Love/Hate: Upgrading to Web2.5 with Local-First

Millions of developers have made their careers in the age of Web2, and most weren't even around for the days of Web1. Yet it's no shocking claim that Web2 still has a lot of problems that weigh it down, despite all the amazing advancements.

For years now, hopes of fixing those problems have largely hung on the promises of Web3 (blockchain and crypto). But just as other classic "version upgrades" (like Python 2 vs Python 3) lingered and stagnated, so too has Web3 largely failed yet to bridge the divide and pull along those of us stuck in Web2.

Big changes are usually the sum of many smaller changes. Web3 is too big a leap for many, so we need to look to more immediate and tangible evolutions.

Luckily, the web has a new way forward. Let's explore how local-first will upgrade the world to Web2.5.

Transcript

1. Kyle Simpson Upgrading Love/Hate: to Web2.5 with Local- First
2. Many “Great Divide”s
3. Web
4. Native vs
5. https://css-tricks.com/the-great-divide/ vs
6. vs
7. Front- End Back- End vs
8. UX DX vs
9. Vim Emacs vs
10. (tabs) (spaces) vs
11. Love! Hate! vs
12. Dichotomy Pick Sides!
13. Privacy Convenience or?
14. Security Performance or?
15. Rich Poor or?
16. Once Subscription (not specifically 37Signals, but…) or?
17. Enough More or?
18. No middle? No bridge?
19. Can the web exist… without giving up our data?
20. Can the web exist… without spinners?
21. Can the web exist… without subscriptions?
22. Can the web exist… without the cloud?
23. Web2 Web3? vs
24. Web2 (too far off) is crumbling…
25. No matter how crumbling one side is… most won’t jump over that huge gap.
26. Web2 needs a bridge… ???
27. …but we’ll come back to that.
28. privacy Web2
29. In the digital space… you are what you do.
30. Either you own that data… or others own the data that defines you.
31. If you don’t have custody of that data… you don’t own that data.
32. If they can block you from that data… you don’t own that data.
33. If they keep charging you for that data… you don’t own that data.
34. If you can’t protect that data… you don’t own that data.
35. privacy performance Web2
36. The web is supposedly “zero install”… and also promises “instant start” UX.
37. But far from actually “zero install”… the web is, in practice, install every visit.
38. And the average page size in 2024 is… ~2.7MB, growing ~7% YoY. https://httparchive.org/reports/page-weight
39. While the 90th percentile page size is… ~10.5MB, growing ~8% YoY. https://httparchive.org/reports/page-weight
40. It’s harder than ever for web devs to… fake zero-install and instant-start.
41. Hosting providers increasingly control… architecture and optimizations.
42. privacy performance access Web2
43. The web increasingly assumes… unlimited internet and power*. *especially with AI.
44. 2/3 of the world’s population doesn’t… enjoy such privileges.
45. Our best plan to address this disparity… is billionaires floating balloons!?
46. privacy performance access cost Web2
47. Users not only spend more on access… but more on devices, just to keep up.
48. Not only is the cost to users going up… but the cost to companies, too.
49. “During a gold rush, sell shovels.” During a shovel rush, sell metal.
50. privacy performance access cost complexity Web2
52. Programmer job security used to be… ensured by writing hard to maintain code.
53. But the complexity arms race… now ensures they’ll be replaced by AI.
54. Web2 is crumbling…
55. but… there is a bridge! Local-First
56. Local-First 1. No spinners: your work at your fingertips 2. Your work is not trapped on one device 3. The network is optional Seamless collaboration with your colleagues The Long Now Security and privacy by default You retain ultimate ownership and control https://www.inkandswitch.com/local-first/ https://localfirstweb.dev/
57. Local-First Data must default to primary storage on the user’s device.
58. Local-First Servers must be optional (or based on ubiquitous, agnostic protocols). IOW, Server-Last
59. https://blog.jim-nielsen.com/2024/allure-of-sync-engines/ @martinkl
60. 1. Create data on the user’s device 2. Optionally, sync data with other devices (servers, peers) How-to: Local-First
61. ▪ Not just data, though ▪ The app should act installed and remain persistently on the device, and should update atomically How-to: Local-First
62. Local-First Any framework will do, as long as it’s not a server-bound framework… …which is most of them.
63. Local-First However… If you put the server in the client… https://jakelazaroff.com/words/building-a-single-page-app-with-htmx/
64. Local-First DXOS ElectricSQL PowerSync TinyBase AutoMerge Triplit Yjs Replicache
65. Users: Privacy, Performance, Resilience Why Local-First?
66. Business: Cost, Complexity, Competition Why Local-First?
67. Local-First Obsidian Ledger Linear Trello (some or all of the local-first principles) Spotify
68. Local-First What about… syncing without any servers?
69. Bluetooth, NFC, Animated QR Codes, Quiet.js, LP2P, WebRTC, Socket Relay P2P Protocol, etc. Zero-Server Sync
70. *Bluetooth, *NFC, Animated QR Codes, Quiet.js, LP2P, WebRTC, *Socket Relay P2P Protocol, etc. Zero-Server Sync * Requires native app wrapper
71. *Bluetooth, *NFC, Animated QR Codes, Quiet.js, **LP2P, WebRTC, *Socket Relay P2P Protocol, etc. Zero-Server Sync * Requires native app wrapper ** Working Group Draft Spec
72. *Bluetooth, *NFC, Animated QR Codes, Quiet.js, **LP2P, ***WebRTC, *Socket Relay P2P Protocol, etc. Zero-Server Sync * Requires native app wrapper ** Working Group Draft Spec *** Requires signaling/relay servers (or P2P)
73. *Bluetooth, *NFC, Animated QR Codes, Quiet.js, **LP2P, ***WebRTC, *Socket Relay P2P Protocol, etc. Zero-Server Sync Web-only… yay!
74. Local-First What about identity without any servers?
75. 1. Your public-key is your identity 2. Local data protection (encryption at rest, E2E transmission) 3. Keys protected by biometric passkeys Zero-Server Identity
76. • mylofi/WebAuthn-Local-Client • mylofi/Local-Data-Lock • byojs/storage • mylofi/Local-Vault Zero-Server Identity
79. Demo
81. Local-Only …except syncing your “local ring” devices Many apps never need to share data…
82. Web2.5 Local-First Web2
83. The cloud is not inherently “The Web”, it’s an implementation detail. Web2.5 “The Web” is interconnectedness and collaboration around our data.
84. 1. Local Identity 2. Surf To Install 3. Peer-First (Server-Last) Web2.5 Principles
85. ▪ Autonomous, user-defined identity ▪ Data ownership ▪ Users mix/share their own data across apps ▪ App ownership ▪ Reduce distinctions between web, apps, and data Web2.5 Goals
86. Web2.5 Local-First Web2 Kyle Simpson Thank You!