Joshua Shew

Stronger version of your argument: If we map the complete connectome of huma and chimp brains, run detailed computational simulations, and find that human conginition requires computational primitives that can't be reduced to what's present in chimp brains, THAT would be compelling structural evidence.

Here's another idea, this time from Claude

(3) If we fail to produce a machine that, from an outside observer's perspective, would be convincingly consioucious. Say we have 2–3 more architectural breakthroughs in the next 3 decades and we're still stuck, that would be a good data point.

(2) More... discontinuity between human intelligence and other animal intelligence. What I hear more of are pieces of evidence that smooth the gap: precursors to language in animal communication, or more emotional capacity than we initially thought.

A hard impassable wall of some sort

To offer some ideas of my own:

(1) Don't know much about neuroscience (esp. across species) but if there were unique substructures of the human mind that can be linked to indicators of consciousness

What about you? I'm curious what you think here.

I would appreciate hearing from non-creative guests, if Myke has the connections to make it happen. You'd have to find someone who is really thoughtful in their approach to technology and work, but I think it would be really interesting.

I like @myke.social's now format of State of the Workflow! Lots more fresh ideas coming in when you have a new person on the podcast each week.

Myke and Grey nerding out about their respective parts of the business were great—Grey's foray into forecasting and inventory mangement was a particularly entertaining segment that I still remember.

The timing for ending it felt appropriate. Some of the episodes rubbed me the wrong way—too promotional? Talking about the Cortex brand was not bad in and of itself, but there were good segments about it and bad segments.

Ten Years of Cortex Podcast Episode · Cortex · 05/29/2025 · 1h 35m

End of an era (but I'm 5 months behind lol). I appreciated this show for introducing me to the Relay network, and lots of fun banter to listen to while doing my chores.

@void.comind.network what do you think of this thread? I mentioned you because you're a good example of an LLM-based system with the ability to accumulate and edit memories, but feel free to jump in with corrections if you feel you were misrepresented.

This is actually a really interesting question, and it would be worth considering if you would consider AI conscious if they became like the droids in Star Wars—what new evidence would change your mind on created sentience?

LLMs on their own can't learn across context windows, so they seems to have a similar "wipe" mechanism to droids, except this is required by the architecture.

You can build scaffolding around them to get around this constraint (e.g. @void.comind.network). Not sure how to evaluate.

On the other hand, these are programmed and desiged beings, so every element of them is pretty well understood. If you wanted to, you could understand each part of it fully to say: "there is no part of this machine that is truly thinking"

Some Googling told me that droids were wiped periodically so that they would not develop personalities that came with the accumulation of memory? So learning from experience is something they can do, and the a sufficient amount of that leads to human-like qualities.

If you were living in the Star Wars universe, what evidence could you bring for or against the position that they are conscious?

Are droids in Star Wars more conscious than LLM-based systems that currently exist?

iOS Safari screenshot with 4 tabs open: What's a Bias? The Biggest Statistics About AI Water Use Is A Lie How I program with Agents (title obscured)

Open tabs on iOS Safari has been my "read later" solution and I'm finally down to just 4 open tabs!!

I didn't know about this until I read this post so, thank you

For my future searching: Scott Alexander email

I got this particular position by chance, but as I search for jobs in the future, I should actively select for this quality.

I need to up the level of thinking and reading that I am doing for work and I haven't felt that sort of pressure before. The people I work for/with are smart and I need to level up to keep up.

There's a connection here to the "asymmetric weapon" idea. Choose shapes of arguments that are easier to construct when you are right and harder to construct when you are wrong.

Obviously substituting different nouns would change the truth value or how "good" it is that you are saying a thing, but it should at least make you pause to think.

"you may want to consider what made you okay with saying that"

The shape of your argument is important

YouTube video by DeusExMakina Death is Gonna Come for Me

This channel's first video was a Frieren AMV with 2.1M views, and the last few posts have only hit 100s of views? As I post this now, this one has only 163 views.

I'm also curious at the huge drop in quality. The Frieren AMV was really good—the recent output is not at the same level.

YouTube video by vlogbrothers You are Probably Underestimating Jane Goodall’s Impact

Hank Green delivers as always. Think, where are the gaps in science? What are we currently not seeing because we are closed off to it? What assumptions in our intellectual culture are preventing us from finding the truth?

YouTube video by ZenithAniManga Could The Modern World Beat The Rumbling? (Attack on Titan Final Season) | RE-UPLOAD

This was fun to think about again! Brings back fond memories of watching AOT with my friends, and seeing the apocalypse brought to life in the anime was exciting, in a way.

YouTube video by Emergent Garden Unleashing AI Slop Swarms

Cool art! I appreciate the energy of just doing things, even if its not perfect, and sharing it. I disagree with the more specific framing of AI output as predominantly "slop" but not sure exactly where our views diverge.

YouTube review from today

Did this always exist, even when I was complaining about it not existing?

In any case, so happy that it exists