(I have signed no contracts or agreements whose existence I cannot mention, which I am mentioning here as a canary)
Drake Morrison
Karma: 922
The Wormtongue Test
I’m always a bit surprised talking through the basic arguments is not done more, especially in places where new people are joining the field. How do you even understand what problem the field is facing without that?
I sometimes reread Optimality is the tiger, and agents are its teeth, and wonder what people would think of it now.
She posted this essay back in 2024 on her old personal site
The legal liability seems like the real thing that forced OpenAI’s hand, but sure, it was probably helped a bit by the marginalist faction, so they get some merits for that. Then they get a lot of demerits for aiding the safetywashing / safety-licensing that maintaining good relations with the labs entails. What’s the relative balance of the merits? I don’t know, seems tricky to measure.
The difference between Schindler and Speer is razor thin, and difficult to tell. I think it right to maintain some suspicion on those claiming to play that strategy.
Seems pretty clearly like in the first case you have mostly similar priors, or world models, and in the second case you don’t.
I like Highly Advanced Epistemology 101 for Beginners as a shorter intro to ~ a quarter of the art
I had a great time playing with you then, and I would gladly have kept at it for hours and tried novel things as stuff got boring. But I was working and had to move on.
I think this obstacle comes up a lot when trying to make spontaneous playfulness happen in a way that keeps its momentum or grows, people have shit to do. When I was a kid I had all the time in the world, now… not so much. I’m reminded of Playing in the Creek
jeweler’s loupe
small tape-measurer
monocular
lifting straps for moving furniture
See the event site for more details
name three examples.
(You’ve said you didn’t want to get into this, so I’d consider it understandable if you don’t. Maybe that’s a later post you intend to write, or have already written? I just want to push back a bit on the implicit consensus of “community elites” without examples)
Immediate taxonomy of subskills that came to mind:
designing for consumers vs designing for producers
technical breadth vs technical depth
managing up vs managing down
Physical strength vs physical dexterity
Dang it, now I wanna read/write posts for all of these
If It’s Worth Arguing, It’s Worth Arguing With Whiteboards
If You’ve Never Bought a Tool You Didn’t Need, You’re Not Buying Enough Tools
Carpathia Day
Plans are Not Promises
You’re referring to the Vickrey-Clarke-Groves mechanism?
Meaningful Questions Have Return Types
Uninterrupted Writing as Metric
I ended up here in desperation after trying a couple of other ways to write up my thoughts as posts. I tried narrating, stream-of-consciousness style, into Claude hoping it would help me, but then it wrote a whole essay itself. It was too much Claude, and not enough me. I didn’t like it. I did find the structuring work that Claude did useful, so I asked it to do this instead. I found it much easier to respond to questions that are already structured than to extemporize right out of my head the heap of chaotic thoughts about a topic.
Or perhaps its a case of trapped priors. Ontology differences mean you don’t interpret the same piece of evidence the same way.
I think you miss how costly it is to do so, especially when your interlocutor is already dismissing your claims about yourself. To use your earlier exampole about buying cookies from your girlfriend: if you happened to not produce artifacts of your thinking when deciding where to buy cookies, the only thing you have is claims about your internal reasoning. Which is opaque to your interlocutor who just got through dismissing your claims about your internal reasoning!
That’s the highest-order hypothesis? Feels like taking an adversarial stance unnecessarily. It can easily be the case the conditions to continue the conversation are too costly. That’s not a threat.