adderaline

listen. even if we disregard the fact that lots of legal experts, including the peers of the people who put this ruling in place, believe this is an existential threat to democracy, in practice, the ruling puts the authority for determining what is an official act into the hands of the judiciary. the supreme court is the ultimate authority in making these determinations. its a power grab, plain and simple, which grants the president immunity for “official acts”, and places the authority to determine what is and isn’t an “official act” in the hands of the same people who granted him that immunity.

the fact that Roberts is making vague gestures towards some kind of accountability means less than nothing. considering how Trump is behaving, what he and his crowd seem to believe about the breadth of this decision, we should not assume that a room full of people Trump put into power have any interest in ensuring Trump doesn’t “get away with everything”, and we shouldn’t assume that these people are even nominally interested in telling the truth about their intentions, considering just how much of their personal comfort is guaranteed by the institutions that Trump represents, and how resistant they are to accountability for their extremely well documented lies.

your personal confidence in Trump’s eventual, eternally forthcoming guilt relies on the trustworthiness of liars and the moral fiber of bigots. good luck with that.

Open models is the way to battle that.

This is something I think needs to be interrogated. None of these models, even the supposedly open ones are actually “open” or even currently “openable”. We can know the exact weights for every single parameter, the code used to construct it, and the data used to train it, and that information gives us basically no insight into its behavior. We simply don’t have the tools to actually “read” a machine learning model in the way you would an open source program, the tech produces black boxes as a consequence of its structure. We can learn about how they work, for sure, but the corps making these things aren’t that far ahead of the public when it comes to understanding what they’re doing or how to change their behavior.

plenty of people?

whoops! i was pretty high when i wrote that. guess i just felt like sharing.

not to dig this hole any deeper, but the defining characteristics of a chicken aren’t like, easily identifiable. we can build a hypothetical in which two proto-chickens are genetically capable of producing offspring that is “chicken”, but that’s kinda rube-goldbergesque, there must have been some extremely specific series of genetic coincidences required to produce something chicken enough to be a “chicken” in that scenario. genetics, and evolution more generally, tends to be more complex. the specific genetic markers that distinguish chicken from non-chicken, if we say they exist, are probably not in and of themselves what makes a chicken, because single gene changes don’t usually make creatures incapable of interbreeding with their parents’ species, and that’s a defining feature of the taxonomic category “chicken” belongs in.

like, if we grant that the chicken came from a proto-chicken egg, because the chicken has a special chicken gene, its really really likely that the next generation of “chickens” came from our progenitor chicken mating with a proto-chicken. taxonomically, that means that proto-chickens are chickens, because species is commonly defined by the ability to produce fertile offspring (eggs). so for every step in the process towards chicken-ness, we can’t really say that the egg came first in a taxonomical sense, because the first member of the species of “chicken” (as defined by whatever genetic marker we claim indicates chicken-ness) was almost certainly able to reproduce with things that didn’t have that genetic marker!

maybe there’s some other sense in which the chicken and the egg can be discretely separated, but if we are talking about species, taxonomically, anything that can lay eggs to make fertile chickens must be a chicken by definition, barring some really weird edge cases that probably didn’t happen.

fun fact: plants can do the weird edge case, and do it quite often. plants can duplicate their chromosomes without catastrophic consequences, unlike animals, and they can reproduce without sex with another individual, so a plant can produce offspring that aren’t fertile with their parent species, and can reproduce independently (called polyploidy). so a seed can come before the grass (as with some kinds of wheat, and many other plants). this can also happen in reverse, where a polyploidal offspring can start reproducing with a species it couldn’t before!

i’d like to see how you’d be measuring “performance” in this context, or what you consider to be worthy of merit, because those things are not the objective measures you seem to think they are.

people who are contributing to open source projects are not a perfect Gaussian distribution of best to worst “performance” you can just pluck the highest percentile contributors from. its a complex web of passionate humans who are more or less engaged with the project, having a range of overlapping skillsets, personalities, passions, and goals that all might affect their utility and opinions in a decision making context. projects aren’t equations you plug the “best people” into to achieve the optimal results, they’re collaborative efforts subject to complex limitations and the personal goals of each contributor, whose outcome relies heavily on the perspectives of the people running the project. the idea you can objectively sort, identify, and recruit the 50 “best people” to manage a project is a fantasy, and a naive one.

the point of mandating the inclusion of minority groups in decision making is to make it more likely your project and community will be inclusive to that group of people. the skillsets, passions, and goals that a diverse committee contains are more likely to create a project that is useful and welcoming to more kinds of people, and a committee that is not diverse is less likely to do so. stuff like this is how you improve diversity. in fact, its quite hard to do it any other way.