The Shala Daily

Anthropics’s CEO just said half of all entry-level jobs will be gone by 2031. In a recent piece for Entrepreneur Loop, the Editorial Team unpacks Dario Amodei’s forecast—and the numbers are already proving him right. AI can now spot 500 security vulnerabilities in the time it would’ve taken a junior analyst months to find. It writes code, reviews contracts, builds financial models. All the work we used to hand to beginners.

The obvious worry is jobs vanishing. The deeper problem is what happens to the entire learning ecosystem. You don’t become an expert by watching expertise from the sidelines. You become an expert by doing mediocre work while someone who’s been there for decades watches over your shoulder. You learn through correction. Through repetition. Through the specific humiliation of screwing something up while someone kinder and wiser shows you the better way.

🪜 Every Rung Matters

In yoga, there’s svadhyaya—self-study, but also study *with* a teacher. The traditional model is parampara: an unbroken lineage of teacher to student, knowledge passed through direct transmission. Not because ancient yogis were gatekeeping, but because some things can’t be learned from a book or a prompt.

The distance between a 200-hour yoga teacher and someone who’s practiced for thirty years isn’t just mat time. It’s apprenticeship. Showing up day after day, making mistakes, receiving adjustments, absorbing not just the *what* but the *why* and the *when not to*. When we eliminate entry-level positions, we’re not just cutting costs. We’re severing the transmission line. We’re saying: figure it out alone, or give up.

🔥 You Can’t Microwave Mastery

The yogic concept of tapas—often translated as discipline or heat—means sustained effort over time. Transformation requires it. You can’t skip the boring parts, the repetitive parts, the parts where you’re just not very good yet.

Entry-level work is tapas. Unglamorous, tedious, essential. That’s where you build the neural pathways, the muscle memory, the intuitive understanding that separates someone who *knows about* a thing from someone who *knows* a thing. If AI does all the entry-level work, where does the next generation build their heat? Where do they develop the capacity to sit with difficulty, to persist through confusion, to gradually become competent?

🧭 Finding Your Path Takes Wandering

Your dharma—your unique path, your contribution—doesn’t arrive fully formed. It’s discovered through exploration, through trial and error, through the messy process of finding what you’re actually suited for.

Entry-level jobs were never just about the tasks. They were about exposure. About discovering whether you actually like legal work or if law school was a mistake. About learning you’re terrible at spreadsheets but brilliant at client relationships. About stumbling into unexpected passions because you were given access to the arena.

I don’t think we should slow AI advancement. That ship sailed. Some of this technology is genuinely useful. But we need to get serious about preserving the apprenticeship model. We need to understand that mentorship isn’t a nice-to-have. It’s infrastructure. It’s how civilization transmits itself forward.

In Ashtanga, we say: practice and all is coming. But practice requires a place to practice. It requires someone to show you how. It requires the patient, inefficient, deeply human work of transmission. If we automate away the bottom rungs of the ladder, we shouldn’t be surprised when no one can reach the top.