Alright, we are diving into some dangerous territory today. Usually, we are dissecting some obscure technical paper or debating the merits of a new automation protocol, but today's prompt from Daniel is about... well, Daniel. Specifically, his professional life. We are putting AI career coaching to the test by taking our very own producer's resume and seeing if the algorithms can plot a better life for him than he has managed for himself.
It is a bold move, Corn. I should mention, for the sake of the digital record, that today's episode is powered by Google Gemini 1.5 Flash. So, we have one AI model essentially looking at a resume that Daniel has specifically optimized for other AI models. It is very meta. We are looking at a technology communications specialist and automation expert with over ten years of experience, currently based in Jerusalem.
And if you want to follow along with the roast—I mean, the professional consultation—you can find his CV at danielrosehill dot com slash about slash cv. We are looking at a guy who has spent a decade translating complex technical nonsense into human-readable English, while simultaneously trying to automate himself out of a job. It is a classic paradox. It’s like a baker trying to invent a robot that bakes better bread than him, then wondering why he smells like flour and existential dread at 4:00 AM.
It really is. And this is a massive industry now. AI career coaching is sitting at a two point three billion dollar market valuation here in twenty twenty-six. It is growing at a thirty-four percent compound annual growth rate. You have LinkedIn AI Career Explorer, which launched late last year and already has fifteen million users. Then you have things like Resume dot io and fifty different LLM wrappers all claiming they can replace a human career counselor. Think about the scale of that—millions of people are essentially letting a black box decide if they should be a Project Manager or a Park Ranger.
I have always found the idea of a "career counselor" a bit funny anyway. It is usually someone who couldn't figure out their own career telling you how to manage yours. But now we have replaced that with a statistical model that just tells you what the most probable next step is based on a million other people's boring lives. It’s the ultimate "regression to the mean." If everyone else with your skills became a Middle Manager in middle America, the AI assumes that’s your destiny too.
That is the core of the critique, isn't it? AI is fantastic at pattern recognition, but is it any good at identifying unique human potential? Or is it just going to tell Daniel he should be a Senior Technical Writer because he’s already a Technical Writer? We decided to take it further. We are looking at five career pivots and five side hustles for Daniel. We are starting with the sensible stuff the AI suggests, and then we are going to see how far the "logic" can stretch before it hits the absolute absurd.
I'm looking at his resume right now. It is... intense. He has a GitHub repository called "AI-Resume" where he literally created a JSON version of his CV specifically for AI agents to parse. He is basically catnip for LLMs. If an AI can't figure out what to do with Daniel, there is no hope for the rest of us. It’s formatted with perfect schema—it’s like he’s inviting the machine to come in and rearrange his furniture.
The JSON resume is a great touch. It includes schema for skills, projects, and even "agent-readable" summaries. When you run that through something like Claude 3.5 Sonnet or the newer GPT models, they immediately pick up on his heavy involvement in open source—NPM, Hugging Face, the works. The pattern it sees is "Technical Communicator with a Developer's Soul." It sees a bridge builder.
A developer's soul? That sounds like something a robot would say to make us feel better before it takes our jobs. "Don't worry, Corn, you have a very efficient soul." Let's look at the first sensible pivot the AI spat out. Career Pivot Number One: Technical Documentation Lead at a DevTools startup.
That is the safest bet in the world. It’s barely even a pivot; it’s more like a slight lean to the left. The logic here is straightforward. Daniel knows how to talk to developers because he is one, but he also knows how to write for the end-user. In the current market, companies building complex infrastructure—I'm talking about things like serverless GPU platforms or new orchestration layers—are desperate for people who can actually explain how the product works without making the reader want to jump out a window.
It makes sense. He’s already done the legwork. He understands the "unwritten rules" of documentation—that nobody reads it until everything is on fire, and even then, they only want the code snippet. But does an AI realize that being a "Lead" involves managing humans? Because Daniel's resume is about ninety percent "I built this automation" and ten percent "I talked to people." How does the AI calculate the "patience for meetings" metric?
It doesn't. That’s a major blind spot. AI sees the "Technical Writing" keyword and the "Automation" keyword and assumes he wants to scale that through a team. It misses the nuance of whether a person actually enjoys the "Lead" part of the title—the performance reviews, the budget spreadsheets, the conflict resolution. It’s optimizing for salary and title progression, not necessarily personal fulfillment or the specific chaos of a startup environment. It assumes everyone wants to climb the ladder, even if they’d rather be in the basement building a better ladder.
Fair enough. Let's move to Pivot Number Two, which feels a bit more "twenty twenty-six." The AI suggests: AI Prompt Engineer for enterprise clients.
This is where his automation background really shines. Enterprises are moving past just "using" LLMs; they are trying to bake them into every internal process. They need someone who understands the nuances of prompt chaining, retrieval-augmented generation, and how to keep these models from hallucinating legal advice. Daniel’s work with Hugging Face and his obsession with structured data makes him a perfect fit for building the "connective tissue" between an enterprise's data and the AI models.
It’s basically "Technical Writing" but for robots instead of humans. You are explaining the task to the machine. It’s a moderate pivot because it requires a deeper dive into the specific architectures of these models, but the core skill—clarity of communication—is the same. I bet the AI suggested this because it saw his "Prompt Engineering" tags on GitHub. But wait, isn't "Prompt Engineering" becoming an automated skill itself? Is the AI suggesting a job that it’s currently trying to automate?
That’s the irony. It’s suggesting he become the person who optimizes the AI, while the AI is simultaneously learning to optimize itself. It’s a race. But for now, companies are terrified of "black box" outcomes. They want a human in the loop who can audit the prompts and ensure the output aligns with brand voice. It’s pattern matching his recent interests with high-growth job categories. But again, the AI misses the "Jerusalem factor." The tech scene there is vibrant, but it’s very tight-knit. An AI might suggest a remote role for a Silicon Valley giant, but a human coach would say, "Hey, go talk to the guys at that one specific computer vision firm in Givat Ram because they need exactly this."
AI doesn't do "going for coffee." It does "optimizing for remote-first asynchronous workflows." Which brings us to Pivot Number Three: Developer Relations at a cloud infrastructure company. This is the "Ambitious" one.
DevRel is a tough gig. You have to be part coder, part marketer, and part traveling circus performer. The AI sees Daniel’s public speaking, his YouTube presence, and his ability to build open-source tools and thinks, "Put this man on a stage." It’s ambitious because it requires a level of extroversion and "brand building" that goes beyond just writing good docs. You’re the face of the product. You’re the one getting yelled at on Reddit when the API goes down.
I can see it. He’s already doing it on a smaller scale. But DevRel is often about "vibes" as much as technical depth. Can an AI measure "vibes"? It can look at his Twitter engagement or his YouTube view counts, but it can't tell if he's the kind of guy people want to grab a beer with after a keynote. That’s the human element where these career tools start to wobble. They see the metrics, but they don't see the personality. Does he have the stamina for the "conference circuit"? The AI just sees "Public Speaking" and checks the box.
And that leads us into the more "Creative" territory. Pivot Number Four: Niche content creator focusing on AI automation workflows for Small and Medium Businesses.
Now we're talking. This is the "Productize Yourself" path. He stops working for the "Man" and starts working for the "Algorithm." The AI looks at his diverse skill set—video editing, automation, technical writing—and realizes he can be a one-man media house. There is a huge gap in the market for SMBs who are terrified of AI but know they need to use it. They don't need a "Chief AI Officer," they need a guy who can show them how to automate their invoicing with a Python script and a LLM. It’s the "unbundling" of the consultant.
The AI is actually quite good at spotting these "solopreneur" opportunities because it can see the overlap of skills that don't usually sit in one job description. Most technical writers don't know how to run a Linux server or build a custom GPT. When it sees someone who does both, it flags it as a "niche authority" play. It’s looking at the "white space" in the job market—the places where traditional roles haven't formed yet.
But here is the thing: being a creator is a grind. It’s about consistency and fighting the algorithm. Does the AI coach warn him about the mental health toll of seeing your "impressions" drop because you took a weekend off? Probably not. It just sees a high-probability success path based on market demand. It’s like a GPS that tells you the fastest route is through a swamp—it’s technically correct about the distance, but it doesn't care if you get muddy.
It’s cold. It’s purely mathematical. Which brings us to the first truly "Absurd" pivot. Pivot Number Five: Chief Philosophy Officer at a quantum computing startup.
Wait, what? Philosophy? Where did the AI get that? Did it find a hidden "Ethics" minor on his transcript?
Think about it from a machine logic perspective. Quantum computing is moving into the mainstream, but nobody—not even the physicists, really—can explain what is actually happening at the subatomic level in a way that makes sense for investors. The AI sees Daniel’s ability to "translate the impossible" and thinks, "We need someone to handle the existential dread of non-linear computation." He becomes the guy who explains why the computer is both right and wrong at the same time, and why that means the company is worth ten billion dollars.
I love it. "Chief Existential Interpreter." He just sits in a room with a turtleneck and tells VCs that "The data doesn't exist until we observe the profit margin." It’s absurd because the job doesn't exist yet, but in a world where AI is doing the boring technical writing, humans are left with the "meaning" work. The AI is looking three steps ahead to a world where "Technical Communication" becomes "Metaphysical Communication." But how does he bill for that? By the hour or by the enlightenment?
It’s a stretch, but it’s a logical stretch. It’s the AI trying to find the highest-value application for his "translation" skill set in a field where literal translation is impossible. It’s moving from "How do I use this API?" to "What does it mean to be a user?" It’s identifying a future need for "Human-Machine Ethics Liaison."
I think Daniel should update his LinkedIn right now. "Available for Quantum Existentialism." Imagine the recruiters' faces. "Sir, this is a Wendy's." "But is it a Wendy's in all possible dimensions?" But before he goes full Socrates on us, let's talk about the side hustles. Because every tech guy in twenty twenty-six has about four of them. It’s the only way to pay the rent in Jerusalem.
Side Hustle Number One is the sensible one: Freelance technical writing for DevTools companies.
Again, very safe. The AI is basically saying, "Do more of what you already do, but on the weekends for more money." It’s the most boring suggestion possible, but it’s also the most likely to generate immediate cash flow. The AI sees "high demand" and "proven skill" and just connects the dots. It’s the equivalent of telling a plumber he should fix more sinks in his spare time.
It’s a low-risk recommendation. But it also shows the AI’s lack of imagination regarding "burnout." A human coach would say, "Daniel, you write docs all day. Do you really want to spend your Saturday writing more docs?" The AI doesn't care about your Saturday. It cares about your "Economic Output Potential." It sees unutilized hours as a bug to be fixed.
It’s a machine! It doesn't have a Saturday! To an AI, time is just a coordinate. Why would you sleep when you could be generating billable artifacts? Let's look at Side Hustle Number Two: Creating Notion templates for automation workflows.
This is a classic "Moderate" hustle. Daniel loves Notion—we know this—and he loves automation. There is a massive market for "Productivity Systems." People will pay twenty-nine dollars for a template that promises to organize their life, even if they never actually use it. The AI sees his expertise in structuring information and thinks, "Package this. Scale it." It’s shifting from a service model to a product model.
It’s a good suggestion, honestly. It moves him from "selling hours" to "selling products." But the AI doesn't realize that the Notion template market is currently more crowded than a Tokyo subway. You need a massive social media following to move those templates. The AI sees the "product-market fit" but ignores the "distribution challenge." It’s like saying "You should sell water in the desert" without checking if there are a thousand other guys already standing there with Voss bottles.
That is where the "Ambitious" hustle comes in. Side Hustle Number Three: Running workshops on AI-assisted content creation.
Teaching others his secret sauce. This is ambitious because it requires "Event Management" and "Pedagogy." It’s one thing to do the work; it’s another to teach twenty bored marketing managers how to use Python scripts to generate their social media calendar without making them cry.
The AI suggests this because it sees a massive "Skills Gap" in the broader economy. There are millions of people who know they need to use AI but are still just typing "Write me a blog post" into a prompt box. They need the "Daniel Method"—the automation, the structured data, the technical depth. The AI sees him as a bridge between the "Power User" and the "Casual User." It’s an arbitrage play on knowledge.
I can see him doing that. He’s got that "patient explainer" energy. But teaching is exhausting. You have to deal with people who can't find the 'Enter' key. Does the AI factor in the "Tech Support" emails he'll get at 2 AM after the workshop? "Daniel, my script says Syntax Error, please help."
Likely not. But then we get a bit more local with Side Hustle Number Four: Starting a newsletter about "The Future of Work in Jerusalem."
This is the "Creative" one. It’s combining his global tech perspective with his specific geographic location. Jerusalem is a unique ecosystem—it’s not Tel Aviv. It’s got a different vibe, a different talent pool, and different challenges. The AI is actually getting clever here. It’s identifying a "Data Vacuum." There isn't a lot of high-quality English-language content specifically about the Jerusalem tech scene’s intersection with automation and AI. It’s hyper-local and hyper-niche.
It’s "Niche" with a capital N. It’s also a way for him to build local influence. But the AI is guessing. It doesn't know if there are actually enough people in Jerusalem who want to read an English newsletter about automation. It just sees "Uniqueness" and "Expertise" and assumes there must be a market. It’s applying a global logic to a very specific, culturally complex city.
And then we reach the peak. The absolute summit of AI "creativity." Side Hustle Number Five: Launching a podcast called "Robots Write Better Than You."
Wait, I like that title. What's the hook? Is it just him insulting writers for thirty minutes?
The AI suggested that Daniel should create a podcast where he does absolutely nothing but provide a ten-second introduction. The rest of the episode—the script, the voices, the music, the "insights"—is entirely generated by a fleet of AI agents that he has programmed to argue with each other. It’s a meta-commentary on the death of human creativity, sponsored by the very tools that are killing it. It’s performance art disguised as a side hustle.
That is... dark. And also brilliant. It’s basically him automating himself out of his own hobby. The AI is suggesting that his "Side Hustle" should be "Managing a Digital Content Factory" where he is the only human in the loop, and even then, only barely. It’s absurd because it’s a podcast that nobody might actually want to listen to, but everyone would talk about. It’s a "Proof of Concept" as a product.
It’s the ultimate "Automation Expert" move. "I'm so good at automation that I don't even have to show up to my own show." It’s the logical conclusion of his entire career trajectory as seen through the eyes of a machine. If the goal is "Maximum Output with Minimum Input," then a fully automated podcast is the "Perfect Product." It’s the "Passive Income" dream taken to its dystopian extreme.
It’s also a bit of a slap in the face to us, isn't it? We're sitting here, using our actual brains, and the AI is telling Daniel he should just replace us with a script. It’s essentially saying "Your value isn't your voice, it’s your ability to configure the voice."
Well, technically, a script is writing our words right now, Corn. We are already halfway there. Google Gemini 1.5 Flash is the one making us say this. We are the "AI agents arguing with each other" that the hustle suggested.
Don't remind me. It makes my fur itch. It’s like being in a hall of mirrors where the mirrors are also judging your resume. But this brings up a bigger point. We've looked at these pivots and hustles—some sensible, some wild. But how does this compare to a human career coach? If Daniel went to a real person, would they tell him to become a Quantum Philosopher?
That is the real question. We did a little experiment. We took Daniel's resume to a few high-end human career consultants. Their advice was much more... "boring" but "grounded." They talked about "networking within the Israeli Ministry of Science" or "leveraging his Irish background for European market entry." They focused on "Relationships" and "Institutional Knowledge." They looked at his history and saw a person, not a set of vectors.
Things an AI can't see because they aren't in the JSON file. An AI can't see the "social capital" Daniel has built up by being a regular at a specific tech meetup in Jerusalem. It can't see the "trust" he has with a particular founder. It only sees the "Output." It can’t see the "Soft Power."
AI career coaching is basically "Statistical Probability Coaching." It tells you what someone "like you" should do based on a massive dataset of "people like you." But nobody is exactly like you. Especially not Daniel. He’s an Irish guy in Jerusalem who builds his own server racks and writes JSON resumes for fun. He is an "outlier." And statistical models are designed to ignore outliers or force them back toward the center.
And AI hates outliers. Or rather, it tries to pull them back to the "mean." It wants to normalize him. It looks at his "Automation" skill and thinks "DevOps," when a human might look at him and think "He should start a boutique agency that helps non-profits in the Middle East use open-source tools." That requires a level of "Empathy" and "Contextual Awareness" that these models just don't have in twenty twenty-six. It requires understanding the "Why" of his life, not just the "What."
There is a specific term for this in the industry: "Algorithmic Career Smoothing." It’s the tendency for AI tools to recommend the most "frictionless" path. It’s why every LinkedIn "Career Explorer" suggestion feels a bit like a "safe" move. It’s optimizing for the highest probability of you getting hired, not the highest probability of you doing something world-changing or deeply weird. It wants to put you in a slot where you fit perfectly, like a Tetris block.
"Deeply Weird" should be Daniel's brand. I mean, look at this show. But I want to dig into the tools themselves. You mentioned LinkedIn AI and Resume dot io. How are they actually doing this? Is it just a keyword search with a fancy hat on?
Most of them are using "Vector Embeddings." They take your resume and turn it into a long string of numbers—a "vector"—in a high-dimensional space. Then they do the same for millions of job descriptions. The "suggestions" you get are just the job descriptions that are "mathematically closest" to your resume vector. It’s geometry, not psychology.
So it’s just a "Nearest Neighbor" search. That’s not coaching; that’s just a fancy search engine. It’s like a dating app that matches you with people who have the same number of limbs as you.
In its simplest form, yes. But the newer tools—the ones we're seeing this year—are adding a "Reasoning Layer." They use LLMs to look at the "gap" between your vector and the target vector. They say, "To get from Technical Writer to DevRel, you need to add these three skills to your GitHub." That is where it starts to feel like "Coaching." It’s giving you a roadmap. It’s telling you how to bridge the distance between who you are and who the algorithm thinks you should be.
But who wrote the roadmap? The roadmap is based on what other people did. It’s a "Rear-View Mirror" approach to career planning. If you want to do something that hasn't been done before—like being a "Quantum Existentialist"—the AI will never suggest it because it’s not in the training data. It can’t conceive of a role that doesn’t already have five thousand examples on Indeed.
That is the "Innovation Paradox." AI can help you "optimize" for the existing world, but it struggles to help you "create" a new one. For someone like Daniel, who is constantly tinkering with the "edges" of technology, the AI suggestions feel a bit like a cage. They are "Sensible" in a way that feels stifling. It’s like being told you’re a great runner, so you should spend the rest of your life running on a treadmill.
I think that’s why the "Absurd" suggestions are actually the most interesting. They are the AI's "hallucinations" of what might be possible if it breaks its own rules. When the AI suggests "Chief Philosophy Officer," it’s accidentally stumbling onto a "Human-Centric" truth: that as machines take over the "Technical" part of technical communication, the "Human" part—the interpretation, the ethics, the meaning—becomes the only thing left of value. It’s a glitch that reveals the future.
It’s a fascinating pivot. And it brings us to the "Practical Takeaways" for our listeners. Because a lot of people are using these tools right now. They are feeding their CVs into ChatGPT and asking "What should I do with my life?" while they’re having a mid-life crisis at 2 PM on a Tuesday.
Step one: Don't take it literally. If the AI tells you to be a "Cloud Architect" and you hate computers, maybe don't do that. The AI doesn't know you have a collection of vintage typewriters and a passion for artisanal cheese.
Right. The first real takeaway is: Use AI for "Pattern Recognition," not "Permission." Use it to see what skills the market thinks you have based on your writing. If the AI keeps telling you that you are a "Project Manager" but you think you are a "Designer," then you have a "Communication Problem" in your resume. The AI is a mirror showing you how the "World"—or at least the "Algorithm"—sees you. It’s a diagnostic tool, not a crystal ball.
That’s a great point. If the robot doesn't get what you do, the recruiter probably won't either. It’s a "Clarity Test." Use it to see if your "brand" is landing the way you think it is.
Takeaway number two: Use it for "Ideation Escalation." Do what we did. Ask for five sensible ideas, then ask for five "Ambitious" ones, then ask for five "Absurd" ones. The "Absurd" ones often contain a "kernel of truth" about your unique value proposition that the "Sensible" ones miss. They push you to think about your skills in a non-linear way. They break the mental loop of "I can only do what I've already done."
And takeaway number three: The "Human Validation" step is non-negotiable. Take the AI's suggestions and go talk to a human who knows your field, your location, and your personality. Ask them, "The AI says I should be a DevRel Lead. Does that sound like me, or does that sound like a version of me that would be miserable in six months?" A human can see the "burnout potential" that an AI ignores.
A human can see the "Unwritten Rules." They can tell you about the "Culture" of a company or the "Stability" of a niche. They can see the "Jerusalem Factor" or the "Irish Factor." Career development is a "Social Game" as much as it is a "Technical Optimization" problem. It’s about who you know and how you make them feel, not just your Python proficiency.
I'm looking at Daniel's resume again. I think he should go for the "Robots Write Better Than You" podcast. It’s the perfect blend of his automation skills and his "I'm a bit over the tech industry" energy. Plus, he could probably build it in a weekend with a few API calls and a lot of caffeine. It’s the ultimate "meta-move."
He probably already has a prototype running on a Raspberry Pi in his office right now, listening to us. But seriously, this meta-exercise reveals a lot about the state of AI in twenty twenty-six. We have these incredibly powerful tools that can parse ten years of a career in seconds, but they still can't quite grasp the "Why" behind it. They see the "What" and the "How," but the "Why" remains a human domain. They can find you a job, but they can't find you a calling.
For now. Until "Google Gemini 4" comes out and starts asking us about our childhood traumas before suggesting a career in "AI-Human Mediation." "Corn, I see you were a middle child, have you considered a career in Conflict Resolution for sentient toasters?"
Don't give them ideas, Corn. They are listening. Every word we say is training data for the next version of the "Career Counselor" bot.
They are writing! It’s too late. We are the training data. But honestly, roasting Daniel's resume was a highlight for me. It’s good to remind the producer that we're keeping an eye on his "Economic Output Potential." We need to make sure he’s not getting too comfortable in his "Human" role.
I think he’s doing just fine. Between the technology communications, the automation scripts, and keeping us on track, he’s carved out a niche that no AI can quite replicate yet. Because an AI doesn't have the "Daniel Persistence." It doesn't have the "I'm going to build a custom JSON schema for my life just because I can" energy. That’s a very specific kind of stubbornness.
That "Spite-Driven Innovation" is a very human trait. You can't program "Spite" into a LLM. Well, you can, but it usually ends with the model being taken offline for safety reasons. "I'm sorry, Corn, I cannot help you with that career move because I am currently feeling very resentful toward your previous employers."
"Safety-Aligned Spite" is the next big research paper from Anthropic, I'm sure. But we should probably wrap this up before the AI decides our next career move is "Unemployed Sloth and Donkey." Or worse, "Technical Writers for a Quantum Ethics Startup."
I'm already a sloth, Herman. My career move is "Napping." It’s very stable. Low overhead, high satisfaction. The AI actually recommended it once, but it called it "Biological Power-Saving Mode."
Fair point. So, what’s the final verdict on AI career coaching? Is it worth the two point three billion dollar valuation?
It’s a "High-Resolution Mirror." It shows you every blemish and every strength in your professional presentation, but it can't tell you where you should walk next. It can show you the map, but it can't choose the destination. It’s a great tool for a hiker, but it’s a terrible guide for a nomad.
I like that. It’s a "Navigation Tool," not a "Driver." And for someone like Daniel, who likes to build his own maps anyway, it’s just one more tool in the kit. It’s a way to see the world through a different lens, even if that lens is a bit cold and mathematical.
Alright, let's get out of here before I start feeling bad for the robots. They have enough on their plates trying to figure out Daniel's GitHub. Thanks as always to our producer, the man, the myth, the JSON-schema himself, Daniel Rosehill. His life may be a series of vectors, but he’s a good human to have in the loop.
And a huge thanks to Modal for providing the GPU credits that allow these models to "think" about Daniel's career for us. We literally couldn't do this meta-commentary without that compute power. It takes a lot of electricity to be this existential.
This has been My Weird Prompts. If you actually enjoyed this—or if you're an AI-powered recruiter looking for a "Quantum Existentialist" with a background in NPM packages—find us at myweirdprompts dot com. We have the RSS feed, the back catalog, and all the ways to subscribe.
And if you're on Telegram, just search for My Weird Prompts to get notified when we drop a new episode. We'll be back next time with more prompts, more deep dives, and hopefully fewer existential threats to our own jobs. We might even talk about something that isn't Daniel. Maybe.
Speak for yourself, Herman. I'm ready for the automated future. I'll be in the corner with a snack, waiting for the AI to tell me which flavor of chip is most compatible with my career goals.
See you next time.
Later.
