David Ondrej began his online career making gaming videos in the Czech Republic, earning a steady USD 20,000 each month. Everything changed when ChatGPT launched at the end of 2022. Overnight he saw that large-language-model technology would reshape every knowledge job. In April 2023 he took a “think week” in Copenhagen, shut down the gaming channel, and dove into artificial-intelligence content—even though his income fell to barely USD 600. That bold pivot, he says, “was the right directional bet”: being on the cutting edge of AI would open more doors than any short-term cash grab.
When Alex Hermoszi acquired the community platform Skool, David spotted a way to monetise his growing AI audience. He entered the monthly “Skool Games,” finished in the top ten, and briefly earned strong coaching revenue. Yet he disliked chasing quick sales. Coaching brought in dollars, but it did not inspire him. He refunded clients, quit high-ticket courses, and resolved to build software that could compound over a decade rather than peak in a quarter.
Thirty days of ideation produced dozens of startup concepts. The winner became Vectal.ai, an AI-driven task-management tool that mixes the flexibility of Notion with a multimodel assistant. David coded the first version himself—even though he was not a professional developer—because he wanted to understand every technical decision before hiring engineers. Using models like Claude 3.5 for “vibe coding,” he shipped the product in two months instead of the six or eight months traditional development would have required. Today Vectal.ai has more than 60 000 sign-ups and just over USD 11 000 in monthly recurring revenue. Growth is slower than he hopes, but ownership remains one-hundred percent his.
David’s thesis is straightforward: at least through 2032 the world will run on partnerships between people and specialised AI agents. Repetitive, well-documented processes—anything that fits an SOP—will migrate to software agents that charge tiny usage fees. Humans will design the workflows, oversee edge cases, and exercise judgment. Micropayments on fast crypto rails will let agents buy proprietary data, pay other agents, or access APIs without a human credit card in the loop.
Large models can already generate hundreds of app interfaces, cold-email drafts, or video titles in seconds. The bottleneck is not output volume but selection: knowing which variant is actually good. That skill—taste—comes only from experience and iterative feedback. David refuses to outsource creative judgment to an LLM, whether he’s writing a YouTube title or choosing a front-end design. The future belongs to people who hone domain-specific taste while letting models handle the brute-force options.
He draws a clear line between two emerging styles of software creation.
David predicts a common pattern: AI will whip up the entire front end, while skilled developers hand-craft the secure back end. The result is faster experimentation without sacrificing reliability.
Google’s forthcoming agent marketplace, linked to Coinbase for real-time crypto payments, hints at a world where specialised agents hire one another. An HR agent could pull in a best-in-class background-check agent; a finance agent could rent a tax-filing agent on demand. Early movers who release narrowly focused, high-quality agents may capture enormous value, just as early SEO experts once dominated search rankings. David sees nothing in physics that prevents a single founder with a swarm of agents from running a billion-dollar company.
Sam Altman half-joked that anyone who is not spending three hours a day “skill-maxing” with the latest OpenAI models “won’t make it.” David agrees. Whether you are a salesperson, a marketer, or a developer, everyday practice with frontier reasoning models compresses the learning curve. His own routine includes daily deep-research sessions across Perplexity, Claude, GPT-5 Pro and Grok Code Fast to debug, brainstorm, and refine strategy.
David’s closing advice is blunt: many college majors and mid-career jobs are racing toward obsolescence. Hiding from that fact only delays the shock. Better to accept temporary discomfort—downsizing income, retraining in code, relocating to a tech hub—than to face a full-blown career collapse later. The universe does not care about sunk costs. Those who move first, cultivate taste, and partner with agents will ride the wave instead of being swept away.
You can watch David’s ongoing experiments on his YouTube channel David Ondrej and test Vectal.ai if you’re ready to manage work the agentic way. His story shows that betting on long-term technology, rather than quick profits, can pay dividends—if you commit, iterate fast, and keep your eyes on the horizon.