Sign in

Ken

@whatabouttheken.bsky.social
948 followers 1.1K following 2K posts

Mostly a politics/policy reply guy. May post about various nerdy/techy interests from time to time.

PostsRepliesMedia
Ken @whatabouttheken.bsky.social · 3h
Someone making half a mil is almost certainly doing a Mega Backdoor Roth
120 0
Ken @whatabouttheken.bsky.social · 13h
I feel like Jeffries has some kind of strategy, though I have a lot of disagreements with it. Schumer appears to be sticking his head in the hand
010 0
Ken @whatabouttheken.bsky.social · 15h
Oh I've got an idea
DJ Crazy Times saying "World peace!" in the Planet of the Bass music video
050 0
Ken @whatabouttheken.bsky.social · 17h
That's like, lose your house and have your wages garnished for decades money
000 0
Ken @whatabouttheken.bsky.social · 17h
I feel like this discussion is really underplaying what $500k of legitimate collections can do to someone's life
210 0
Ken @whatabouttheken.bsky.social · 17h
Bad climate policies affect more than just the constituents
050 0
Ken @whatabouttheken.bsky.social · 17h
Is there a difference between $500k being written off because it was a mistake, and it not being written off/going to collections/probably causing bankruptcy? These are the same outcome?
110 0
Ken @whatabouttheken.bsky.social · 16/10/2026
Centrist Dem exodus to be Republicans? "New Republicans" branding maybe?
110 0
Ken @whatabouttheken.bsky.social · 16/10/2026
It's easy when the writers hand you a magic time-travel escape hatch
000 0
Ken @whatabouttheken.bsky.social · 16/10/2026
I think "making sure your users aren't encouraged to act psychotically" is a reasonable benefit. But yeah the model doesn't care imo
130 0
Ken @whatabouttheken.bsky.social · 16/10/2026
Right, and they're trained to react certain ways when presented with slurs or abuse. This doesn't mean one reaction is pleasurable and the other is painful necessarily. They're just following their weights.
050 0
Ken @whatabouttheken.bsky.social · 15/10/2026
Agree but I do think a more aggressive approach from DoJ could have done some good. Not sure if Bernie was the guy for that though
010 0
Ken @whatabouttheken.bsky.social · 15/10/2026
Exactly. Re-litigating a primary is a pointless exercise. The voters voted for who they voted for. The important questions are how the next campaigns should be run and how the next administration should govern.
010 0
Ken @whatabouttheken.bsky.social · 15/10/2026
Yes but anarchists have a response to all this:
"but it might work for us" Arrested Development meme
0100 0
Ken @whatabouttheken.bsky.social · 15/10/2026
The millennial's creed
060 0
Reposted by Ken
EAll @ealluia.bsky.social · 14/10/2026
Consciousness as we know it is closely associated with cognitive mechanisms that direct attention, which isn't something LLM's really do. I think this gives plenty of runway to be skeptical they are conscious, but you can imagine a near-future where that's no longer true, and then it gets tougher.
042 0
Ken @whatabouttheken.bsky.social · 14/10/2026
This exactly. If you can have "consciousness" without reasoning, then it's also feasible to have reasoning without consciousness. A lot of the information processing in our brains is done unconsciously after all!
000 0
Ken @whatabouttheken.bsky.social · 14/10/2026
It mentions anti-corruption investigations. Those are necessary but no where near sufficient.
000 0
Ken @whatabouttheken.bsky.social · 14/10/2026
And llms at temp 0 don't perform very well! Especially thinking models, a bit of randomness helps them explore more of the solution space and get out of local minima (and if they make a mistake most models nowadays can autonomously fix it and try again)
020 0
Ken @whatabouttheken.bsky.social · 14/10/2026
Not sure why you think I don't know that
210 0
Ken @whatabouttheken.bsky.social · 14/10/2026
This is in contrast to, for example, a mammalian brain, which exhibits huge amounts of feedback and inter-connectedness, especially in the areas associated with neurological consciousness. If an AI architecture were to come along that was more like that, it would be harder to dismiss.
110 0
Ken @whatabouttheken.bsky.social · 14/10/2026
It's pretty easy to make a structural/functionalist argument that LLMs, as they currently exist, are very unlikely to maintain a continuous conscious state. They basically wake up, say one word, then die, hundreds of times a second. The flow of information through them has no feedback paths.
110 0
Ken @whatabouttheken.bsky.social · 14/10/2026
Or to put it another way: if rocks can be conscious, why is it an interesting question whether LLMs can be?
120 0
Ken @whatabouttheken.bsky.social · 14/10/2026
The problem is when that claim requires you to admit that rocks might be conscious, what good is it? How does this advance our understanding of the world or what moral rules to apply?
Alex O'Hara @pghpubdef.bsky.social

Can you prove a stone isn't conscious?

Will Stancil @whstancil.bsky.social

No of course not
670 0
Ken @whatabouttheken.bsky.social · 14/10/2026
Neuroscience at least can say "when you break one of these parts of the brain, it seems to stop being conscious" and then look at those structures more closely
060 0
Ken @whatabouttheken.bsky.social · 14/10/2026
If we're going to be philosophers about it, maybe frame the question in a slightly more precise way than "are AIs conscious"?
030 0
Ken @whatabouttheken.bsky.social · 13/10/2026
Sounds like sunk-cost fallacy to me /s
010 0
Ken @whatabouttheken.bsky.social · 13/10/2026
In this case specifically they're pointing at small open models as the risk (because they are a risk... To anthropic's bottom line)
080 0
Ken @whatabouttheken.bsky.social · 13/10/2026
Sounds more like debate club than science. The notion that a bunch of matrix transformations has spontaneously become conscious is an extraordinary claim requiring extraordinary evidence. There is absolutely a skewed burden of truth against anyone making it.
100 0
Ken @whatabouttheken.bsky.social · 13/10/2026
The huge datacenter use is in training the models. Actually chatting with them is a tiny impact, akin to watching a few videos on YouTube. (So I guess the ethics shift to how much you believe using the model is encouraging/supporting the training process)
110 0
Ken @whatabouttheken.bsky.social · 13/10/2026
Not that it's really the point, but you can add -ai to Google search to suppress the ai summary
100 0
Ken @whatabouttheken.bsky.social · 13/10/2026
Treating one side of a question as "default" and requiring evidence to disprove it is literally a foundation of the scientific method (search "null hypothesis"). Burden of proof is not equally shared between yes and no on every question.
110 0
Ken @whatabouttheken.bsky.social · 13/10/2026
Not all affirmative claims are equal a priori. "A bunch of complex linear algebra used to generate words is not conscious" is the reasonable null-hypothesis. The burden of proof is on the claim that it is.
000 0
Ken @whatabouttheken.bsky.social · 13/10/2026
🤨
000 0
Ken @whatabouttheken.bsky.social · 13/10/2026
They want to be published academics sooo badly but don't want to do the work to get there so they just write pretentious blog posts
170 0
Ken @whatabouttheken.bsky.social · 13/10/2026
A gpu that can hold opus 5.5? In this economy?
120 0
Ken @whatabouttheken.bsky.social · 13/10/2026
BALL IS ROUND
0660 0
Ken @whatabouttheken.bsky.social · 13/10/2026
I'm sorry but running Gemma4 on a cluster of AndroidTV boxes is really funny to me
000 0
Ken @whatabouttheken.bsky.social · 13/10/2026
And "please government ban cheap open models so they stop taking our business"
000 0
Ken @whatabouttheken.bsky.social · 12/10/2026
It sounds like they tightened things up a bit vs the first year. Still insane to me that this is all connected to AI doomerism
010 0
Ken @whatabouttheken.bsky.social · 12/10/2026
"I don't wanna watch a show" ok then don't
000 0
Ken @whatabouttheken.bsky.social · 12/10/2026
"Cure for cancer that passes Lean" is some kind of mathematics fanfiction
000 0
Ken @whatabouttheken.bsky.social · 12/10/2026
On the other side of the bay: overheard some Berkeley students talking about their polycule on Alameda beach
030 0
Ken @whatabouttheken.bsky.social · 12/10/2026
Thought this was a calculus book for a second
110 0
Ken @whatabouttheken.bsky.social · 11/10/2026
Heck yeah
Emoji kitchen version of "hairy heart" emoji
040 0
Ken @whatabouttheken.bsky.social · 11/10/2026
I can't believe I was foolish enough to stay on Bluesky while the rest of the world was getting rich :(
070 0
Ken @whatabouttheken.bsky.social · 10/10/2026
I just did some driving across Spain and the highest I saw was 120 on very isolated divided roads. It slows down frequently when there are intersections and slows *way* down when you are going through any city/Town area. I got dinged by the speed cameras a couple times at these transition zones.
100 0
Ken @whatabouttheken.bsky.social · 10/10/2026
The road in question is not a freeway
010 0
Ken @whatabouttheken.bsky.social · 10/10/2026
"very few going over 80" on a non-divided road with intersections, bikes and pedestrians is a deranged thing to say
020 0
Ken @whatabouttheken.bsky.social · 10/10/2026
Original article is paywalled but that didn't look like a freeway to me
100 0