The Next Decade of Software Development
summary
Forecasts that matter: cloud becomes default, Moore’s Law slows and chip design shifts to many-core and domain-specific silicon — Apple’s M3/M4 packs 25 billion transistors on a 3 nm die and high-end GPUs hit ~82 TFLOPS, while Frontier tops exascale. Large models power tooling (GitHub Copilot yields 25–30% productivity gains, ChatGPT reaches ~100M users in months). AlphaFold already accelerates protein design. AR prototypes (Meta’s Orion with puck and nerve‑read wrist device, ~$25k demo) and flaky quantum runs (a 53‑qubit run in ~200s) promise domain breakthroughs rather than instant general AI.
AI-written summary · about
chapters (14)
- 0:00The evolving tech landscape
- 2:01Cloud and mobile foundations
- 5:19The limits of Moore's Law
- 10:48New hardware architectures
- 15:06Programming and client stacks
- 20:08Low-code and containers
- 23:31The artificial intelligence era
- 26:14GitHub Copilot and LLMs
- 30:39Hype vs. reality in AI
- 33:46Generative AI applications
- 38:10Dynamic user interfaces
- 40:28The future of AR hardware
- 47:43Speculative quantum computing
- 57:29Conclusion and call to action
related talks
01h 07mThe Next Decade of Software Development
01h 07mKeynote - The Next Decade of Software Development
26mKeynote: Artificial Innovation Accelerates Artificial Intelligence
34mThe Future of Computing: AI technologies in the exascale era
22mAdvancing the Human Experience
52mComputing Performance 2022: What's on the Horizon
41mThe kaleidoscopic futures of AI and security
12mKeynote: What's New in AI