Apple's secrecy and reluctance to publish research on AI hurt its ability to hire best researchers, some say
Apple's Deep Learning Curve — The company's secrecy is hurting its AI software development. — In the world of artificial intelligence, one of the year's biggest coming …
Context & Ripple Effects
In late 2015, at the peak of the deep learning hiring frenzy, Bloomberg flagged a problem unique to Apple among the big labs: its culture of secrecy meant researchers who joined couldn't publish, present, or build reputations — the currencies of the field. The reporting framed this as a direct drag on recruiting top AI talent.
The decade since reads as a vindication of that thesis. Coverage traced organizational dysfunction inside Apple's Siri group in 2023, detailed in 2024 how the AI team failed to collaborate with product teams or get computing resources, and by early 2026 documented at least four more researcher departures to Meta and DeepMind — the exact rivals an open-lab culture feeds.
First-order effects
- Top deep learning candidates weighing offers face a tradeoff Apple's rivals don't impose: joining Apple means forgoing publication and citation credit, so Meta, DeepMind, and Google win the contested hires.
- Apple's existing researchers have weaker external visibility than peers at publishing labs, compounding retention pressure inside its Siri and machine learning groups.
Second-order effects
- As internal accounts of Apple falling behind accumulated — including [[a:885863|reporting on John Giannandrea's tenure and Craig Federighi's reservations about generative AI]] — the company has been forced to compensate through acquisitions like Siri rather than organic research recruitment.
- Competing labs gain a compounding advantage: every researcher Apple loses to Meta or DeepMind strengthens those labs' publication output, which in turn attracts the next round of candidates.
Third-order effects
- If the pattern holds, secrecy functions as a structural tax on consumer-hardware companies competing in AI: they either relax publication norms or cede the research talent market to open labs, with product quality gaps — Siri chief among them — as the visible symptom.
- The 2015 warning anticipated a durable rule of the AI talent market that outlasted any single product cycle: strategic legitimacy through research openness is now a precondition for lab competitiveness, not a marketing choice.
The trend: AI talent increasingly flows toward organizations that publish research openly, making corporate secrecy a measurable competitive liability in the race for machine learning capability.