LarsDu88

32 ms·

751.
by LarsDu88 2y ago
Made a mistake here and have since corrected it
752.
by LarsDu88 2y ago
I go over the downsides in the article, but there is nothing fundamentally wrong with using pip and venv. It's just that if you've ever worked in other programming ecosystems, it would be immediately apparent that things could be
753.
by LarsDu88 2y ago
I lol'd at this. At my last job, we lost days if not weeks to "solving environment" due to poetry and many folks not understanding that ceiling pinning is BAD when you use a tool like poetry to manage your deps.
754.
by LarsDu88 2y ago
Original author here... wait, what? I believe pip (or at least older versions of pip) makes it possible to have cycles. Are you analyzing dep graphs made with older versions of pip?
755.
by LarsDu88 2y ago
I made a mistake about Rye when I posted the first draft of this article. That has since been corrected. Rye is incredibly young, and when I first encountered it, I looked at its source code and found it was simply calling piptools under th
756.
by LarsDu88 2y ago
I mention this ceiling pinning footgun in the article. It's an enormous pain in the ass to explain to folks, and some software engineers I've met are totally incredulous that that's "not the right thing to do" Poetr
757.
by LarsDu88 2y ago
fixed
758.
by LarsDu88 2y ago
I just wrote a post on the myriad python package management options that you might find interesting: https://dublog.net/blog/so-many-python-package-managers/
759.
Python Has Too Many Package Managers(dublog.net)
131 points by LarsDu88 2y ago|163 comments
760.
by LarsDu88 2y ago
Python actually now has over a dozen package managers, none of which are as good as Cargo in Rust. Here's my attempt at a nearly comprehensive rundown.
761.
by LarsDu88 2y ago
The book Loonshots says as much. Extrapolating from monkey brain sizes to human social interactions is a bit reminiscent of phrenology IMO, but it is an interesting observation. Anecdotally, I've found it to hold true, particularly in
762.
by LarsDu88 2y ago
This might be a bit of an isolated example. Numerous battles against the Mongols and Turks in the middle ages start with a haphazard charge, followed by the cavalry getting out of position and completely crushed. Seems like the success in t
763.
by LarsDu88 2y ago
Have you ever read the book "Loonshots" There's an entire chapter about how primitive human societies probably maxed out at about 150 individuals. Companies that exceed this headcount basically need to change themselves into
764.
by LarsDu88 2y ago
I should note that its only something like 1/3rd of Tibetans who have this variant, and something like 1/200 Han Chinese, Vietnamese, Laotians, etc who have this variant (If you go back tens of thousands of years, basically everyo
765.
by LarsDu88 2y ago
Enough diagnostic tools? Hah. Most cancer detection assays are pretty much garbage if you were to compare them to the latest advances in say deep learning image recognition models. You can actually compare the sensitivity/specificity f
766.
by LarsDu88 2y ago
This is really interesting for another reason (which is kind of buried at the end of this article). A variant of the gene EPAS1, which occurs at low frequencies in east asia, and relatively high frequencies on the Tibetan plateau seems to b
767.
by LarsDu88 2y ago
This idea of "curing cancer" as if cancer was a monolithic disease really needs to die. Cancer is just a catch-all term for somatic cell mutations that become unmanageable. There are as many ways to get "cancer" as there
768.
by LarsDu88 2y ago
The little anecdote about Peter Thiel about shitty physicists going into biology reminded me of a chapter from Richard Feynman's book, "Surely Your Joking Mr. Feynman". So it turns out Richard Feynman, after working in Los Al
769.
Semantic encoding during language comprehension at single-cell resolution(nature.com)
2 points by LarsDu88 2y ago|1 comments
770.
by LarsDu88 2y ago
New nature paper shows tungsten microelectrodes inserted into single neurons of conscious humans reveals individual neurons encoding words into semantic encodings (e.g. responding to only certain semantic categories but not others)
771.
by LarsDu88 2y ago
That's funny. I recall years ago watching a documentary on the Vindolanda letters stating that a complement of Syrian archers from Hama (the area just north of Homs) was stationed in Britain during Hadrian's time: https:/&#x
772.
by LarsDu88 2y ago
You should try the quest3. Virtually no setup
773.
by LarsDu88 2y ago
I honestly don't think we've seen the end of AR/VR yet. The tech continues to improve year over year. There are rumors the prototype Zuck plans to show at Meta Connect this year are mindblowing
774.
by LarsDu88 2y ago
I had to take a second look at this: https://www.datacenterdynamics.com/en/news/meta-to-operate-6... 340,000 H100s 600,000 H100 equivalents (perhaps AMD Instinct cards?) On top of the hundreds of thousands of lega
775.
by LarsDu88 2y ago
I think people wildly underestimate how protectionist people - particularly educated software engineers and PhDs will get once an AI model directly impacts their source of wealth. Red state blue collar workers got their candidate to pass ta
776.
by LarsDu88 2y ago
Ok, I might have misread some rumored ballpark figures. And most of the GPUs will be used for inference rather than training. Still 6 GPT-4's every 90 days is pretty amazing.
777.
by LarsDu88 2y ago
True that, but I think in a very short amount of time, using dedicated general purpose GPUs just for inferencing is going to be mega overkill. If there's dedicated inferencing silicon (like say the thing created by Groq), all those GPU
778.
by LarsDu88 2y ago
According to Jensen it takes about 8000 H100s running for 90 days to train a 1.8 Trillion param MoE GPT-4 scale model. Meta has about 350,000 of these GPUs and a whole bunch of A100s. This means the ability to train 50 GPT-4 scale models ev
779.
by LarsDu88 2y ago
I do some 3d modeling on the side with my side project ( https://roguestargun.com ), and I suspect those 800 polygons with good topology may be more useful to a lot of 3d artists than blobby fully textured SDF derived models. A lo
780.
by LarsDu88 2y ago
This is crazy impressive, and the fact they have the whole thing running with a PBR texturing pipeline is really cool. That being said, I wonder if the use of signed distance fields (SDFs) results in bad topology. I saw a paper earlier this

More ›