Hacker Newsnew | past | comments | ask | show | jobs | submit | verdverm's commentslogin

I hope Google follows their lead, Google Cloud has way to much space dedicated to "in platform ads"

Is this ~cat fishing~ crab potting the NFT crowd?

no NFTs but sadly yes it makes too much sense so might have to do it.

Tangential to this cool project, Hytale has an amazing modder experience and toolset built into the game. Everything is JSON driven, with hot reload, so we can build tooling in any language (without needing to name said language, lol to that thread here, ahem, I'm using CUE, ahem). They have flags to dump the JSONschema for everything, including mods, as well.

More related, Worldgen v2 is pretty amazing compared to Minecraft. What is the basis for the Worldgen in this project? Not too different from a 3d voxel game? I'm pretty new to it and have a general curiosity still.

https://hytalemodding.dev


Initial research indicates that agent written instructions and skills decrease their success rate in completing tasks.

This intuitively makes sense because they make many mistakes and largely lack the context and capability to know what needs to be done.


Most people don't write Rust, so being in Rust is a disadvantage for those people. Most users don't care what language is used.

Requisite link to satirical study

"Parachute use to prevent death and major trauma when jumping from aircraft: randomized controlled trial"

https://www.bmj.com/content/363/bmj.k5094


An app you can use to play back their audio on a short delay that messes with the brain

https://github.com/Pankajtanwarbanna/stfu


Pretty sweet. Do you have it hosted anywhere? Seems github doesn't want to let you load HTML directly (for obvious reasons lol)

Not mine, I'll keep this limitation in mind when I redo my personal site and add it in so it's easy to remember (for me at least ;)

Looks real pretty, I like the aesthetics too. Various click handlers would be nice, for example expand/collapse or navigate to some link.

Context, Argo Workflows DAGs (and UI generally) sucks real bad. This would be a great improvement.


Thanks! Right, I asked Claude to write a ROADMAP.md and it was super long and a lot of YAGNI. I try to work much more intentionally but I got some great ideas from it.

I'm thinking first of my own needs and click/mouse handlers are definitely yes. It would be nice to enable popups when hovering over nodes/edges.


Gave it a star and a watch, looking forward to seeing the progress.

What are you using it for?


I looked at CUE and lattices and I realize CUE fits right in with some other work I'm doing!

For now I made it possible to draw Hasse diagrams. That seemed like a fun way to extend the project. I did not add any CUE examples, that seemed out of scope but who knows, there could be some applied examples in the future.

I see that lattices are commonly drawn with straight line, but that is mostly historical (how it was done). I decided to allow bezier edges as well because, why not?


Both personal and work projects. This is personal: https://github.com/23min/FlowTime I see DAGs everywhere...

> I see DAGs everywhere

lol, me too, though the lattice has enabled a more refined view for me

"Squint harder if you can't recognize the cube." ~ image caption

https://cuelang.org/docs/concept/the-logic-of-cue/


Hmm… I need to read around on that! Was working on a policy drift detection project but never heard of CUE before.

Join us in Discord! I'm a big CUE fan, https://cuetorials.com is one of my outputs.

My personal project is https://github.com/hofstadter/hof, a CUE centric developer experience, and perhaps where I will try out your project first, for visualizing CUE things.


https://victoriametrics.com/blog/log-collectors-benchmark-20...

1. The test setup is divergent from the real world (single node k8s cluster)

2. Their product is #1 in every metric they measured, but is also missing features from those compared against. Will those features change the results?


Disclaimer: I am affiliated with VictoriaMetrics.

> The test setup is divergent from the real world (single node k8s cluster)

The setup was chosen to simplify the suite, so it can be easily run anywhere. In real world, log collectors are mostly deployed as deamonsets and this is what was tested during the benchmark. The vlagent was initially developed to run as a deployment, though. So I don't think changing setup will affect its performance.

> Their product is #1 in every metric they measured, but is also missing features from those compared against. Will those features change the results?

Depending on what features will be involved into the testing. In the benchmark, all collectors are doing the same job: collecting logs, parsing JSONs, shipping log records. So they are even in used features. Of course, the #1 product is missing features for log transformations, but these features aren't used during testing and shouldn't affect performance of other log collectors.

The bottom line of the post has "Should I switch" section explaining what's missing yet in product #1 for transparency.


Using a single node cluster is not representative, full stop.

Meh, I don't get it—what's stopping you from running the same benchmark on a Kubernetes cluster and sharing your own benchmark results, instead of just claiming that this benchmark is crap?

The company selling a product and trying to convince me with benchmarks should be the one putting in the honest effort. The point is to nudge them towards doing this themselves, because people who know about things (how to run distributed system benchmarks) will silently close the tab. HN gives feedback to each other so we can improve. Do you have a problem with that?

Another thing we try to do here is avoid comments that lower the level of discourse, it's in the guidelines.

It seems you are biased to OP, your comment does more harm than good to their effort. (why would I want to use this project if the people around it make snide remarks)

https://news.ycombinator.com/item?id=44680602


I have migrated away from NextJS for similar reasons of being too special case in the ecosystem. Next in particular is designed for the way Vercel runs applications and comes with extra pain if you don't run on edge. (see middleware as an example, I think they renamed it since I left)

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: