wenc

34 ms·

571.
by wenc 3y ago
> In one news release regarding the cashless policy at Death Valley, the park service said that the $22,000 in cash collected by the park during the previous year took more than $40,000 to process. > “Cash handling costs include an ar
572.
by wenc 3y ago
Brazil also has an interesting relationship with teeth. When I was there, I noticed people brushing teeth after lunch. This happens even at corporate offices -- the restrooms would be full of people brushing and flossing after meals. People
573.
by wenc 3y ago
That said random forests are an internal ensemble, so I guess that could work. In my mind an ensemble is like a committee. For it to be effective, each member should be independent (able to pick up different signals) and have a greater than
574.
by wenc 3y ago
Dropout is different from ensembles. It is a regularization method. It might look like an ensemble because you’re selecting different subsets but ensembles combine different independent models rather than just subset models.
575.
by wenc 3y ago
Thank you for catching that. Corrected.
576.
by wenc 3y ago
They are comparing a non-ensembled transformer model with an ensemble of simple linear models. It's not surprising that the ensemble models of linear time series models will do well, since ensembles optimize for the bias-variance trade
577.
by wenc 3y ago
Apple Delays Work on Next Year’s iPhone, Mac Software to Fix Bugs https://www.bloomberg.com/news/articles/2023-11-07/apple-del...
578.
by wenc 3y ago
TIL. Most of those streetcar suburbs on that list that I'm familiar with are actually pretty dense. They would have been considered neighborhoods in a big city (analogous to say Lincoln Park in Chicago).
579.
by wenc 3y ago
DuckDB is ok at JSON data but it hasn’t worked around all the corner cases like Pandas has. Converting from a hierarchical format like JSON to a tabular format is always fraught. I find Pandas to be slightly more robust and faster. However
580.
by wenc 3y ago
From this: https://www.statista.com/statistics/1202559/japan-rent-for-a... About US$755.
581.
by wenc 3y ago
S3 is really high latency though. I store parquet files on S3 and querying them through DuckDB is much slower than file system because random access patterns. I can see S3 being decent if it’s bulk access but definitely not for random acces
582.
by wenc 3y ago
Tokyo actually builds a lot of new housing which accounts for why housing is affordable there. People think Tokyo is an expensive city -- and relatively speaking it is -- but it's a lot cheaper than the small US city I'm living in
583.
by wenc 3y ago
I mean it would be really nice if we solve really large scale optimization problems (mixed integer nonlinear) to global optimality, fast. But we already do today with some gross approximations and some suboptimality.
584.
by wenc 3y ago
Least squares is quadratic. Quadratic means square terms.
585.
by wenc 3y ago
I’ve never used RapidMiner so I couldn’t say. So Tableau isn’t a full fledged statistical tool. It does have all the basic statistical functions but it doesn’t really focus on that. What it is is a visualization tool. You can actually see a
586.
by wenc 3y ago
Could you list a few? I tried looking once but nothing came close to Tableau.
587.
by wenc 3y ago
I've been using Tableau since 2013 to deploy dashboards, which in my case are front ends to ML model output. (for me, I used it as a Streamlit before Streamlit existed) People think of Tableau as a dashboard tool, which it is. But it&#
588.
by wenc 3y ago
I hear what you're saying, but I have a different experience. I prefer to order from a kiosk and to scan my own groceries. It's not like I don't like the human interaction, but when I have only a few items, or I want to be ab
589.
by wenc 3y ago
They work well if there are lots of people like you (ie you are a normie). All machine learning algorithms struggle at the edges — they’re very good at predicting aggregate behavior. If you have eclectic tastes there’s probably not enough d
590.
by wenc 3y ago
> Personally I had a very hard time even breathing in then holding my breath for 3~4s (heart pounding, growing panic state), so I found that "anything goes" helped me a lot in achieving 4-7-8, and in turn 4-7-8 helped me a lot
591.
by wenc 3y ago
Surprised they didn't call out box-breathing. It works for me. It's so simple: 1. inhale 4 seconds. 2. hold breath for 4 seconds. 3. exhale 4 seconds. 4. hold breath for 4 seconds. (repeat) I do this when I encounter any stressful
592.
by wenc 3y ago
CSS sticky headings (the solution) won't work on the user agents that don't support it anyway, so they wouldn't be missing out. :)
593.
by wenc 3y ago
What? Most people actually do want a second-hand car. https://www.statista.com/statistics/183713/value-of-us-passe... I've never bought a new car in my life. A new car is great but loses on average 10% of its
594.
by wenc 3y ago
Paul Graham wrote an article about the difference between a Maker’s schedule vs a Manager’s schedule. [1] Powerful people are on a manager’s schedule. Meetings are a unit of work for a manager, and they freely schedule meetings because ther
595.
by wenc 3y ago
I'm curious where you did French immersion? As far as I can tell what my friends' tell me (their kids are in French immersion in Ontario), French immersion teaches standard French, not the Quebecois patois (which we call le joual)
596.
by wenc 3y ago
Sure. Just ask ChatGPT 4 (not the free 3.5) to help you formulate some simple MIP constraints. My prompt was: (and it gave me the right answer) "I have to formulate an MIP. I have a list of items i \in I each belong to groups g \in G.
597.
by wenc 3y ago
I’m not proving anything but I’m using ChatGPT to come up with MIP optimization mathematical formulations. It’s ingested all the techniques from journal papers so it actually comes up with some really solid formulations based on what I tell
598.
by wenc 3y ago
Thanks for pointing that out. It was my mistake (corrected in parent post)
599.
by wenc 3y ago
> come up with the most efficient data structure to store highly sparse data (say vector), that will need to be access only sequentially. There's something even more efficient -- a sparse array. I worked in the sparse linear algebra
600.
by wenc 3y ago
One thing it doesn’t do which matters to me is properly rendered math like ChatGPt does. I asked it for mathematical formulations and so far the answers are not as good as ChatGPT 4.

More ›