VHRanger

6 ms·

1.
by VHRanger 4mo ago
I'm curious why seemingly none of those projects tried using browsers JS/wasm execution as a sandbox instead
2.
by VHRanger 4mo ago
Nothing more reliable than good old SPLARGLE kitchenware
3.
by VHRanger 4mo ago
Do you have a better source if I speak fluent French by any chance? Thanks for the recap
4.
by VHRanger 4mo ago
A lot of those are spending 10-100k a month when they could be spending 1k-10k a month with a more lean infra
5.
by VHRanger 5mo ago
On vllm with a 5090 I get 120-180TPS with the awq 4 bit quant + MTP speculative decoding For gemma4 26B, same quantization, I get >200TPS. Also note that qwen is extremely inefficient in reasoning; the reasoning chains are ~3x longer tha
6.
by VHRanger 5mo ago
The issue is that benchmarks that look insightful will end up being gamed by labs quickly (Goodharts law) The best LLM benchmarks test around the margins of those behaviors, tasks that are difficult and correlate with usefulness while being
7.
by VHRanger 5mo ago
Radeon 9700 pro or intel arc b70 (both $1000-1400, 32GB, 650GB/s bandwidth), or ryzen AI max 390 (more vram, less bandwidth) The local inference space is pretty good nowadays.
8.
by VHRanger 5mo ago
The S3 API doesn't work like normal filesystem APIs. Part of it is that it follows the object storage model, and part of it is just to lock people into AWS once they start working with it.
9.
by VHRanger 5mo ago
That's not surprising; Opus & Sonnet have been regressing on many non-coding tasks since about the 4.1 release in our testing
10.
by VHRanger 5mo ago
There's presumably engagement on those two. It's better to have a smaller core of highly engaged people than a mass of disengaged eyeballs glazing over.
11.
by VHRanger 6mo ago
Pytorch is such a maddening mess of half implemented research features in a state of Heisen-deprecation, Jax becomes more appealing to me by the day.
12.
by VHRanger 6mo ago
Remember when we had the term "spyware" for a class of malware? I remember
13.
by VHRanger 6mo ago
If rust is not in the HN title and fire emojis in the readme, it doesn't come from the Rust region of France. It's just sparkling memory safe high performance software
14.
by VHRanger 6mo ago
Mexico has much lower economic development and higher crime. Canada has similar levels of economic development. It's really not that complicated: controlling for general crime levels, guns drastically increase murder rates.
15.
by VHRanger 6mo ago
Yeah I'm another pop os user. Cosmic works great for a laptop. But it's a PITA for a desktop. It doesn't deal with multi monitor setups well. There's a recent new bug where the system hardlocks on monitor power state cha
16.
by VHRanger 6mo ago
Of course within the USA the state levels laws will do little. There's free movement between states! Compare the USA to Canada, where you can't bring a gun easily. You'll see Canadian murder rates being very low. Even control
17.
by VHRanger 6mo ago
Sure poverty explains crime, and murder is the ultimate crime. That said, my point was that a place like Rio, where you feel alertness at a physiological level by the constant lack of security, still has a murder rate around Houston, a vast
18.
by VHRanger 6mo ago
Ah, some egregious misuse of statistics! Murder is a byproduct of crime. Crime is, largely, downstream of economic conditions with some obvious caveats. New Hampshire has the 2nd lowest crime rate of the USA states. You could make the same
19.
by VHRanger 6mo ago
And I presume that doubled price is before you look at the workstation class motherboards, which also tend to be much more expensive. Thanks for the info on the hardware quirks, useful to know! We seem to be arriving at a cambrian explosion
20.
by VHRanger 6mo ago
It's really hard to push DDR5 past 6000MT/s on 4+ DIMMs it seems. I had to get everything top spec to fit 4 channels of 6000MT/s on my 9950x (asus proArt motherboard and the top tier trident neo RAM sticks) -- otherwise it&#x
21.
by VHRanger 6mo ago
Youre aware that the rest of the planet have stricter gun laws and the American problems are fairly unique? This is even after controlling for things that exacerbate crime like high economic inequality. For instance, Brazil [1] (a much poor
22.
by VHRanger 6mo ago
GDDR7x doesn't come in dimm factor? In general soldered ram seems to get much higher bandwidth than removeable ram. See ryzen AI Max vs 9950x max ram throughputfor example
23.
by VHRanger 9mo ago
paperless-ngx?
24.
by VHRanger 9mo ago
This t5 is multimodal. Also a hint: you can create a finetuning dataset from a frontier LLM pretty easily to finetune those t5 and effectively distill them pretty fast these days
25.
by VHRanger 9mo ago
In general encoder+decoder models are much more efficient at infererence than decoder-only models because they run over the entire input all at once (which leverages parallel compute more effectively). The issue is that they're general
26.
by VHRanger 9mo ago
Encoder/decoder is much, much more efficient for finetuning and inference than decoder-only models. Historically T5 are good when you finetune them for task specific models (translation, summarization, etc).
27.
by VHRanger 9mo ago
A store of value is an asset with as close to 0% volatility in price as possible. Bitcoin is a speculative asset: it has very high price volatility. It is not a store of value in the proper term.
28.
by VHRanger 10mo ago
Same, I got 96GB of high end 6000MHz DDR5 this summer for $600CAD and now it's nearly triple at $1500CAD
29.
by VHRanger 10mo ago
Kagi has Mistral as well
30.
by VHRanger 10mo ago
If you send a message in our discord we can give you a beta key I think. Costs nothing to ask over there, say Matt R sent you

More ›