AI summary by TwinMind
The Post-AWS Future
Theo - t3․gg
7m 41s
Industry Shift from Horizontal to Vertical Teams
- Development industry moving away from large horizontal teams (backend/frontend) to small vertical teams
- Facebook era initiated this change - teams now organized around specific problems/user solutions
- Vertical teams include: product managers, designers, frontend/backend engineers, occasional platform engineers
- Major companies adopted this model: Facebook, Amazon, Twitch
- Platform engineers help with deployment consistency but average engineer works on vertical team
Transition from Generic to Specialized Solutions
- Historical approach: build generic solutions for maximum user coverage
- Previous trade-off: 9/10 perfect for 40% users vs 5/10 perfect for 90% users
- Generic approach made sense when building software was extremely time-intensive
- Now seeing emergence of specialized solutions that excel for specific user types
AWS vs Specialized Alternatives
- AWS provides solutions for vast majority of developers/problems but not optimally
- Live video example:
- AWS IVS: obnoxious setup, good quality
- MUX: specialized video product, easier setup, good quality
- Web hosting example:
- AWS: CDK and Amplify (complex relationship)
- Vercel/Netlify: GitHub integration, one-click deployment, built on AWS
- Many services solve problems better while built on top of AWS
- AWS can't compete with specialized focus - they build "one size fits most" solutions
Small Team Efficiency and Platform Abstraction
- Small vertical teams shouldn't rely heavily on platform teams
- Facebook/Twitch examples: constant collaboration between platform and vertical teams needed
- Complexity of underlying infrastructure requires platform team support
- Personal experience: Confidence to move from multi-hundred person company → 10 person → 2 person company
- Smaller teams moving faster using specialized tools that abstract platform team functions
- Vercel recognized deployment team pattern and built better deployment product
Future Market Structure Predictions
- Industry becoming more niche-focused
- Companies targeting specific user sets they understand well
- Profit reality: Vercel will never make AWS-level money due to smaller customer base
- Smaller focused teams can still achieve multi-billion dollar valuations
- No future mega-companies: Unlikely to see Microsoft/Apple/Amazon-sized companies again
- Rapid solution building for specific users prevents monopolistic takeovers
- Nobody can replace AWS directly - Planet Scale, Vercel, etc. built on AWS
New Company Landscape
- Future companies won't reach multi-hundred billion dollar valuations
- Lucky to achieve unicorn status ($1B)
- Expect many more $100M-$1B companies with smaller teams
- 5-person focused teams vs 500-person teams solving internal problems
- Small teams enable more focus and agility
Developer Experience Evolution
- Tools mentioned: Vercel, Planet Scale, Clerk, Axiom, Upstash
- Small teams don't need platform people for deployment/scaling
- Companies rolling own infrastructure/Kubernetes for 5 users can't compete
- 2-person team with auto-deployment via Vercel can iterate faster
- Better feedback loops result in superior software/user solutions
- Personal note: Life improved since avoiding AWS console
- AWS will be replaced by hundreds of specialized companies building on top
