jszymborski

4 ms·

31.
by jszymborski 2mo ago
My bike is $100... If I can take off the $2k electric bike conversion kit and lug it with me, I'd sleep a lot better
32.
by jszymborski 2mo ago
I think they'd prefer it replaced with female in every and all cases :(
33.
by jszymborski 2mo ago
no, this is the appropriate response to hearing the words "HN" :P
34.
by jszymborski 2mo ago
Likely inverted "meta" to avoid collision with HTMLs meta tags
35.
by jszymborski 2mo ago
As is this whole blog post... Very grating prose
36.
by jszymborski 3mo ago
This is indeed how it works in some countries, but depending on how strong a nations social net that isn't necessarily true.
37.
by jszymborski 3mo ago
Beyond all the other more salient objections one might have with this pitch, there is just the very practical fact that there really isn't enough signal from "noninvasive" sensors to get the sort of data they are promising. I
38.
by jszymborski 3mo ago
It's not easy to disentangle and it varies by country and institution, but those positions are not normally directly funded by public research grants.
39.
by jszymborski 3mo ago
> Scientific discovery is bottlenecked. Yah, by funding and how we award it, not by an imaginary lack of undergrad and grad students. Scientific funding requires a shotgun approach and many national science funds try to pick winners as o
40.
by jszymborski 3mo ago
Super impressive results, and an awesome write up. Thanks :)
41.
by jszymborski 3mo ago
From the article > In the days of generative image models, it's still fun to see what can be achieved using traditional image transformation techniques. > But the really fun part of the project was in combining a number of techni
42.
by jszymborski 3mo ago
margin:0 auto is the (probably incorrect) incantation I have always reached for
43.
by jszymborski 3mo ago
If the error bars are so large that the old and new analyses aren't statistically different, then the margins are too uncertain to make any claim in the original work... which is sorta the claim that the new work is making.
44.
by jszymborski 3mo ago
It's annoying they didn't plot the new and old "perceived" curves on the same figure, but if you pay attention to the y-axis, there is a very big difference. In the old plot, the bottom quartile has about a 50 percentage
45.
by jszymborski 3mo ago
I've switched over to unsloth studio and am hitting those speeds now, thanks for the tip.
46.
by jszymborski 3mo ago
bahaha, LMStudio's badge looks a lot like a B but the tooltip says it is indeed an 8 (I just assumed it was B for "byte", but that's a poor initialism considering "bit"). https://dl.jszym.com/sh
47.
by jszymborski 3mo ago
No, it's what I get repeatedly. Recall that it's QB_0 though. I'll get back to you when the Q4_K_M quant finishes downloading. EDIT: I get 259 tok/s with the Q4_K_M quant
48.
by jszymborski 3mo ago
I think sometimes the quantization is the source of the conflicting reports since those are often not included in discussions. Honestly, 27B is too slow for me to use very often, so I just do 35B-A3B and if it is underperforming I'll s
49.
by jszymborski 3mo ago
Thanks for flagging, this is on my local rig and it's driving my display too. I'm curious now, will take a closer look. These are the tok/s as reported by LMStudio. EDIT: Updating llama.ccp gets me 58 tok/s on Gemma 31b
50.
by jszymborski 3mo ago
On my RTX 5090, Qwen3.6-35B-A3B is my go to for coding, but Gemma-4-26b-a4b (Q4_0) is blazing fast and super high quality. I sometimes use the dense cousins of these models (Qwen3.6-27B QB_0 and Gemma4-31B-QAT Q4_0) but they can be very slo
51.
by jszymborski 3mo ago
I don't even know where to begin with this comment...
52.
by jszymborski 3mo ago
First and foremost, if DHH meant "junkies" or "homeless" by "gypsies", a term widely regarded as derogatory used to describe Romani people [0], then already there is plenty to take umbrage with. Second of all,
53.
by jszymborski 3mo ago
K2.6 is cheaper than GLM5.2 (at least on DeepInfra) and I've found it works as good as Sonnet for my purposes. Both tend to think themselves into circles a bit and aren't super token efficient, but I've found GLM5.2 much wors
54.
by jszymborski 3mo ago
Datacenters that have not and might not be built to accommodate for future AI demand that may or may not grow to the extent or as fast as the companies anticipate.
55.
by jszymborski 3mo ago
In fairness, the original crop of detectors were based on perplexity scores and were entirely useless (famously, the declaration of independence often came back as 100% AI generated). I'm not convinced we'll ever have full-proof d
56.
by jszymborski 3mo ago
This comment is for like three people on HN, but anybody remember when Paul Neave had a website/blog/showcase (c.2007-2010ish?) where the borders of the page had similar jelly effect? That was achieved with Flash back in the day.
57.
by jszymborski 3mo ago
For open weight models, you can choose from a few providers. Each have their own caveats, none of ToS'/Privacy Policies I entirely trust, nor do many make renewable energy claims.
58.
by jszymborski 3mo ago
Big if true! One of the caveats here is that the file size seems to balloon 341 times.
59.
by jszymborski 3mo ago
I'm just saying they should just drop the dubious claims and just say "I made a font that I think looks cool".
60.
by jszymborski 3mo ago
I think this illustrates that you can just do stuff without claiming it is useful. Like couldn't you just make this font and call it something like double-entendre or something?

More ›