pasteleft
pasteleft on Hacker News
Isn't daily conversation the biggest strength of ChatGPT????
There are plenty of coding agents, but I thought ChatGPT is the only one frontier model for any conversations..
These days games are about 20GiB, you can't put the entire data in the cartridge while having the same price with the digital one.
Also, even before GKC you didn't have a physical ownership because games now have a online update. You only have a physical ownership for v1.0.0; newer versions required online connection anyways.
Interesting project, but why didn't they just run the test for every input that is less than 64MB?
I bet it is cheaper (and way more understandable) than generating 180 million lines of proof with frontier AI...
Didn't GitHub broke the entire CI system by switching to "main" branch :thinking_face:
Anyway, I think it'll be the same.
Tools will support SHA-256 quickly and we might have a migration program that converts SHA-1 repo to SHA-256 repo.
The only problem is that git (and related tools) will get twice as big...
"Safe Rust" being closer to "optimized C++".
And those "optimized C++" usually incldues assembly code, so it's a huge improvement.
Of course, Rust is not a magic, so just porting to Rust wouldn't make this performance improvement.
It seems their AI overfitted code to Rust compiler to find safe Rust code that compiles to efficient assembly.
Either Muse somehow bypassed the MacOS's security or Muse was able to access Messages without explicit permission.
It is Muse's fault, but I also agree that it is MacOS's fault too.
Clearly the Antrophic don't even look at the single line of the code, so I'm curious why do they even need a Javascript (+ engine) for the Claude Code.
They can just refactor it into C or Rust so they can ship a real binary instead of a script and a runtime.
I don't understand what you are talking about. Vibecoding IS a human error.
I'm not sure why they didn't just extend plugins, but having mod support is definitely a plus for everyone.
AGENTS.md seems to simply showcase what mods are capable of.
Jev HAS hallucinations and it doesn't attempt to solve hallucination at all.
For three choices problem (A,B,C), what Jev guarantees is that it will give the choice in a defined schema (type-safe). It never guarantees that the choice is correct (hallucination).
Thank you for adding a 5-line AGENTS.md. Other projects force you to install skills or MCP even when they're just doing the same thing.
This is really disappointing. I love programming community because they just share or open(-source) their knowledge with an absolute no cost. There's tons of resources about how to build a OS, DB, Apps, etc. I'm curious why would you think differently about AI workflows.
All of these Skynet-style AI attacks are unrealistic and impractical. What’s worse is that discussing these things distracts from issues that are far more realistic, dangerous, and HAPPENING RIGHT NOW: bad people using AI to do bad things.
Soon people will just prompt "Hack the HuggingFace" with their local models, and you don't need an autonomous agent swarms for that.
Why would a malicious group want to build a self-running LLM hoping it would somehow trigger a nuclear war? They would just prompt to trigger a nuclear war; it's more predictable, faster, and easier to control.
Clearly it will make this even worse. Soon every HN posts will have some comments like "86% AI. Skip it."
The whole point of the debate is that "those around them" didn't agree to send their so-called "de-identified" data to Apple. In fact, "those around them" had no chance to agree/disagree at all.
LLM is "stochastic parrot next word prediction machine"; it's just that this "stochastic parrot next word prediction machine" have proven to be smarter than most people. I mean, this already happened with AlphaGo too.
I don't understand any of this. If Claude is going to add some patterns (or some probablity distribution) in generated text, how can this not reduce code quality?