Hacker News: Inviz

New comment by Inviz in "Denial of service and source code exposure in React Server Components"

Inviz — Thu, 11 Dec 2025 22:40:47 +0000

have you seen the code of next.js? its completely impenetrable, and the packages have legacy versions of the same files coexisting, it's like huge hairball

New comment by Inviz in "Should CSS be constraints?"

Inviz — Wed, 10 Dec 2025 23:21:07 +0000

If I am to provide a summary of why layout shouldnt be linear constraints, is that it can't faithfully represent content overflowing onto multiple lines. I.e. it's inherently one dimensional. I.e. you can't really have a layout that adaprts to the screen size without creating a lot of separate breakpoints. This is a big limitation, that for example flexbox doesnt have. When I left Grid i immediately went and reproduced a lot of stuff we've been doing in (new then) flexbox layout engine, and i was like: Oh my god, this is so much more powerful.

New comment by Inviz in "Should CSS be constraints?"

Inviz — Wed, 10 Dec 2025 23:18:38 +0000

Hello Jon! This is Yarik. Cool to see you doing all this machine learning work these days.

New comment by Inviz in "Claude 4.5 Opus’ Soul Document"

Inviz — Tue, 02 Dec 2025 23:16:57 +0000

Is there a consensus about "Dont do it" negative prompts vs "Do it this way" positive prompts? So it's negative when there's a hard line, and positive when it's being nudged towards something?

New comment by Inviz in "The CRDT Dictionary: A Field Guide to Conflict-Free Replicated Data Types"

Inviz — Tue, 02 Dec 2025 06:11:23 +0000

Well, OT has its CAP-like theorem, which says that you can make some sacrifices to get some qualities. Excluding time component from the OT and removing the "first edit wins" rules, allows for very flexible OTs with deterministic tie breakers (i.e. when two people are writing a word in one place, sort words alphanumerically to brea the tie). But in order for this to work for P2P environments, the list of actions can't be collapsed/optimized until all parties rebased their changes against all other parties. In centralized systems, peers can throw away the log of ops that happened before the server's last acknowledged state

New comment by Inviz in "The CRDT Dictionary: A Field Guide to Conflict-Free Replicated Data Types"

Inviz — Mon, 01 Dec 2025 00:24:14 +0000

is there a modern operational transformation implementation that would work in P2P environment? i've had my own, but it had some convergence issues in complex fuzzied scenarios, couldnt fix before my enthusiasm ran out. I still feel like OT is more elegant and easier to integrate than CRDT. But i'm not finding good lightweight implementations.

New comment by Inviz in "Tell HN: Happy Thanksgiving"

Inviz — Thu, 27 Nov 2025 23:22:16 +0000

In these 15 years, HN was a website that shaped me and my worldview. It's a social circle that inspires me and broadens my perspective.

New comment by Inviz in "Waiting for SQL:202y: Group by All"

Inviz — Sun, 16 Nov 2025 23:57:44 +0000

What's wrong with GROUP BY 1,2,3?

New comment by Inviz in "JSON River – Parse JSON incrementally as it streams in"

Inviz — Wed, 15 Oct 2025 01:48:21 +0000

Please consider it a feature request

New comment by Inviz in "JSON River – Parse JSON incrementally as it streams in"

Inviz — Mon, 13 Oct 2025 23:48:09 +0000

I want to ditch stream-json so hard (needs polyfills in browser, cumbersome to use), but I need only one feature: invoke callback by path (e.g. `user.posts` need to invoke for each post in array) only for complete objects. Is this something that json river can support?

New comment by Inviz in "Sampling and structured outputs in LLMs"

Inviz — Tue, 30 Sep 2025 06:44:18 +0000

yes but you can put it on nested objects. Each type: object can have its own propertyOrdering, so you control whole tree

New comment by Inviz in "John Jumper: AI is revolutionizing scientific discovery [video]"

Inviz — Tue, 30 Sep 2025 00:34:07 +0000

I have a mildly psychotic friend who think that he uncovered the secrets to everything with AI. Quantum theory and Jungian archetypes, together with 4 dimensions - great mix

New comment by Inviz in "Sampling and structured outputs in LLMs"

Inviz — Fri, 26 Sep 2025 12:38:58 +0000

nested field have their own propertyOrdering

New comment by Inviz in "Sampling and structured outputs in LLMs"

Inviz — Tue, 23 Sep 2025 14:20:30 +0000

gemini api has propertyOrdering field for that

New comment by Inviz in "The attr() function in CSS now supports types"

Inviz — Fri, 22 Aug 2025 02:48:32 +0000

hmm... where to start?

  - Variables cascade and cause restyle of whole sub-tree.
  - Attributes often are already on the html element, e.g. title, aria, data attributes, so dont need inline styles
  - Not mixing data with presentation

Just to name a few

New comment by Inviz in "The new shape of Mixxx 3.0 – Open Source DJing"

Inviz — Thu, 07 Aug 2025 00:43:28 +0000

I wish this had youtube video mixing support. This is something i cant find in other software.

New comment by Inviz in "Cerebras launches Qwen3-235B, achieving 1.5k tokens per second"

Inviz — Wed, 23 Jul 2025 14:22:14 +0000

I contacted their sales team before, cerebras started at $1500 a month at that time, and the limits were soooooo small. Did it get better?

Edit: Looks like it did. They both introduced pay as you go, and have prepaid limits too at $1500. I wonder if they have any limitations on parallel execution for pay as you go...

New comment by Inviz in "Context Rot: How increasing input tokens impacts LLM performance"

Inviz — Tue, 15 Jul 2025 01:50:35 +0000

Cursor lifted "Start a new chat" limitation on gemini and i'm actually now enjoying keeping longer sessions within one window, becuase it's still very reasonable at recall, but doesnt need to restate everything each time

New comment by Inviz in "A job queue in two lines of JavaScript"

Inviz — Tue, 15 Jul 2025 01:46:38 +0000

I wonder if people who are claiming its not the case ever tried to use promises in hardcore ways like this, chaining 10s of thousands of calls. You often hit some bullshit, i.e. Promise.race is a big cause of issues.

Even if you are doing `then().then()`, something _else_ in the code base could retain the promise somehow and then your whole chain isn't GCable!

For example https://github.com/rxaviers/async-pool library that implements concurrency for async iterators (and uses promise.race) subtly creates GC pressure which you dont see until you make a whole lot of calls and suddenly its slow

New comment by Inviz in "A job queue in two lines of JavaScript"

Inviz — Mon, 14 Jul 2025 10:33:45 +0000

yeah but it creates closure as well, which typically references some objects that isn't easy to GC. So if you have long-living promise with some data captured in closure, it will not be cleared. So doing then().then().then() may not release objects referenced in callbacks that resolved time ago.