News for hackers. Stuff that matters.

wilonth

Re:The Deletion of Docker.io/Bitnami on Thursday August 28, 2025 @04:57AM (#45049971)
I never understood the point of Bitnami. Every time I tried one of their image / package, it's a complicated mess full of custom and strange stuff, really hard to work with.

Instead of a simple package of the software based on some familiar base, you get some weird enterprise garbage that follows strange conventions and a nightmare when you need to customize anything.

This won't stop any app store spammers, if they can spam the app store it's no problem for them to get 20 fake devices and bypass this. Again it's the genuine indie developers getting punished.
Re:Why Nvidia Keeps Winning: The Rise of an AI Giant on Friday July 07, 2023 @06:12AM (#36629591)
Apple was able to break the CUDA lock-in with their Metal computing api (MPS), in no time at all. Within months, now the major AI libraries like Pytorch and Tensorflow all support Apple GPU without a hitch.

What's taking AMD so long? They just can't do software I guess?

Re:InternLM – new open source 7B LLM on Thursday July 06, 2023 @08:54AM (#36615021)
7B params would take 14gb of gpu RAM at fp16 precision. So it would be able to run on 16gb GPUs with 2gb to spare for other small things.
Trying to understand how the 86 billions of neurons in humans achieve consciousness is like trying to understand how the billions of weights in ChatGPT interact with each other.

Basically impossible impossibility. For a system which derives its behavior from complex interactions between those billion components, you can only understand its origins how it was tuned, and some high-level concepts of its workings. (Which we already achieved both for human brain and ANN).

Not sure what the neuroscientists are even researching at this point, has there been any major findings from neuroscience in the last 10 years? (With similar kind of impact as Transformers in 2017?)

Re:From Word Models to World Models on Friday June 23, 2023 @02:19PM (#36450439)
"Hush hush, I'm gonna sacrifice the queen to do a surprise checkmate!" Agent said
Re:From Word Models to World Models on Friday June 23, 2023 @02:14PM (#36450374)
Was excited for a moment, thought it was related to this https://worldmodels.github.io/.

World models are meant to be for simulating environments. If this was something like testing if a game agent with llm can form thoughts as it play through some game it would be very interesting. Maybe someone on HN can do this?

Re:Why I started (and stopped) making games on Tuesday June 20, 2023 @04:26AM (#36401159)
Yeah Godot editor is snappy and smooth, unlike Unity and Unreal. However Godot's 3D and physics implementation leaves a lot to be desired. Not sure if Godot 4 solves those issues.
Dude database is supposed to be on disk not RAM.
True, 90% of doctor diagnosis are garbage and maybe even worse than ChatGPT. The 10% is for when you have a visible problem that can be cured with a surgery, they are pretty good at that. A doctor with GPT-4 could be way better than one without.
Re:Ubuntu stops shipping Flatpak by default on Wednesday March 29, 2023 @09:12AM (#35356742)
Some apps chose to distribute with snaps first, past and present. And without snaps they might've pushed a more unified or better experience with something better.

Canonical has money to do lots of good, too bad they waste it on terrible engineers and terrible projects.

Re:Ubuntu stops shipping Flatpak by default on Wednesday March 29, 2023 @08:40AM (#35356373)
Wow they're still pushing snaps, the project with some the worst engineering I've ever seen.

- Extremely slow at doing anything, even the most basic commands.

- Ridiculous auto-update mechanism (you can't even disable it wtf).

- Random, nonsense limitations (why can't I open dot files and dot directories???).

So terrible that for most apps that I installed with snaps I end up installing the deb version later on.

What an abomination, it is a devil that's hurting the Linux desktop everyday.

Can I get the name?
Re:Artificial synapses 10k times faster than real thing on Wednesday August 10, 2022 @03:53AM (#32408964)
Breaking news: website which displays "Hello World" runs faster than Google.
Solana programs have a "final" flag that could be turned on to make it immutable. This project doesn't have it turned on so it's open to update.
Re:Cozy Futurism on Monday February 28, 2022 @11:49AM (#30501966)
Except that the US government has Trillions to spend and they absolutely do not need Musk's wealth to do this. 20 billions is nothing for the US government's budget, why don't they end homelessness? Again, a socio-political problem that money alone cannot solve.
Re:Cozy Futurism on Monday February 28, 2022 @11:33AM (#30501751)
Why are people even arguing about space travel vs solving poverty? Just imagine if all of Musk's wealth is seized by the US government, would that solve poverty or homelessness or even make a difference? Obviously not, Musk's billions are barely a drop in a bucket. The US government printed and spent Trillions during the pandemic, and that barely made some people's life better.

Most of the problems that we're having on Earth, including poverty, homelessness and climate change, are in fact socio-political problems and they can't be solved with any realistic amount of money. The only way to solve them is through either strong socio-political coordination or very advanced technology (AI, nuclear fusion etc.)

Re:V Playground on Saturday March 30, 2019 @02:33AM (#19528117)
It has 2 modes, one for debug (compile directly to machine code) and one for release (compile to c)
Re:C++20 Design Is Complete: Modules, Coroutines in C++20 on Sunday February 24, 2019 @02:24AM (#19237553)
We can then skip that 98 and go straight to C++101.
Re:A Go Custom Flutter Engine Embedder for Desktop on Friday January 18, 2019 @02:01PM (#18941869)
Its usage in fact has nothing to do with Go or any other languages, Flutter is a framework for Dart and Dart only.
Re:Dr. Elon and Mr. Musk: Life Inside Tesla's Production Hell on Friday December 14, 2018 @09:19AM (#18680945)
Why do you guys say this is a negative article about Musk? Did you read the whole of it? It's quite a fair story about the downs and ups of Tesla in the past couple of years, and some of Musk's weird behaviors.

In the end, as the article mentioned, Tesla got through the 'production hell', October's financial reports were positive and Musk's emotions went back to (more) normal levels.

Re:Ask HN: Who wants to be hired? (December 2018) on Monday December 03, 2018 @12:40PM (#18590848)

    Location: HCMC, Vietnam  
    Remote: Yes  
    Willing to relocate: Yes, anywhere maybe  
    Technologies: C++11, Javascript (Nodejs, koa, React, Redux, React Native), GraphQL, Python, Go, ElasticSearch, Postgres, AWS, Linux, Bash, OSX  
    Résumé/CV: https://www.visualcv.com/y5dcchc68lf38dj7l22k/pdf 
    Email: [email protected]
Re:Show HN: An “automatic” way to learn Chinese on Saturday November 17, 2018 @11:37AM (#18476161)
That's strange, haven't heard of such an issue. What's your device and android version? You can try reinstalling the app or test it on another phone.
Re:Show HN: An “automatic” way to learn Chinese on Saturday November 17, 2018 @01:39AM (#18474319)
Update: The app now has a way to turn off the sound in the last (settings) tab.
Re:Show HN: An “automatic” way to learn Chinese on Saturday November 17, 2018 @01:32AM (#18474296)
I certainly have no intention of shaming or talking him down, just trying to suggest an option, if it comes across that way, my bad.
Re:Show HN: An “automatic” way to learn Chinese on Saturday November 17, 2018 @01:27AM (#18474286)
The app uses Expo, those libraries just come with Expo by default and can't be excluded, but they are unused and not enabled. Only Google Analytics is in use.
Re:Show HN: An “automatic” way to learn Chinese on Friday November 16, 2018 @10:08AM (#18469203)
No plans for now, I'm having a bit too much to work on atm, no teammates :(
Re:Show HN: An “automatic” way to learn Chinese on Friday November 16, 2018 @10:05AM (#18469189)
Thanks for checking out, I'll definitely look into the font size issue. I would also want the writing feature myself but just haven't got around to implementing it, hard to figure out where to put it too.
Re:Show HN: An “automatic” way to learn Chinese on Friday November 16, 2018 @09:55AM (#18469126)
I think you should try learning the characters for once, maybe this time you won't have so much difficulty remembering them.
Re:Show HN: An “automatic” way to learn Chinese on Friday November 16, 2018 @09:53AM (#18469110)
Unfortunately none at the moment. For the second question, it's just the simplest, most straightforward design to me.

Older »