DisheveldYoshi
disheveledyoshi.bsky.social
DisheveldYoshi
@disheveledyoshi.bsky.social
Trapper turned poaster
Ran off on the plug twice
My strong belief is there will be what appears to be a very competitive GOP primary in 2028, and then everyone will get washed by whichever candidate(s) have the last name Trump
December 11, 2025 at 9:12 PM
Lying to congress is a federal crime, regardless of whether you’re under oath
December 11, 2025 at 6:00 PM
Reposted by DisheveldYoshi
There’s a sentence on page 2 that is longer than Bluesky’s character limit.

And the sole knowledge that sentence conveys is “Olivia was born on January 6th.”
December 8, 2025 at 2:15 PM
I feel like it’s worth emphasizing the framing takes her being attractive as the “action” on her part that shifts this from a) something being done to her, to b) something she’s seeking

It’s on its face abuse justification
December 7, 2025 at 6:11 PM
PredictIt got basically shuttered under Biden ~2022 by the CFTC, but the courts reversed it beginning under Biden and finally under Trump

I don’t really get it, but this was surprisingly not Trump but the courts
December 3, 2025 at 11:11 PM
They were originally ran by universities, as nonprofit information markets for education / research

PredictIt is online and first to do just gambling (i.e., “How many times will Trump tweet?”), vs political questions that should be efficient markets because everyone has equal information
December 3, 2025 at 11:06 PM
Possibly they’re casting Hagseth to take all the blame, hoping he’ll play along to provide cover for them and then get a pardon?

If so, he seems to not have gotten the memo given he’s throwing Bradley under the bus

Surely the scheming viziers aren’t just making it up as they go…
December 2, 2025 at 5:34 AM
Why are ct senators like this
December 2, 2025 at 5:28 AM
Full belly laugh at this. God they’re so fucking stupid
December 2, 2025 at 2:39 AM
Again, you’re not actually addressing anything I’m saying and stumbling your way backwards into science denialism. Moreover, you’re a deeply incurious person and very boring as a result. Tata
December 1, 2025 at 5:22 PM
You are the only one saying algorithms are magic, either because you refuse or are incapable of engaging with the substance of the critique in front of you

If this approach makes algorithm magic, then everything from your brain to your atoms are also magic
December 1, 2025 at 5:04 PM
For example, such a worldview would lead to statements we previously assessed as true to become untrue as a result of discovering more true facts, none of which contradict the original claim

That’s obviously a bad framework for assessing truth

bsky.app/profile/dish...
What if we suddenly eliminate all “hidden components” of the brain? It seems like you’re saying all the information we consider provably true today would suddenly be not true because we’ve gained more information (so the standard changes). That’s a weird approach to truth!

bsky.app/profile/mani...
Unlike in your field, there are no hidden components of an algorithm.
December 1, 2025 at 5:03 PM
I mean, whether we consider CS its own science onto itself or not is entirely orthogonal to “how do we assess truth for CS questions”
December 1, 2025 at 5:01 PM
Yes, I know. My response is that you need to provide an argument for why we should throw away how we typically asses truth in this one case, and only this one particular case

And my second point is that such an epistemological framework inherently contradicts how we expect truth to work
December 1, 2025 at 5:01 PM
What if we suddenly eliminate all “hidden components” of the brain? It seems like you’re saying all the information we consider provably true today would suddenly be not true because we’ve gained more information (so the standard changes). That’s a weird approach to truth!

bsky.app/profile/mani...
Unlike in your field, there are no hidden components of an algorithm.
December 1, 2025 at 12:25 PM
Ultimately, this whole conversation seems to be an epistemological exercise where you’ve decided perfect knowledge about a system (or at least what you think constitutes that) changes *how* we know something to be true about that system, and that is a rather bold and non-obvious claim
December 1, 2025 at 12:06 PM
Why are you imposing a separate set of standards onto CS vs every other science?

The scientific method and our approach to knowledge are in service of best understanding the world. The study that began this conversation takes those approaches and applies them to LLMs in a medical context
December 1, 2025 at 12:02 PM
Yes, but he was elected president and the Supreme Court ruled the president couldn’t be tried or jailed for anything

The only thing that keeps Trump out of jail is the fact he had majority support, i.e., he won an election
December 1, 2025 at 5:20 AM
As soon as capital and elites realize Trump is going to jail, they’ll give up because they’re cowards
December 1, 2025 at 1:11 AM