quadrature

7 ms·

61.
by quadrature 3y ago
I don’t think so. Karpathy is incredibly productive, I’m assuming they’re referring to obsidian being open but Karpathy having worked for OpenAI on closed source AI.
62.
by quadrature 3y ago
nope, google had a discussions tab in search. Looked something like this https://www.ghacks.net/wp-content/uploads/2014/01/discussion...
63.
by quadrature 3y ago
Is your workload an analytical or transactional workload ?.
64.
by quadrature 3y ago
Its not about being price conscious, its about finding an efficient way to replicate data into your data warehouse.
65.
by quadrature 3y ago
I haven't used it but polars has a lazy api https://docs.pola.rs/user-guide/lazy/using/#using-the-lazy-a...
66.
by quadrature 3y ago
Thats pretty cool, what kind of stuff do you work on ?.
67.
by quadrature 3y ago
Yeah strings are very expensive, saw huge gains in my submission after removing them. Parallelization gets the runtime down to 30s , After removing strings you get to around 15s on the eval machine, the rest of the tricks are still very rel
68.
by quadrature 3y ago
> It's either 6 minutes with 20 lines of AWK written in five minutes. Or 6 seconds in 300 lines of java written in five days. Classic "XKCD: Is It Worth the Time?". Not quite, it runs in 3 minutes written in java in 5 minu
69.
by quadrature 3y ago
It depends on the objective. You can test stuff like is there any set of navigations that results in an exception in the backend. This is what Sapienz from facebook did, it fuzzed their mobile applications to find paths that caused errors a
70.
by quadrature 3y ago
Apologies, I thought you were asking a general question about ML. Will let someone else comment on the specifics here.
71.
by quadrature 3y ago
Yes. Think of it in terms of fitting a line to some points y=mx+b. Training is finding the right slope m and intercept b of the line to get a good fit to the points. Inference is when you take an x coordinate and find the y value using the
72.
by quadrature 3y ago
not made up as in generated. it was sampled from different instances of time. This is the article's explanation > Coates was moving when the photo was taken, so when the shutter was pressed, many differing images were captured in th
73.
by quadrature 3y ago
1) Orders are complex and require many steps to process correctly. 2) Orders are only the writes, we have to serve many magnitudes more read traffic and not all of it is cached (invalidations, inventory counts).
74.
by quadrature 3y ago
What kinds of issues did you see?. Do you think there are better alternatives to the disruptor ?
75.
by quadrature 3y ago
Good point, i like the balance houdini has here where you can drop into vex when you want.
76.
by quadrature 3y ago
Agreed, It's really impressive how well they nailed the atmosphere of Ian Hubert's work.
77.
by quadrature 3y ago
They likely created a large portion of these assets themselves. There might be some kitbashing but the amount of detail here is certainly doable for a motivated 16 yr old. Theres a lot of creative ways to add details using scans, texture pr
78.
by quadrature 3y ago
from https://arxiv.org/abs/2310.04378 it sounds like its a form of distillation of an SDL model. So im guessing it can't be directly trained, but once you have a trained diffusion model you can distil a predictor
79.
by quadrature 3y ago
maybe its tracking your hand and aiming the projection ?. Can't really say much based on the landing page, I doubt most of the interactions are as smooth as they are pictured. But it is refreshing to see a new take on mobile devices.
80.
by quadrature 3y ago
Fair, but this is how innovation happens. It's built on the shoulders of previous research.
81.
by quadrature 3y ago
Very cool, thanks for sharing. Anything.world looks very useful, great work on this release!.
82.
by quadrature 3y ago
I don't know how the rigging works, but the animation might follow some principles from learned motion matching https://www.ubisoft.com/en-us/studio/laforge/news/6xXL85Q3bF... I don't think it
83.
by quadrature 3y ago
Hadn't heard of that dagster feature, thats really cool.
84.
by quadrature 3y ago
This is very cool!. Last time I checked duckdb didn't have the concept of a metastore so do you have an internal convention for table locations and folder structure ?. What do you use for reports/visualizations? notebooks ?.
85.
by quadrature 3y ago
If anyone here is using DuckDB in production i'd love to hear what your stack looks like over the entire lifecycle of extract->transform->load.
86.
by quadrature 3y ago
The App stores provide distribution for software. This is network, storage and eyeballs. Unity doesn’t provide any value for the marginal install that it is charging for. A better comparison would be to Unreal which charges a royalty after
87.
by quadrature 3y ago
Thanks for the reply!. This is really interesting research, hope to see more from your team.
88.
by quadrature 3y ago
Exactly, since its not doing whole program synthesis im thinking it could be done with fewer parameters. However program synthesis is part of the loss function.
89.
by quadrature 3y ago
I wonder if this requires a 7B parameter model.Assembly has a small grammar and is very constrained compared to natural language.
90.
by quadrature 3y ago
Anyone know if there is something similar for games ?. my parents like playing classic games on their phones but they're usually filled with scammy ads.

More ›