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

even for Pentagon thing, Dario said he doesn't object military AI, but said Claude is not ready YET. I speculate he was afraid of reputational damage from cases if Claude would guide missiles on elementary schools.

It was ideological goal: they wanted to build query optimizer which would always be smarter than human, but looks like they admitted defeat after many years of stubbornness.

> We have had a champion in our team

there are good actors, which are empowered by AI to produce positive impact, but often there are N times more bad actors, which push crappy code to close feature requests fast, increase performance LoC-like metrics, etc.


most industries have high cost of entrance unlike software, so decision makers are way more careful on how to move forward.

In software + GenAI now every housewife can build some App over evening.


> And if the product is good it will succeed.

it needs to win marketing landscape, hyper-overcrowded by thousands of competitors, slop-gened over weekend.


Could you imagine Obsidian being posted on HN today, if it weren't really popular already? There's no way a tiny team working on a note taking program would make it out of new, no matter how good it was. I wouldn't click the link, myself.

Maybe that gigabyte is occupied by useful information: traces/memory?

Traces and memory are text. A gigabyte of text is an insane amount. That is an equivalent of tens of millions of lines of code, or hundreds of millions of AI tokens.

A gigabyte is a lot of memory. Even the largest context windows are a small fraction of that with any sane engineering discipline.

For each LLM interaction they likely have bunch of thoughts traces, tool calls, etc, which don't go to context, but still can be retrieved.

But I obviously don't know for sure.


Nope. Used to render on the terminal like a game engine.

https://x.com/trq212/status/2014051501786931427


This kind of immediate-mode rendering is quite standard for TUIs. Although immediate-mode rendering tends to be significantly simpler and use less memory than retained-mode rendering, at the cost of some redundant computation. So I am not sure if this is the reason for the bloat.

It’s possible that it doesn’t play well with JS garbage collection, since it recreates the whole UI structure for every frame (which tends to not to be an issue in the languages immediate-mode is usually employed).

But yes it’s a bit more akin to game renderings than web rendering. Which can be totally fine if done well.


I haven't tried to make a TUI admittedly, but double buffering is the oldest technique on the planet. A TUI doesn't even need to pay the cost of a lot of pixels since its effective resolution is much lower

Long long time ago, I used to do some graphics stuff in 320x240, which uses a whopping 64KB per buffer, and still has more resolution than a terminal.

In 1GB I could probably fit all the buffers to double-buffer all the TUIs in a whole country. Well, maybe not. But it's likely not that far off.


> render on the terminal like a game engine

All those CPU to render this effect

https://x.com/cyrilXBT/status/2060617507615207904


ULTRACODE! NOW WITH MORE BLINKENLIGHTS!

(to be read with the Unreal Tournament announcer voices, see https://www.youtube.com/watch?v=MwxjYFqP35A )


How on earth are you spending more than 50us on a UI like this from start to finish? What the actual hell? 11ms to construct a scenegraph of this complexity? I don't even know what to say to that.

In comparison, I have around 3ms total latency when streaming 4k 144hz from headless machine in my basement to upstars :-D

At least that is what Moonlight client claims.


Frankly that's an insult to gamedev. Literally every game engine I can think of could do better. Probably even Unreal Engine could do better.

If I saw our UI show up in the profiler eating 5ms of CPU time every frame, I'd send whoever was responsible to QA hell until they find some way to redeem themselves. Not even fancy animated 3D UIs, like what you get in Death Stranding, eat up these kinds of resources. Not even remotely close.

I sorta remember Quake console running on an 486dx2 ..

LOL right? This is all that needs to be said about the engineering behind Claude Code

Do game engines constantly have buffer issues?

Depends on if they're written with Claude

I think un-unified memory issue is solved by software layer in datacenter setting: model is distributed across multiple GPUs in the same server, or across multiple servers if model is extra large.

> The thing about Apple is that as the "IT" guy for my family, its ecosystem is the one which needs the least attention from me. > It really just works.

chromebooks likely need even less attention


Well, and so does a chalkboard, but if what they want is a computer...

Now we have promises, observables and signals.

I would be more happy if it would be just one of those..


Each one of these solves a different problem.

Promised - async

Observables - streams

Signals - reactivity


In theory that’s true (although observables are for reactivity too), but Angular uses observables for its http library and http requests are very much not streams. It’s one of the main downsides of working with Angular, the http library is mediocre and does come with the added overhead and complexity that rxjs brings.

Until this release (if you only use stable features) using forms meant dealing with observables too, even if you just want to read data when submitting a form and validating some data on change/blur.

And often you’ll find that your data from promises, observables and signals need to interact with each other, which can be annoying.

Fortunately the situation with signals and their async usage is improving, and iirc the Angular team wants to make rxjs optional, but until it is Angular can be a confusing mess on some points.


I partially agree, there is an overlap between signals and rxjs, however the core business is different- observables are about data manipulation, while signals are about efficient state management.

Regarding angular I agree, rxjs was a bad choice for data management, and before signals arrived I abandoned rxjs in favor of mobx in my angular projects. However you could roll your own http client, we used axios, and using DI it’s a drop in replacement.


> Observables - streams

> Signals - reactivity

The r in rx stands for reactive.


The react in react stands for reactivity, however it is not.

React reacts to changes in state or properties by automatically updating the UI. What's not reactive about that?

Its entire state management is not reactive, it’s always on push, not pull. You always need to call setState to get render changes.

But why is push vs pull the definition of reactivity?

I suppose we can say that there are different kinds of reactivity. Signals is one kind. Observables à la rxjs is a different kind (the whole model of programming with rxjs was referred to as "functional reactive programming"). Observables are push-based. Signals, as I heard, are a more complex primitive, which, under the hood, is push-pull.

React's reactivity model may be crap; but this doesn't make it non-existent.


Maybe push pull wasn’t the best metaphor, but the point is that everything can be reactive, it only depends on how much boilerplate you need to write to achieve the desired result.

Since react doesn’t have a true reactive model, you need to subscribe to changes manually (use effect) to create computations, while in signals it’s a primitive (computed).

I actually created a lib that operates signals over reacts state management (https://roypeled.github.io/react-logic/), so I removed the boilerplate to create a true reactive system.

If you want, you can create reactive system just from JS primitives, using callbacks. But that doesn’t make JS reactive by nature.


> you need to subscribe to changes manually (use effect) to create computations

How do you mean? Since the render function reruns during every update to state/props, derived/computed values can be calculated from the updated state/props during rendering.


When all you need is a synchronous operations, yes. When it involves async, batching, buffering, and user input, it becomes much more complicated, and every step needs to be setup manually.

UI is reactive, not state. You push changes to state and UI reacts to it.

A derived state is certainly reactive.

Of course you can have reactive state, your complaint however was:

"The react in react stands for reactivity, however it is not." [because] "Its entire state management is not reactive"

React is primarily an UI library, not full state management library. And its UI is reactive.


Agreed, the OP said that the r in rxjs stands for reactivity, so my point was the the names have little bearing on the actual design patterns achieved with the libs

Ah, now I see what you meant.

The chance is that they cache out during IPO, and will lose interest to increase capacity, some/many contracts will be canceled, and demand being reduced.

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

Search: