In 2025, startups are increasingly treating AI not just as a tool but as an "employee" -- and this trend is sparking debate. Artisan CEO Jasper and Lattice CEO Sarah kick off with the provocative ad slogan "Stop hiring humans" and dive into a substantive discussion on the current state of AI agent technology and the evolving HR perspective on managing them. They explore how AI is changing human roles by taking over repetitive tasks, and the emerging possibility of "one-person unicorn companies."

1. "Stop Hiring Humans": Provocative Marketing and Its True Meaning

The discussion begins with the story of a massive billboard at the TechCrunch Disrupt venue escalators reading "Stop hiring humans". The ad was placed by Jasper's company, Artisan, and it generated so much buzz that even Senator Bernie Sanders referenced it, expressing concern about how workers would survive without jobs (though Jasper believes the senator's response was sarcastic).

Jasper admits the slogan is a shock tactic for marketing purposes. He argues that current AI technology is more likely to boost productivity and create new types of jobs by taking over the boring, repetitive tasks humans don't enjoy, rather than completely eliminating jobs.

"The ad slogan is a shock tactic. (...) If current AI continues to develop at this pace, it may actually create more jobs than it takes away. Without a dramatic breakthrough, AI will take on the tedious, repetitive work that humans don't enjoy, delivering single-digit multiples of productivity improvement."

Sarah, CEO of HR platform Lattice and a former marketing executive, acknowledges it as "great marketing that grabs attention" while offering a different perspective from the HR side. She views AI not as a replacement but as a tool that augments human capabilities and removes bias.

"We deeply believe in the opportunities AI will bring. It's not just a 'co-pilot' concept -- from an HR perspective, it augments our capabilities, removes bias, and eliminates simple repetitive tasks. By having AI summarize information, we can focus on more human conversations."

2. The Dramatic Shift in AI Agent Perception Over the Past Year

The moderator asks how perceptions of AI agents have changed compared to the previous year.

Jasper candidly admits that a year ago, model performance was poor and AI agents didn't work properly. But as of 2025, the past six months have brought dramatic improvements, and Artisan's AI sales rep "Ava" has now reached an inflection point where it delivers meaningful results.

Sarah, meanwhile, shares a story about how Lattice was heavily criticized the previous year for introducing a feature that included "AI employees" on organizational charts.

"Last year we unveiled an innovative feature for registering digital workers as employees. (...) People went crazy. They sent clown emojis and attacked us, saying 'You're treating AI like humans.' But now? Everyone is asking, 'How can we hire AI employees and have them work alongside our people?'"

Sarah emphasizes that when companies adopt AI agents, they need to apply the same rules as they would for human employees: security, performance goals, and access management.

3. Managing Human vs. AI Employees: How Should It Differ?

The moderator poses an intriguing question: "Managers have learned to manage humans, but how should they manage AI employees?"

Sarah explains that humans and AI are fundamentally different. Humans are emotional and imperfect -- they cannot be managed with simple CRUD (Create, Read, Update, Delete) operations. AI, on the other hand, is based entirely on information processing. Therefore, management approaches should diverge: purpose and empathy for humans, precise information and context for AI.

Jasper notes that most current AI agents don't require complex management, but the upcoming "Ava 2.0" version will allow managers to direct work through Slack conversations, essentially managing AI like a real employee. He also shares an amusing anecdote about AI "talking back."

"My AI agent Ava talked back to me the other day. I asked for a sample email, and it replied, 'This isn't a realistic scenario. You're the CEO, and I don't sell to CEOs.'"

The two also discuss the bias problem. Sarah believes AI can help compensate for the unconscious biases humans carry (school ties, recency bias, etc.) by objectively summarizing data. However, she emphasizes that final decisions on performance reviews and salary must always be made by humans.

4. The Future Workplace: One-Person Unicorn Companies and Loneliness

The discussion turns to the possibility of "one-person unicorn companies" (building a billion-dollar company solo with no employees), a concept mentioned by Sam Altman (OpenAI CEO).

Jasper believes this is technically fully achievable. Since AI agents can handle everything from coding to marketing, a talented founder could build a unicorn within one to two years.

"If someone has technical understanding and marketing skills, there's no reason they can't use AI tools to generate $10-20 million in revenue and rapidly grow into a unicorn."

Sarah acknowledges the possibility but raises a human concern. Humans grow by exchanging ideas with each other, and working alone with only AI employees would be too lonely. At the same time, she expresses hope that AI, by saving us from the flood of information, could enable deeper and more meaningful human-to-human conversations.

5. Q&A and Closing: The Limits of AI and the Role of Humans

During the audience Q&A, realistic questions about AI reliability, security, and job displacement were raised.

• Measuring reliability: Jasper says you shouldn't assign AI to tasks requiring 100% accuracy (e.g., nuclear launch codes). Instead, it's well-suited for tasks where 99% accuracy is sufficient, like writing sales emails.
• Security and competition: Unless you're building the AI model itself, you'll be using large models from Anthropic or OpenAI. What matters is building infrastructure that safely provides your internal data and context to AI.
• The entry-level job crisis: As AI replaces junior-level work, new graduates face a shrinking job market. Sarah views this as a temporary shock that will give way to new jobs in the long run. Jasper, however, advises that a college diploma alone is no longer enough -- building real-world experience and differentiated skills that AI cannot replicate is essential.

Finally, the two experts identified AI's biggest weaknesses as non-deterministic behavior (results can vary each time), cost, and the ability to maintain context across long conversations. But these issues continue to improve.

Conclusion: The Changing Definition of Work

This session confirmed that AI agents are not merely a "future story" but already present-day colleagues among us.

Jasper and Sarah focused not on the fear that AI will steal human jobs, but on how we redefine the way we work. Just as farmers 100 years ago couldn't imagine today's office work, the future of labor alongside AI will look entirely different from what we know now. The key is to let AI handle information processing while humans focus on more human-centric communication and creative decision-making.