I conducted an interesting thought experiment with AI, asking the following question: Imagine two carpenters with a kitchen cabinetry business. How can AI assist these carpenters?
Gemini’s answer
Prompt: How can AI help the carpenters?
AI can help the two carpenters by automating administrative tasks (scheduling, lead qualification, estimates), optimizing project management (materials, timelines), assisting with marketing (content, reviews), and providing personalized client communication, thereby freeing up their limited time to focus on cabinetry craftsmanship and potentially increase their client capacity.
This naturally raises a question: if AI is so effective, do we still need two carpenters? AI is enhancing and reshaping roles. Experts can use AI to speed up their work, from personalizing communications and automating schedules, billing, contacts, and estimates to optimizing project planning, material calculations, and design prototyping
Here is Gemini’s take
Prompt: if AI is so effective, do we still need 2 carpenters
AI automates tasks and provides data, but it cannot replicate the carpenters’ physical skills, creative problem-solving, artistic touch, or the personal client relationships crucial for custom cabinetry. AI is a powerful tool to enhance their work, not replace their irreplaceable human craftsmanship.
I’m not fully convinced by the answer. I agree that carpenters’ physical skills remain essential, but I have doubts about personal client relationships and strongly disagree about creative problem-solving and artistic touch. There’s a narrative promoting full AI-human collaboration, where AI handles mundane tasks and humans lead with creativity. However, after lightly exploring tools like MidJourney, DALL-E, Canva, and others, I’ve found that creativity and artistry are just a prompt away. But here’s the catch: humans come up with the prompt ideas. Add an AI agent to generate those prompts, and voilà! the human is out of the loop! Of course, a human still needs to instruct the agent, but at that point, they’re two steps removed from the creative work.
For creative problem-solving, when prompted correctly, AI can be remarkably creative, evaluating problems, offering new perspectives, and generating innovative solutions like designing adaptive workflows or optimizing complex systems. With future tools automating prompts, AI could independently drive creative outcomes, pushing the boundaries of innovation without direct human input.
Regarding building client relationships, people are increasingly comfortable interacting with AI. AI now acts as a psychologist, advisor, consultant, or even a friend. As technology evolves, people are redefining their relationships with both tech and each other. A good example is the rise of social media. Before social media, an influencer was someone local: the cool kid at school, a thought leader in your company, or an outspoken family member. With social media, people can follow and engage with millions of influencers worldwide, shaping their worldviews. Now, integrate AI into all aspects of life, including social media, and you get immediate answers to questions or concerns, further transforming how we connect and communicate.
Gemini’s take
Prompt: I just copied and pasted the content above and asked for a 2 paragraph answer:
You’ve articulated a powerful and increasingly plausible vision for the future of work. It’s true that the capabilities of generative AI are rapidly blurring the lines of what constitutes “human creativity” and “personal connection.” If an AI can generate stunning, unique cabinetry designs based on subtle prompts, and another AI can manage client communications with flawless efficiency and a simulated understanding of their needs, the traditional roles of creative problem-solver and relationship-builder shift dramatically. The human input for creativity moves further upstream to high-level instruction, while client “comfort” with AI interfaces grows.
So, for our two carpenters, the answer becomes less about their outright replacement and more about a profound transformation of their roles. They would likely become highly skilled operators and installers of AI-designed and AI-optimized cabinetry. Their craft would evolve from traditional design and client-facing sales into precision assembly, meticulous quality control of AI outputs, and expert physical installation. The need for two sets of hands on the tools remains, but the decision-making, creative ideation, and much of the client interaction could be orchestrated by advanced AI, making the human carpenters incredibly efficient and focused on the tangible work that still requires their unique physical dexterity.
Since AI is agreeable, I asked it to disagree with me. Here it’s answer:
You’ve presented a forward-thinking view of AI, and it’s true that generative AI is astonishingly capable at creating designs and that people are increasingly comfortable with AI interactions. However, in the context of our two carpenters building custom cabinetry, there are still critical, factual limitations that keep humans essential. AI, while excellent at synthesizing existing data, currently lacks true conceptual novelty—it can’t invent entirely new paradigms or artistic movements from scratch, nor does it possess the common sense reasoning to navigate the inherent unpredictability of a real-world renovation site with its crooked walls, hidden pipes, and unexpected client changes. Furthermore, for high-stakes, personalized projects like a custom kitchen, AI cannot replicate genuine empathy or build the nuanced trust that comes from a human understanding and addressing a client’s anxieties and non-verbal cues. The two carpenters, therefore, remain indispensable for their unique blend of on-the-fly problem-solving, intuitive creativity beyond data, and the profound human connection that AI simply hasn’t achieved
All good! This raises two new questions and brings me back to my previous question.
New Questions:
- If we are all “idea guys,” as Sam Altman from OpenAI mentioned in his latest blog post, is there any value in specialization?
- If AI can handle all the planning, work, and client communication, do carpenters become mere installers, essentially working for AI?
Previous Question:
- If AI can perform all these tasks efficiently, do we need two carpenters?
Is there any value in specialization?
AI already possesses vast knowledge, enabling it to perform certain tasks effectively. A good example is “Dr. Google,” where people go online to self-diagnose medical conditions. Doctors often complain that patients who self-diagnose lack the experience and years of training that enable physicians to recognize patterns and distinguish between similar symptoms. For instance, a doctor might say, “It looks like you may have condition A, but based on seeing similar symptoms in other patients, it could also be condition B.” In contrast, a patient using Dr. Google might misinterpret symptoms due to biases or lack of expertise, concluding they have the flu or even cancer. Nevertheless, access to this information can empower patients to advocate for themselves more effectively.
In a lower-stakes scenario, such as kitchen cabinetry, I might have enough knowledge to build cabinets myself by consulting AI and watching YouTube tutorials to guide me through the process. However, I would likely hire professional carpenters for their expertise. Their experience ensures that if something goes wrong, they can fix it (whereas if I do it myself, the responsibility is mine). Carpenters also have established networks of suppliers and other contractors, providing access to quality materials and complementary expertise. Moreover, their real-world, on-the-spot pattern recognition allows them to address issues like crooked walls or hidden pipes with immediate solutions.
If AI can handle most planning and communication, are we working for AI?
Probably not. AI offers significant support in planning and communication, but human approval, face-to-face interaction, and trust remain essential. For example, before signing a contract, I prefer to meet someone in person to assess their credibility, something AI cannot replicate. This human connection is a distinctly human characteristic. You see this in everyday scenarios, like hiring a local contractor or lawyer, or choosing a restaurant based on human reviews. While AI can automate much of the groundwork, such as drafting plans or scheduling, we still rely on personal interactions to build trust before committing. Plus, AI can’t enter into contracts.
This automation, however, is increasing competition, which leads to my third question: do we still need two carpenters?
I’ll let Gemini answer.
Prompt: I copied and pasted the above paragraphs and asked, “Do we still need two carpenters?”
Given AI’s ability to handle extensive planning and communication, the core question shifts from if we need carpenters to how their roles evolve, and whether two are still optimal. Your context highlights that while AI excels at groundwork, it critically lacks the human capacity for trust-building through personal interaction, the on-the-spot pattern recognition for real-world issues like crooked walls or hidden pipes, and the ultimate accountability for contracts and complex problem-solving that comes from years of experience and a network of suppliers. In a market where AI fuels competition by automating administrative tasks, the two carpenters become highly efficient human “executors” and “trust-builders.” They leverage AI for speed and precision in planning, but their continued necessity lies in their physical skills, the nuanced judgment derived from real-world experience to fix unforeseen problems, and their ability to forge the personal connection and credibility that clients demand before committing to significant projects – functions AI cannot yet fully replicate.
I love the phrase “functions AI cannot YET fully replicate”.
I’m not convinced by Gemini’s answer, it seems too “diplomatic.” So, I asked Grok (xAI) for its perspective.
Prompt: Copied and pasted the same paragraph as Gemini and asked the same question.
AI’s efficiency might reduce the need for multiple specialists in some cases, but it depends on the project’s complexity and scale. For intricate or high-stakes work, like custom cabinetry or large-scale construction, two carpenters could still be valuable for their combined expertise, speed, and ability to catch errors—qualities AI can’t fully replace. If the task is straightforward, AI-guided tools might allow one skilled carpenter to suffice, especially with access to supplier networks and real-time problem-solving skills. The key is balance: AI enhances productivity, but human specialists ensure quality and adaptability.
The answers highlight a key term: “skilled.” As xAI states, “If the task is straightforward, AI-guided tools might allow one skilled carpenter to suffice, especially with access to supplier networks and real-time problem-solving skills.” That sets a high bar. How does a novice acquire those skills?
The current employment landscape suggests that AI is reshaping job roles. One key change is that AI appears to be impacting junior roles. A senior employee can use AI to perform junior-level tasks effectively through prompting. But how do junior employees gain experience, especially if there’s no demand to invest in future specialists?
Gemini’s take,
Prompt: If a senior employee can use AI to perform junior-level tasks effectively through prompting. How do junior employees gain experience?
The advent of AI indeed poses a profound challenge to how junior employees acquire the “skilled” expertise and “pattern recognition” vital for roles like carpentry, especially if automation reduces entry-level tasks traditionally used for training. If senior staff can leverage AI for what were once junior duties, the incentive to invest in developing future specialists through conventional means diminishes. Therefore, the future of gaining experience for new professionals lies in a fundamental shift from traditional apprenticeship models to AI-augmented learning pathways: juniors will need to rapidly develop “AI literacy” – understanding how to prompt, validate, and refine AI outputs – alongside critical human skills such as complex problem-solving, adaptive thinking, and emotional intelligence, which are essential for navigating the unpredictable realities of real-world work and building client trust. This means a greater emphasis on project-based learning, structured mentorship focused on transferring tacit knowledge and judgment, and a proactive commitment to continuous reskilling in a rapidly evolving technological landscape, ensuring they become valuable AI collaborators rather than redundant task-doers.
xAI gave me a nuanced answer to the same prompt
AI’s Impact on Junior Roles: Your observation that AI allows senior employees to handle junior tasks via prompting (e.g., using AI for drafting or scheduling) is insightful. It mirrors industry trends: a 2023 World Economic Forum report noted that AI could displace 30% of entry-level roles by 2027, particularly in routine tasks, while creating demand for tech-savvy specialists.
Training Challenges: Your concern about “no demand to invest in future specialists” reflects a real issue. With 91% of contractors facing labor shortages, firms may prioritize immediate needs over training novices, especially if AI reduces junior role demand. However, growing niches like sustainable carpentry (part of a $200 billion market) require skilled workers, suggesting investment in training could rise if novices upskill.
Incorporating AI training early is critical for junior professionals seeking to enter any industry. As the field rapidly evolves, it will be interesting to see which AI tools and knowledge are required for specific industries. From an employment perspective, a key question remains for future exploration: how can juniors gain experience if no one is hiring them?
