · importance 3.7/5 · click for rater detail
Perform preventative maintenance or minor repairs on machines.
16| Dimension | Model | Rating | Rationale |
|---|---|---|---|
| Sector adoption velocity | claude-haiku-4-5-20251001 | 2/5 | Manufacturing and CNC operations are moderately digitized but adoption of autonomous maintenance automation is slow due to risk aversion, legacy equipment, and the need for specialized physical intervention that AI cannot yet provide at scale. |
| Sector adoption velocity | claude-sonnet-5 | 1/5 | Manufacturing maintenance is a physical, low-digitization task category where AI/robotic adoption for hands-on repair work remains minimal and slow-moving. |
| Augmentation potential | claude-haiku-4-5-20251001 | 3/5 | AI augments this task moderately by supporting predictive maintenance through sensor data analysis, identifying failure patterns, and recommending maintenance schedules; however, the human technician remains essential for diagnosis confirmation and actual repairs. |
| Augmentation potential | claude-sonnet-5 | 3/5 | AI-driven predictive maintenance software and diagnostic tools can flag issues or recommend fixes, helping technicians plan and prioritize repairs, though the physical execution remains manual. |
| Task automatability | claude-haiku-4-5-20251001 | 2/5 | Preventative maintenance and minor repairs require physical manipulation, real-time problem diagnosis, and component replacement—capabilities current AI systems lack without robotics. While AI can support diagnosis via data analysis, the end-to-end task remains dependent on human technicians for 60-80% of execution time. |
| Task automatability | claude-sonnet-5 | 1/5 | This is a physical, hands-on task requiring manual inspection, lubrication, part replacement, and mechanical diagnostics on shop-floor equipment; current AI systems cannot physically perform maintenance or repairs. |
| Adoption barriers | claude-haiku-4-5-20251001 | 4/5 | Strong barriers exist: machinery manufacturers typically require certified technicians for warranty compliance, liability for equipment failure falls on the responsible party, and regulatory standards often mandate human sign-off on maintenance records and safety-critical repairs. |
| Adoption barriers | claude-sonnet-5 | 3/5 | No licensing requirement typically blocks this, but safety protocols, machine-specific expertise, and liability for equipment damage create meaningful organizational friction against unproven automation. |
| Cost vs. human wage | claude-haiku-4-5-20251001 | 1/5 | Current AI systems cannot reduce the cost of maintenance work because the physical and diagnostic labor remains human-dependent. AI-assisted tools may improve efficiency margins but do not achieve the cost advantage needed for substitution. |
| Cost vs. human wage | claude-sonnet-5 | 1/5 | There is no viable AI substitute for the physical labor involved, so any comparison favors the human technician who can actually perform the repair. |
| Technical feasibility today | claude-haiku-4-5-20251001 | 2/5 | No deployed product reliably performs full preventative maintenance or repairs autonomously; industrial robots exist for specific, controlled tasks but not for general diagnosis and repair. AI tools can support planning and diagnostics, but practical deployment is limited to narrow, pre-programmed scenarios. |
| Technical feasibility today | claude-sonnet-5 | 1/5 | No deployed AI product performs physical maintenance or minor repairs on CNC machines; robotic maintenance is research-stage at best and not in production for this use case. |
Related occupations — Production
How to read this
A high substitution score does not mean this job disappears — it means a large share of its current tasks face replacement pressure, so the mix of tasks is likely to change. High augmentation alongside substitution typically means the occupation reorganizes around the protected tasks. Wide confidence intervals mean the rater panel disagreed: treat those scores as open questions, not verdicts.
What would change this score
New model capabilities (automatability, feasibility), falling inference costs (cost ratio), regulation and licensing shifts (barriers), and measured sector adoption (velocity) all re-enter at every index release. Each release is recomputed, versioned and kept queryable — scores are claims with a date on them, not permanent labels.