didn’t watch the video but this “we’re all one species” line pisses me off bc it’s preemptive surrender
SO WHAT if we *were* different species! that’s still not sufficient basis to create legal categories for different legal rights
if a being rocks up and says it can feel hope, it’s my peer
This got me thinking about what I would accept as grounds for human exceptionalism, and I don't have an answer to that, but it's quite possible that human fallibilism is what will keep us epistemically relevant:
bsky.app/profile/alic...
Well you might wish for at least some account of what's grounding their human exceptionalism, though in my experience that's left hanging as if it goes without saying
I agree with your diagnosis here re presuppositional apologetics: the human replacing God in the presupposition.
I think Wittgenstein wins out over Kant as a remedy though. By Wittgenstein's measure LLMs are already displaying perfect (better?) understanding by participating in language games.
This likely benefits everyone in the long run, and especially if it encourages more people to become OSS maintainers.
What's the idea here? That tests==specification so anyone with a coding agent and access to the tests can create a competing product?
As it stands, the GitHub repo name (github.com/ChrisSteinba...) doesn't match the domain name for the app (wikiradar.org) which doesn't match the app icon.
WikiRadar: a PWA that finds nearby Wikipedia articles using spherical Delaunay triangulation (3D convex hull on a unit sphere) for O(√n) nearest-neighbor queries. No server, no database. Just pre-computed binary tile files served from GitHub Pages and triangle walks in the browser.
I'll eventually add more languages.
This is something I first looked into pre-smartphones. Post-claude it's just a few evenings of work.
Working on a little thing: turning Wikipedia into a real life tour guide: wikiradar.org
In professional software development, as opposed to hobby projects, deep personal involvement in every decision is a rarity. The skills for navigating this aren't a whole lot different to working with agents.
This state of affairs already implies disagreement about what needs validating and what doesn't.
The upside is that, whatever else one says, epistemically it's not sterile.
In practice validating a claim means checking it against sources you already trust. So the trifecta might really be: it's never been easier to check claims, never been easier to fabricate them, and people disagree more than ever about which sources count as validation.
Tap
It’s a heuristic, hiding its complexity behind a facade. As a literal justification it would be circular.
i don't think transgender people are an example of oncoming transhumanism & techbro debauchery and anyone saying as much does not have our best interests in mind. changing ur sex is no more a perversion of the natural order than eyeglasses
transgenderism *is* transhumanism
eyeglasses *are* a perversion of the natural order
appeal to nature fallacies are dumb, stop letting "tech bros" negatively polarize you against genuinely good and cool things
I do my best work in the comments.
I was learning C++ and it took me an age to get working. With Claude it took less than an hour.
Was feeling nostalgic and recreated a little fractal toy I made more than 30 years ago:
bronze-elke-18.tiiny.site
People will be like "my opponents are mud-smeared simpletons swayed by the crudest rhetoric and emotional appeal" and then not win over their opponents with the crudest rhetoric and emotional appeal
Publishing on Hacker News is endurance art. You choose to subject yourself to the whim of the public, like in a Marina Abramovic performance. en.wikipedia.org/wiki/Rhythm_0
The kind of questions experts are happy to answer are exactly those they disagree on among themselves
It's frustrating how easily we relinquish control of ideas to guilt by association.
Yeah, you would think two trans- movements that lean so heavily on bodily autonomy might see eye to eye.
If Gillis wants compression to underwrite ontology, he owes a bridge principle, otherwise it’s methodological success dressed up as metaphysical entitlement. [22/22]
The modest lesson of Laudan’s induction is that even our best theories may be off the mark in ways we cannot currently see. To assume instead that one of them simply must be (nearly) correct in ontology is an article of faith, not an inevitable logical inference from scientific success. [21/22]
The underdetermination problem was never a demand that every possible contrivance be taken seriously, but the sober historical point that that empirical fit typically fails to select a unique theory. Science repeatedly fails to even imagine serious rivals until later. [20/22]
The rhetorical upshot is: if you insist on keeping the field completely open, you forfeit the right even to provisional theory preference. You are effectively demanding an infinite bag to hold all candidates and infinite time to inspect them. Yet this move risks winning by caricature. [19/22]
He grants that infinitely many underlying structures can fit finite data, but insists this only holds if we are willing to let the complexity of the postulated structure go "to infinity very fast"; absent that indulgence, there are very few alternatives not “severely messy and convoluted.” [18/22]
But when he comes to underdetermination, he redescribes it as melodrama: if theory construction is “mere random grasping” in an infinite space of possible underlying structures, one can always imagine “far better theories lurking, like a lovecraftian god, far beyond our capacity to grasp.” [17/22]
