HNHacker News
TopNewBestAskShowJobs

kcb

2,877 karma · joined February 23, 2019

submissionscomments
kcb··on C# now beats Rust on WASM size
Not saying I do or I like it, but I suspect that is the trajectory. You don't ask the model to choose the backend, it's just not an important descision for someone looking to make a web app pop up in their browser. So at very least whatever the model picks will be the entry point for most people into software development.
kcb··on C# now beats Rust on WASM size
The models aren't writing new applications in C# unless you direct it to.
kcb··on macOS 15.8.2 Breaks OpenGL
Everything is universally supported in a small enough universe.
kcb··on macOS 15.8.2 Breaks OpenGL
Unfortunately there is no universally supported newer API thanks to Apple.
kcb··on Llama.cpp on a jailbroken PS5: run any GGUF model natively on the console
An N150 for $385 is a joke.
kcb··on Tesla Roadster
Yea, i was going to reserve one of these sub 1 sec 0-60 vehicles with rocket motors and suction fans, but then I didnt because of the ergonomics of the door handle design.
kcb··on Tesla Roadster
It was fine? I'd rather this than hyper rehearsed CGI human Apple launch video. At least its just real people talking like humans.
kcb··on 3rd Yandex Cloud Data Center Was Hit
Yea, there was even that idea to launch 1000s of satellites into LEO on a reusable rocket system. Crazy. How did people fall for that? There's not even that many satellites in total today.
kcb··on Trying the software factory pattern
Missed the part in July when the token bill starts rolling in
kcb··on A Tesla ran a stop sign and killed a man, Full Self-Driving/Autopilot was on
It's electrek, 95% chance any front page Tesla negative story is from them.
kcb··on QBittorrent breaks out of sandbox to commit crimes
You are still the one operating the computer. It's up to you to keep your eyes on the monitor at all times ready to intervene if it does anything illegal.
kcb··on Xiaomi: New CPU matches Apple cores single threaded, much faster multithreaded
No. its has a 96 bit bus.
kcb··on Cerebras CS-4
Nvidia is at this time a pretty well run company tech wise. They are going to keep iterating on the inferencing hardware stack over the next five years too.
kcb··on CFTC declares market emergency, orders Kalshi to continue to operate in New York
Isn't the act of not offering something in NY = to not offering it nationwide?
kcb··on Poland now 6th-largest EU economy, ahead of Switzerland and Belgium
Holistic...chooses one single state....

https://en.wikipedia.org/wiki/List_of_French_regions_by_Huma... Guess France is a third world country too.

kcb··on Muse Code and Muse Spark 1.2
Open the weights.
kcb··on Qwen3.8-Max: A New Bar for Coding and Cowork
If I recall most of the open weight models have provisions about large scale commercial hosting. So probably licensing issues.
kcb··on Europe EV Sales BEVs Jump 50% & Reach 26% Market Share
> However, unlike Teslas, you can _also_ plug and charge them in other networks.

why?

kcb··on Europe EV Sales BEVs Jump 50% & Reach 26% Market Share
Android Auto and Carplay are better than dealing with just crappy infotainment systems but they still don't make the overall software experience good. You end up with disjointed interface with functionality split between the automakers UI.

FSD is unmatched. Auto lane-keeping and adaptive cruise control don't provide the reactivity that FSD does even for straight highway driving.

Most objective reviews comparing a Model Y and Chinese EVs are far less conclusive than you propose.

kcb··on Kimi-K3 Technical Report [pdf]
You also need a place to put it. With liquid cooling and extremely dense power capability. Your typical colo or closet server room isn't going to cut it.
kcb··on Chinese chipmaker shares surge 470%
Terrible performance on $20,000 worth of GPUs.
kcb··on Nvidia DGX Spark as a daily driver
The igpu is a tiny RDNA2 GPU. Pretty certain the 9950x cpu will outperform it in compute tasks.
kcb··on Inkling: Our Open-Weights Model
Gemma 4 wouldn't really be a competitor to this. Gemma has the dense 31b model, but this has like 25x the total number of params.
kcb··on Microsoft fire idTech team at Id software
Why? What unsuccessful game does Blizzard have?
kcb··on AMD Ryzen AI Halo – $4k AI Dev Kit
Ability to run any OS is a pretty nice benefit versus the spark.
kcb··on Department of Commerce has lifted export controls on Claude Fable 5 and Mythos 5
LLMs are still easily replaceable. If the SOTA frontier model provides meaningful impact for your critical business function, then worse case you flip the switch to the next most capable model.
kcb··on Department of Commerce has lifted export controls on Claude Fable 5 and Mythos 5
You really think China doesn't have massive amounts of capital expenditure related to AI? They're actively bootstrapping an entire chip industry.... https://www.reuters.com/world/china/china-prepares-295-billi...
kcb··on AMD Strix Halo RDMA Cluster Setup Guide
No reason to use fiber on short runs like that. DAC cables are cheap and better in pretty much every way over short distances. You're probably thinking of RJ-45 NICs and SFP modules which are known to run pretty hot.
kcb··on DSpark: Speculative decoding accelerates LLM inference [pdf]
Nvidia's Nemotron 3 Super also shipped with MTP.
kcb··on DSpark: Speculative decoding accelerates LLM inference [pdf]
Google and Nvidia...
Page 1 of 34Next →