Solid 2.0
SolidJS 2.0 is being received as a major evolution of the reactive JavaScript framework, introducing a unified async/signal graph, new server RPC primitives, and a “built to be written by agents” philosophy aimed at both human and AI-driven development. Commenters compare it extensively with React, Svelte, Vue, Remix and Qwik, praising its performance, fine-grained reactivity and reliance on standard TypeScript tooling, while noting trade-offs like larger bundle sizes and a still-maturing ecosystem of component libraries. The release also prompts broader reflection on AI-generated framework code, agent-oriented tooling, and how emerging patterns like universal renderers and headless UI kits shape the future of frontend development.
Overall reception
- Many commenters are excited; several say Solid has been their primary framework for years and they avoid React now.
- Some remain indifferent or fatigued by frontend framework churn, likening these debates to low-level implementation details they no longer care about.
Ecosystem & component libraries
- One thread claims Solid’s ecosystem is thin; others counter with a long list of headless and styled UI libraries (Kobalte, Ark, Shadcn ports, etc.) and new support in Vercel’s UI SDK.
- Discussion clarifies what “headless UI” provides: shared, tested behavior (focus, keyboard nav, etc.) without visual styling, helpful for consistent design systems.
AI, agents, and “built to be written by agents”
- Some see AI-focused positioning as marketing; others argue these features (better errors, stronger patterns) equally benefit humans.
- Multiple comments report that agents perform better in codebases with strong architectural patterns, strict types, and linters—frameworks help, not hurt.
- There’s a recurring theme that AI pressures teams into “grown-up” practices: planning docs, task IDs, enforced tooling, and clearer tutorials, often written partly “for agents.”
Comparisons with React, Svelte, Vue, Qwik
- Several developers describe moving from Svelte to Solid, preferring plain TS/JSX and simpler tooling over Svelte’s custom compiler/tooling stack.
- Vue’s upcoming Vapor mode is mentioned as reaching Solid/Svelte-level performance; among top-tier frameworks performance is seen as “good enough,” making ergonomics and features decisive.
- React is critiqued for hook “coloring” and constraints (e.g., conditional hooks), while Solid’s signal graph and async integration are praised.
New features: async graph, server RPC, live/broadcast
- The unified async + reactive graph is widely praised: it can detect waterfalls, bad effect loops, and poor graph shapes.
- The server RPC / live / broadcast model is seen as a cleaner alternative to LiveView/htmx by some, but documentation wording around
broadcastis criticized as opaque “slop.”
Tooling, DX, and docs
- Solid’s use of standard JSX/TypeScript is highlighted as a big win: works with existing TS tooling, linters, formatters, and editors without special language servers.
- Some feel Solid 2 docs hide the best concise tutorials inside “for agents” materials (e.g., buried skill files), making human onboarding harder.
Performance, bundle size, and resumability
- Benchmarks referenced indirectly suggest Solid is already among the fastest; further gains compete with Svelte and soon Vue Vapor.
- However, multiple comments note a significant Solid 2 bundle-size increase even for trivial components (roughly 4–5x brotli’d in one test), attributed largely to the new runtime and scheduler.
- Some argue that unless a framework is resumable, performance differences are mostly about where you pay hydration costs and personal preference.
Signals, reactivity, and ergonomics
- Signals are praised as a paradigm shift: a simple dependency tree enabling fine-grained updates.
- Developers contrast Solid signals favorably with RxJS Observables in Angular, citing simplicity and debuggability.
- There’s some discomfort with explicit memoization patterns (
createMemo, etc.), seen by a few as emblematic of why they avoid JS-heavy frameworks.
Universal renderer & non-DOM targets
- Solid’s long-standing universal renderer is highlighted; examples include WebGL renderers for smart TVs and a watchOS UI renderer using QuickJS and a Swift host.
- Some note that this closes or reverses what used to be a React-only advantage in native-like rendering.
AI-written core & sustainability concerns
- One commenter expresses unease that much of Solid 2’s core was reportedly written by AI, worrying about deep understanding, long-term maintainability, and sustainability.
- They still view Solid 2 as a major conceptual advance but feel personal “AI crisis” anxiety about adopting heavily agent-authored infrastructure.
Politics, ethics, and vendor choices
- A tangent emerges about boycotting Vercel due to its CEO’s political meetings; some describe actively migrating projects away.
- Others see this as overreaction; the thread reflects divided views on mixing infrastructure choices with geopolitical ethics.