Octane – React’s programming model, compiled
octanejs.dev114 points by nnx 13 hours ago
114 points by nnx 13 hours ago
> Measured, not vibes
Sort of ironic considering that there are a LOT of vibe tells in the copy throughout this page. It's a neat looking project, it's a shame that every other sentence has Claude's fingerprints all over it.
Which is ironic because the author works at OpenAI!
Yeah. I thought it sounded like OpenAI more than Claude, though maybe they’ve converged. It’s been a while since I used Claude.
There's enough inbreeding^W distillation happening that whatever OpenAI tone you're thinking of is probably Claude's tone (and vice-versa)
claudes still claude. still as load-bearing as ever. the ladies have no idea how absolutely right i am every day
Would look more appealing without the AI flavour. If you're vibe coding, do yourself a favor and write website copy yourself - it's the only evidence of human input or oversight your project will have.
It’s especially grating in this case because Dominic Gannaway, the creator of Octane, is the real deal (creator of Inferno, former React core team member, former Svelte core team member); so the project landing page looks like generic 2026 slopcoded abandonware, but it is in fact a real project made by someone who’s one of the great minds of frontend framework development (with heavy Claude use apparently, to be clear).
Sad to see formerly talented software developers become mediocre and bland, losing their unique flavor and personality. Maybe there's still some insight or idea that's their own creation, but it's obscured by the same text and code generation service, a soul-less corporate voice. It's possible language models will mature to better emulate that ineffable quality of having a "heart" or "taste", but it would still be empty without a person on the other side.
We’re looking to address the feedback with the website copy. Thank you for the feedback!
I still think code quality itself reigns supreme. I couldn't give less of a damn about a website.
Code quality at one point in time is useful, but what you want is dependability over time and responsiveness to security fixes; this means looking at the maintainer and their attention to detail.
Sure, the website by itself doesn’t mean anything, but placed into context it could mean the maintainer doesn’t put much effort into clear, unambiguous communication.
Just took a glance so I may be completely wrong, fell free to ignore, but if I understand correctly, tsrx is just syntactic sugar for normal tsx code and does not enable anything new?
If this is the case: I think introducing tsrx as part of Octane muddies the water a bit and makes me more hesitant to try (even if it is optional).
The idea to make tsx files more readable may be interesting but in that case it should go into its own separate project, and maybe just put a lang toggle in Octane docs and a page about tsrx compat, that way if people are intrigued by the new syntax they can just look it up but the docs stay focused.
agreed, tsrx is neat as supported but personally i'm not a fan of it, makes me think of angular and more domain specific language stuff. the whole tsrx angle makes me wary of this project entirely, which is a shame, since octane does sound pretty neat. i'll still check it out, but this was indeed my initial reaction upon seeing tsrx here
TSRX allows for better runtime generated guarantees, especially around loops which generates faster runtime logic from it. We can’t assume too much statically when reading something.map as the method might have different heuristics
I recognize this comparison style. Definitely generated by Claude. haha
- please post 10 new frontend frameworks daily to HN
- i ll take that over the AI, GPT, LLM, Agent stuff on HN anyday now
An interesting feature is this:
> A current-state getter. useState and useReducer return [state, update, getState], so a delayed callback can read the latest value instead of a stale capture.
There are currently a few workarounds in React for this, and using other state management libs also give you non-reactive access to current state, but this is a big help in certain situations.
For example, a callback / event that you pass to all your child elements, that needs to access the current state when run, therefore normally needs to be recreated with the new state captured (in useCallback) whenever state changes, which then causes all child components to re-render. Non-reactive access to current state in callbacks avoids this entirely.
There are some hacks that fix this by wrapping the updated callback in a ref and updating it with an event (for example, https://usehooks-ts.com/react-hook/use-event-callback useEventCallback as implemented various places) and returning a stable reference to a function that calls that saved one, though I always wonder if there are downsides to this.
The copy reads like it is react but improved. Which begs some questions: what strings are attached and why wouldn't the authors of react adopt this? What reasons would developers have to not switch from react to octane?
Does it have any overlaps with React Compiler — like the auto-memoization that they have?
Also, is there any read on the downsides / shortcomings / caveats of Octane vs. vanilla React?
The Octane compiler has overlap in terms of memoization but also goes far further
It seems like an attempt to port Svelte's features to React-like api?
The model is a bit different from React (no deps, hooks can be under ifs) — do coding agents recognize it or they get confused when they see that code?
Their llms.txt spells it all out: https://octanejs.dev/llms.txt
But it doesn't have the repeating yourself/over-emphasising that you sometimes see in LLM instructions when someone is trying to fight the model's default assumptions
I think the Github link [1] does a much better job at explaining what it is.
Looks interesting. But does anyone really like the useState useContext etc. React model? Might as well make a better one if you are going to be compiling no?
It would be nice to see how this differs from SolidJS.
It took a while to render when I clicked the Differences from React button on an iphone 17.
Same. I think it’s just a link, and its content took a while to load. Looks like it’s the docs site vs the marketing page.
I'm surprised how good Vue Vapor mode is from the performance comparison table.
Funny how Vue Vapor 3.6 beta and Ripple 0.3 show the exact same performance (1.0) as Octane, but their bars are a tiny bit longer so the Octane team can pretend to have a justification to put their implementation on top.
Classic statistics dark pattern.
We just run a script that runs the benchmarks and updates the chart. It’s rounds numbers too. We’ve been as honest as possible in all the benchmarks - there’s no point in lying about a metric.
This is an interesting problem. A similar approach that worked for us was to only load the tools the agent actually needs for each task.
This looks like an interesting project but reading these AI slop web pages is so exhausting. At least turn your brain on long enough to explain what your project does
Interested in experimenting with Octane, wish that docs emphasized more the SSR aspect of it, because I'm looking for a Ruby/Phoenix/Php alternative in TypeScript land, and one with React-like apis would be cool.
It has all the same SSR APIs as React. The Octane website is rendered with Octane and the source is in the GitHub repository if you’re interested to see how it works
Weren't there RedwoodJs, BlitzJs and few others that did that already?
theres gonna be a lot more at this rate. everyones tribalistic preferences about how web frameworks should be made... is never going to get solved. now everyone can just spend a week of their claude max plan making whatever framework they want. ive done it like twice now in the past year just for shits and giggles, it is fun but i wont pretend like i know where any of this is going, i think a hybrid of ssr and a blend of "the client should do some work" is probably the prevailing strats that will scale. something like aggrid is a very performant js datatable so i personally am not inclined to try and reinvent that from the backend, rather a backend has full coverage over its api surface but largely just sets/configures/sends data to but lets aggrid do what aggrid does. component models and diffing changes and all of that though and sending lil dom updates... shouldve never been purely on the front end to begin with. front end runtimes are about to get way smaller.
my favorite part of using whatever this weird ssr hybrid landscape ive been in is just not using graphql lol. i can just rawdog sql now for most things like the lords of data intended. its fantastic
[dead]
[dead]
Why are people still building these frameworks? Can't we just use LLMs to build the best JS framework?