Tech Jobs Crisis, The Language After Kotlin, AI Writes CUDA Kernels
W07

Tech Jobs Crisis, The Language After Kotlin, AI Writes CUDA Kernels

Tech employment significantly worse than 2008/2020 recessions, Kotlin's creator on the next programming language, Codex and Claude enable anyone to write CUDA kernels

566articles
25+sources
Y

In this issue: Tech jobs collapse → What comes after Kotlin → AI writes CUDA kernels → Kent Beck's engineering philosophy → Pragmatic Summit recordings → US tariff chaos → References


Tech Jobs: Worse Than the 2008 or 2020 Recessions

One of the most-discussed technical topics on Hacker News this week: tech employment is now significantly worse than either the 2008 or 2020 recessions.

Key data (from job market analysis):

  • Tech job openings declined more than during either comparison recession period
  • AI adoption is accelerating some position reductions
  • Block (formerly Square) announced 40% layoffs, officially stating "not due to AI" — but the timing is telling
  • Large company "efficiency" drives meet AI acceleration amid macro uncertainty

Anthropic's research "Labor market impacts of AI: A new measure and early evidence" provides early quantitative evidence — AI is genuinely changing labor market structure, not just improving productivity.


The Language After Kotlin: JetBrains' Next Move

The Pragmatic Engineer deeply interviewed Kotlin's creator about what the next programming language will be:

  • Kotlin's success came from pragmatism: solving Java's pain points rather than chasing theoretical purity
  • The next generation language needs native AI collaboration support: type systems that let LLMs better understand and generate code
  • Kotlin Multiplatform (KMP) is Kotlin's next big bet

AI Enables Anyone to Write CUDA Kernels

HuggingFace Blog published an exciting piece: how Codex and Claude enable anyone to write custom CUDA kernels.

Previously, writing CUDA required deep GPU architecture knowledge — a rare skill. Now:

  • AI can generate efficient CUDA cores from natural language descriptions
  • This lets more researchers do GPU-level optimization without being limited to PyTorch APIs
  • HuggingFace demonstrated several cases: 2-5x training speedups, kernels entirely AI-generated

Kent Beck's Engineering Generational Philosophy

Kent Beck tackled a sensitive topic this week:

"Generation Gap or Just Rude?":

  • The divergence between young engineers' and senior engineers' attitudes toward AI
  • Not an age issue, but different tolerance for uncertainty
  • Senior engineers' advantage: ability to recognize when AI output is wrong

"Genie Session: Codex for Mac":

  • Kent's personal experiment: using AI Genie to help write a Mac app
  • Finding: AI's strength isn't new features — it's "accelerating things you already know how to do"

Pragmatic Summit Recordings

The Pragmatic Engineer released Pragmatic Summit recordings and summaries. Core debates:

  1. Does AI reduce or increase the number of engineers? Current evidence shows both
  2. Software development's "factory" trend: Agent assembly lines vs. craftsman culture
  3. The 2027 engineer role: More architecture decisions, fewer implementation details

US Tariff Chaos: Courts Order Refunds, Customs Refuses

  • Judge orders government to refund over $130B in illegal tariffs
  • But customs officials are rejecting refund applications
  • Nintendo officially suing the US government for tariff refunds plus interest
  • Tech industry remains in tariff hell — even with automated refunds, logistics and pricing chaos continues

References

Tech Jobs

Kotlin

CUDA Kernels

Kent Beck

Pragmatic Summit

Tariffs

AIJob MarketProgramming LanguagesCUDADev Tools