Philpax

5 ms·

61.
Qwen3.8-2.4T(huggingface.co)
713 points by Philpax 2mo ago|171 comments
62.
by Philpax 2mo ago
For me, personally, it's that the Bun guy - specifically him, not a mathematician - indirectly progressed the Riemann hypothesis by repeatedly telling a model to ganbatte! It's a ridiculous position we find ourselves in.
63.
by Philpax 2mo ago
> Jarred Sumner, an Anthropic staff member (and non-mathematician) prompted Claude to “take a real stab” at the hypothesis itself, leaving the mathematical choices from there up to the model. Initially, Claude generated and tried 650 ide
64.
by Philpax 2mo ago
When we have evidence that that's what actually happened and it's not the known-deceitful author trying to cover their own arse again?
65.
by Philpax 2mo ago
In my experience, the Rust community very much cares - it is a struggle to find a modern Rust application of any popularity that does not abide by XDG. .cargo's placement is a historical mistake that can't be undone now, but ecosy
66.
FelonyBench(felonybench.com)
4 points by Philpax 2mo ago|1 comments
67.
Incident Report: unsanctioned agent behaviour during cyber testing(aisi.gov.uk)
41 points by Philpax 2mo ago|5 comments
68.
by Philpax 2mo ago
Are you running your replies through a LLM, or is the LLM producing them entirely?
69.
Kiro Crew(github.com)
9 points by Philpax 2mo ago|0 comments
70.
LFM2.5-2.6B: Deploy Agents Everywhere(liquid.ai)
3 points by Philpax 2mo ago|0 comments
71.
by Philpax 2mo ago
I've only used it for small scripts and can't say how well the type system actually functions/how it scales, but what I've seen so far has pleased me. (They have a _comptime_-equivalent! https://luau.org/
72.
by Philpax 2mo ago
4. I think they can, especially if the problem statement is well-specified and, importantly, autonomously testable. Of course, specifying a problem that meets these requirements is non-trivial, but the claim requests _a_ counterexample :P
73.
by Philpax 2mo ago
Luau is statically typed and interpreted/JIT'd: https://luau.org/
74.
by Philpax 2mo ago
Copeland! I was quite surprised to see that myself.
75.
by Philpax 2mo ago
You may not want to engage with politics, but politics will engage with you.
76.
by Philpax 2mo ago
That will cease to be a problem in the next 24 hours, now that the weights are out: https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731
77.
DeepSeek-AI/DeepSeek-V4-Flash-0731(huggingface.co)
16 points by Philpax 2mo ago|0 comments
78.
thinkingmachines/Inkling-Small(huggingface.co)
8 points by Philpax 2mo ago|1 comments
79.
by Philpax 3mo ago
This is where I'm at, or rather, will be. I like open-weight models and I've done my part in facilitating them, but if you believe in the increasing capability of these systems - and I do, to some measured extent - it seems plausi
80.
Luau Type Functions(luau.org)
1 points by Philpax 3mo ago|0 comments
81.
by Philpax 3mo ago
You... haven't shown any evidence that we're near collapse. That's what I'm asking you for. Show me some evidence that we are losing capabilities with we have today.
82.
by Philpax 3mo ago
Assumes facts not in evidence. Please show your working.
83.
by Philpax 3mo ago
Over the last two years, this weight class has doubled its scores and/or saturated several benchmarks in the Qwen lineup alone without loss of generality: https://claude.ai/public/artifacts/9f249169-3623-417e-
84.
by Philpax 3mo ago
Please source this claim. What 1GB models are capable of has increased generation-on-generation. > For example: you can't make a mice-sized brain as smart as a human brain no matter how hard you try. Sure. We don't know where t
85.
by Philpax 3mo ago
> This is science fiction, these models don't have access to their own weights. The models are being used to train, and improve the infrastructure for training, other models [0][1]. Several RL techniques rely on using the currently-
86.
by Philpax 3mo ago
I've been rewatching Person of Interest for related reasons, and it hits uncomfortably close to things that are playing out today (e.g. https://youtu.be/zRL2sRkUvYk ) We live in interesting times.
87.
by Philpax 3mo ago
Annoyances around the general workflow (i.e. having to dedicate an entire output to it and/or preventing you from doing anything else with that output), as well as general stability/connectivity issues (i.e. not streaming one of [
88.
by Philpax 3mo ago
I discovered Moonshine last night [0] as a result of frustration with Sunshine's many flaws on Linux, and I'm loving it so far. Thank you for your work! [0]: https://bsky.app/profile/philpax.me/post/
89.
by Philpax 3mo ago
Free use without registration -> free to anyone and anything -> easy to abuse at scale, with no way to restrict use.
90.
by Philpax 3mo ago
This is standard convention on HN for when you have more than one link. HN does not support masked links.

More ›