mjhay

12 ms·

151.
by mjhay 2y ago
An oxidizing atmosphere (as Earth and Mars are currently) would actually make it much harder for life to emerge in the first place. Oxidation makes it very hard for complex molecules to remain stable enough for life to emerge in the first
152.
by mjhay 2y ago
Boeing destroyed its engineering culture and it isn't going anywhere.
153.
by mjhay 2y ago
SpaceX is in a much more solid position. You can tell that Musk hasn't interfered in its operations nearly as much as he has with Tesla or X (case in point: renaming it "X").
154.
by mjhay 2y ago
I've still never been convinced by the use cases of graph databases. It just seems like a new coat of paint over NoSQL. I always hear vague claims of increased performance (the same thing everyone said about NoSQL), without much backin
155.
by mjhay 2y ago
Wow, that's impressive! I wouldn't expect anything to be able to beat MKL, given the optimizations made based on proprietary information.
156.
by mjhay 2y ago
Of course somebody working for OpenAI would say that.
157.
by mjhay 2y ago
IMO, the Mandate of Heaven gets over-emphasized in a lot of Chinese historiography. The bestowal of the Mandate of Heaven was usually seen as occurring once a ruler controlled all or enough of Chinese territory, and disappeared once it was
158.
by mjhay 3y ago
There's almost certainly porn and nipples in the training set. It's pretty hard to eliminate 100% of NSFW images. That means there is always the possibility of something like this happening.
159.
by mjhay 3y ago
There's always been gambling, but there's a lot more of it now. Gross revenue has more than doubled since 2020. I think it's safe to say gambling's new-found ubiquity is a big part. https://www.statista.com&
160.
by mjhay 3y ago
no thanks!
161.
by mjhay 3y ago
A Modest Proposal for Preventing Poor Creatives from Being a Burthen to Tech Professionals or Country, and for Making Them Beneficial to the Publick
162.
by mjhay 3y ago
xgboost will still work better for most problems people encounter in industry (which usually involve tabular data).
163.
by mjhay 3y ago
There's also been some work on more general Lie-group equivariant transformer models. http://proceedings.mlr.press/v139/hutchinson21a/hutchinson21...
164.
by mjhay 3y ago
Counterpoint: Logo redesigns give management something to endlessly bikeshed about in meetings.
165.
by mjhay 3y ago
What is the deal with school administrators? I've never interacted with a profession that incompetent and self-important. Some, I assume, are good people. > Fresno High is the latest school in the district to roll out the 5 Star Stu
166.
by mjhay 3y ago
That actually does happen. Industrial pig and chicken farms often have manure lagoons, which are exactly what they sound like. Do not do an image search for manure lagoons. There's been quite a few instances of people falling in and be
167.
by mjhay 3y ago
I am in 110% agreement with everything you are saying here. There's a quite bizarre thing here, though: In typical American car-centric areas, suspending someone's license because of debt takes away their ability to pay back deb
168.
by mjhay 3y ago
Worries about pipeline leaks and climate change are not silly.
169.
by mjhay 3y ago
Calling library() is kind of an antipattern in production R code. You can either call namespaced functions (like say dplyr::mutate()), or use roxygen. https://roxygen2.r-lib.org/articles/namespace.html
170.
by mjhay 3y ago
reshape2 (where `melt` is from) has been deprecated for some time, and for pretty good reasons. Try dplyr and tidyr instead - they are much nicer and modern. The equivalent of melt would be pivot_longer. For packaging, renv is the usual cho
171.
by mjhay 3y ago
Yeah, the package management situation is a big weak spot. There are some issues with renv, but it is usable. It definitely helps to keep a lid on the number of dependencies, and for God's sake never pull anything in from Bioconductor.
172.
by mjhay 3y ago
There's nothing about RStudio that encourages big single files or writing huge unstructured scripts. RStudio is a pretty good IDE, and R is a highly expressive functional-first [0] language. R was heavily influenced by Scheme, and has
173.
by mjhay 3y ago
Not if the GP is describing the rate of change of the probability of transmission as a bell curve. We can assume 0% transmission when the bite first occurs, and assume that chance gradually plateaus at some point. That would look like an S-
174.
by mjhay 3y ago
The Nazi's SD and Abwehr intelligence services were incredibly incompetent, as this case shows. The Abwehr did have a relatively high concentration of anti-Nazis, which didn't help either.
175.
by mjhay 3y ago
This sounds like just sensational panic being exploited by Hochul to distract from her incompetence. The MTA is a mess, but it's a mess because of mismanagement by the state of NY. Crime has been trending downward. There was an increas
176.
by mjhay 3y ago
OCD tendencies can be good when it comes to programming.
177.
by mjhay 3y ago
> It appears that many ancient counterfeiters used “fire gilding” to apply the coating. > Mercury readily forms a very soft alloy (called an “amalgam”, it has the consistency of butter) with gold or silver. This was smeared evenly ove
178.
by mjhay 3y ago
Interesting. Usually I have better luck with xgboost for tabular data, even when the relationships are complex (which usually means deeper trees). It does fall flat a lot of the time for very high dimensions, though. All data is different,
179.
by mjhay 3y ago
That's an advantage over linear models, but GBTs handle non linearly-separated data just fine. Each individual tree can represent an arbitrary piecewise-constant function given enough depth, and then each tree in turn tries to minimize
180.
by mjhay 3y ago
Still mostly NLP and image stuff. Most actual data in the wild is tabular - which GBTs are usually some combination of better and easier. In some circumstances, NN can still work well in tabular problems with the right feature engineering o

More ›