feoren

68 ms·

901.
by feoren 3y ago
Completely agree, but you're implicitly assuming that developers have a lot of leeway in doing this design in the first place. Many non-technical managers seem to assume that developers should not have any say in the design and functio
902.
by feoren 3y ago
Very good points here, especially about the fact that the single "length" degree of freedom is much less to lose in very high-dimensional spaces. However, I don't agree that large vectors would end up being "more similar
903.
by feoren 3y ago
> I had actually rang alarm bells about him during the interview process but was overruled by HR because his resume was so amazing. Yikes. I have a great cost-cutting proposal for your company, then: don't bother with interviews. Ju
904.
by feoren 3y ago
> Is it that you just dislike it when other people exercise their rights? Is it that you don't bother to exercise those same rights yourself? Do you think this works as well when you replace "exercise one's rights" wi
905.
by feoren 3y ago
70% of the way through this article they drop "Since we have unit norm vectors ..." and later "Always remember to normalize your embeddings." I found this strange and surprising. It seems to me that in a semantic embeddi
906.
by feoren 3y ago
If you turn 100kg of CO2 into Methane you get about 36kg. Methane is somewhere around 25x (potentially 31x) worse for global warming than CO2 by weight, so you end up with 909kg of "Carbon Dioxide equivalent", or CO2e. Let's
907.
by feoren 3y ago
> It is allowed to make a profit if it does. A nonprofit is subject to the non-distribution constraint: any revenues that exceed expenses must be committed to the organization's purpose, not taken by private parties. https:/&#
908.
by feoren 3y ago
> 2. Have you contributed funds to OpenAI? If not, where did your sense of entitlement come from? Is your argument that we need to have given a company money before we can be opposed to their unethical business practices? We're talk
909.
by feoren 3y ago
This is the age-old favorite toddler game to enjoy flummoxing their parents. Somehow whenever mine does this with me we end up on "well, because the early makeup of the solar system ... ". Similarly to how repeatedly following the
910.
by feoren 3y ago
Is there? I scrolled through the readme 4 times looking for one and couldn't find it.
911.
by feoren 3y ago
So 90,000 times smaller. At first glance you made it look 90x smaller because I couldn't tell if "mcg" was a typo for "mg" or not (without clicking the link). I've never seen anyone use "mcg" to mean
912.
by feoren 3y ago
We can decide to make laws based on the society that we all want to live in. If our objective is maximizing the amount of creative work available to society, it may make sense to say that creators of those works enjoy monopoly on the reprod
913.
by feoren 3y ago
Have you read any of the sources you linked to? They debunk most of what you just said. "Ultra pure" water is very slightly more "watery" than regular tap water. Any effect that tap/bottled water has on anything,
914.
by feoren 3y ago
> Every CEO I’ve worked under—no matter how compassionate they were on a personal level—has had a policy of absolutely crushing voices of dissent. I think there's a really interesting confluence of factors here. (A) CEOs tend to be
915.
by feoren 3y ago
> you will get a 4xx/5xx response code, but you will still get billed for the request and whatever the servers generated and you didn't get. That's borderline fraudulent. Borderline!? They're regularly charging custom
916.
by feoren 3y ago
Have you tried gluten free ginkgo biloba bee pollen salt lamps? Sorry, I had to. But here's an actual real suggestion that may or may not be any better. It's a working memory trainer that I feel has slightly helped improve my own
917.
by feoren 3y ago
Which is why I go out of my way to avoid advertisements as much as possible.
918.
by feoren 3y ago
I really like early returns for this reason. if (someEdgeCase) { return SpecialValue; } const normalCase = a + b ... It allows you to drop a case from working memory as early as possible. I believe this would be a +1 in cyclomatic co
919.
by feoren 3y ago
In fact, the "factoring out" process shouldn't even be that hard: find the input vector that forces the ANN to output the copyrighted work verbatim. There should be some simple method of "baking in" the first step o
920.
by feoren 3y ago
An mp3 encoding of a wav file of a copyrighted song is still copyrighted, despite those exact bits never having existed before, and being created entirely by a computer.
921.
by feoren 3y ago
Some thought experiments: What happens if we train a neural network on a single, copyrighted work? Say it has one input node (or even zero, if you like), and regardless of this input, its output is always exactly the copyrighted work it was
922.
by feoren 3y ago
Compiled object code of a bunch of code you didn't write . I don't know why programmers are so eager to forget that copyright is not at all about what something is , and all about where it came from . It'd be hard to asser
923.
by feoren 3y ago
I think the biggest problem with this utter nonsense is this: > In other words, one can assume with 95% certainty that any individual human would be within the last 95% of all the humans ever to be born "any individual human" r
924.
by feoren 3y ago
For those who disagree, please give an example of a domain that is devoid of important "concern-based" operations.
925.
by feoren 3y ago
I don't see that as disallowing "for those who disagree with me: what do you disagree with, exactly?" Isn't that a reasonable question? Besides, isn't downvoting someone for nonsensical, petty reasons (and leaving n
926.
by feoren 3y ago
> When someone suggests this kind of thing, I'll ask them "how do we diagnose performance problems with this technology when there are 100,000 concurrent users and millions of data elements?" I don't understand; the e
927.
by feoren 3y ago
But it is true, much much more often than you probably realize. Just look at your tables and think about how repetitive they are. The reason you can't come up with a lot of "column-based" (as you put it, which is still narr
928.
by feoren 3y ago
The actual best process: talk to each other. Every other set of rules about how/when to branch/merge/PR/rebase feels like an attempt to get around having to just actually talk to each other and know what's going on.
929.
by feoren 3y ago
Management Theory Null: All management is fundamentally parasitic and only exists to perpetuate the organizational structure designed to extract as much wealth as possible from both the Actual Contributors (preferred over "Individual C
930.
by feoren 3y ago
> The correct solution is to come up with some kind of legal regime that's actually designed for software Sure, I agree with that. You told me to revise my expectations for something much less unrealistic than that, though. But let&

More ›