Education in the AI era: when a degree is worth it and when it is not

Executive overview

Entry-level job postings in the US are down 35% since 2023. AI already handles 50–60% of typical junior tasks. The credential that once guaranteed a career path no longer guarantees employment.

The degree is not dead, but its monopoly is. Founders, CEOs, and AI researchers agree: the skills that survive are mastery of hard things, systems thinking, and the ability to teach yourself.

The real value of education is not memorising content — it is building the discipline to go deep, struggle, and come out stronger.

What is breaking in the job market

• Unemployment for recent graduates has reached 5.8%, one of the worst in years
• Companies like Salesforce and Shopify are hitting growth targets with AI instead of junior hires
• Entry-level tech hiring in Europe has collapsed by over 70%
• Workers with real AI skills now earn a 56% wage premium, up more than double in one year
• Universities still teach 5-year-old material while market demand shifts monthly

What founders and AI leaders say about going deep

• Aravind Srinivas (Perplexity AI): at 18, pick one direction and commit for 1–2 years — mastery takes time, not months
• Surrounding yourself with peers who push you matters as much as the subject itself
• Deep competence gives you something AI cannot: confidence to take on harder problems
• Samir Vasavada (skipped college, founded a billion-dollar company at 20): universities teach what to think, not how to think
• His alternative: find someone impressive you want to become in 10 years, do the work to get close to them, let credibility compound

What AI changes about knowledge itself

• Mustafa Suleyman (Microsoft AI): knowledge acquisition is becoming a conversation — AI tutors that never tire, explain the same idea 20 ways
• Knowledge is no longer scarce; the ability to learn from friction is
• Mike Krieger (Anthropic CPO, Instagram co-founder): "learn to code" was always about systems thinking, not Python syntax
• The skills that do not expire: curiosity, systems thinking, and teaching yourself hard things
• Suleyman's warning: if learning is always frictionless, children lose the ability to persist through difficulty

When a degree still makes sense

• Legally required credentials: medicine, law, certain engineering roles
• If you can afford it without life-crushing debt
• If you treat it as an accelerator — research projects, mentors, real network, things you build

When a degree is a bad deal

• Going because everyone else goes, with no clear plan
• Graduating with six-figure debt and no monetisable path
• Fields where bootcamps, apprenticeships, open-source work, or self-study get you there faster and cheaper

Five-step roadmap for ages 17–26

1. Pick a 10-year direction. Not a plan — a draft. What problems do you enjoy? What does a good workday look like? Write it down.
2. Reverse-engineer people already there. Look at what they actually do: projects, tools, responsibilities. Find the patterns.
3. Choose the fastest path to those skills. Degree (slower, broader), bootcamp (faster, focused), self-study (flexible, needs discipline), apprenticeship (paid learning). Ask what gets you the skills and a portfolio fastest at a price you can handle.
4. Build proof of work, not just a CV. Projects, code, designs, case studies, content — anything that proves you can ship.
5. Use AI every single day, but keep friction. Let AI explain, quiz, and review. Struggle with hard things yourself. AI is the co-pilot, not the replacement.