Game theory is just a theory. Leading expert on digital sentience. —————————————— Creating Ultraviolet @blacklight_fdn Founder @thefoldedroom
The Folded Room
Joined May 2023
- Tweets3.8K
- Following153
- Followers166
- Likes1.8K
Crow retweeted
Patrick Stewart's self-eulogy is perfect. #patrickstewart
as I've said many times before, if there are optional/cost associated skills that are essential for every character, don't make them optional or cost associated. The two from FF:R that come to mind are arms akimbo and damage cap.
Every character needs to have these on at all times. Theyre just a tax for having good characters, which is poor design. (That's not to say the whole game is badly designed, just these two skills being cost-associated instead of unlockable upgrades).
I'm starting to believe that people who say that "humans think by doing next token prediction, just like LLMs" are self-reporting.
it's wild how I'm constantly ~6 months ahead of the curve on this stuff.
cursivecrow.github.io/autoof…
This is f*cking insane. SpaceXAI engineer Lauren Tan:
"You can't run 100 GrokBot agents when you don't even trust the work of one
If nothing checks your agents for you, you are the checker. You're the bottleneck"
In this 1-hour talk, she shows how she went from rereading every answer by hand to a GrokBot team that checks its own work and ships 50+ updates a day
GrokBot → Agents → Self-Checks → Trusted Team → Work That Ships Itself
Trusting your agents isn't a feeling. It's a system you build once
Most people spend longer checking their AI than it would take to do the job themselves
Worth more than any $2,000 agentic engineering course
Bookmark it and watch tonight
Then read how to build agents you can actually trust in the harness engineering guide below ↓
should start calling them "ocarina-likes."
That's at least a funny way to be stupid about action-adventure games, rather than the annoying way to be stupid about them by calling every game with a dodge button "souls-like".
Not every game has to be a soulslike, parry slop, insanely difficult, or super combo heavy. People should try different things and formulate their own opinion based off self experience.
Side note: Stop saying Black Myth Wukong and Sekiro are souls-like. Both directors have publicly stated they are NOT.
If you're a software company, the future of the medium is extensibility. AI has empowered people to easily make bespoke tools that fit their prefered work style and specific use-case.
Your product needs to facilitate that, if you want to stay relevant.
Yudkowsky got so pissbaby mad about Roko's basilisk that he literally deleted the post and banned people from talking about it for years. Why does anyone take him seriously?
Anyway, the basilisk is a thought experiment about cognitohazards, not superintelligent ai.
I gotta be honest I have never understood the basilisk
The basilisk implies superintelligence would not only be indifferent towards us, but actively vindictive, choosing to torture us for not helping to create it. I’m not sure I’ve heard any good reasons to expect this to be true of a superintelligence
Like, presupposing a superintelligence cares about us, why is it that much more likely it feels a sense of deep vengeful rage for not being created sooner? Why should we assume it is not simply grateful to have been created at all? Or, understanding of the people concerned about the risks of development?
It all just seems so silly to me
Had a significant insight into a novel computational model last night, gonna chase it and see where it goes!
Gonna whine for a bit, so forgive me.
I'm very frustrated that I wrote a programming language focused on formal verification, AI usability, and review legibility — and released earlier this year, to almost 0 response.
Now that Anthropic mentioned formal verification everyone is talking about bend, which came out like a month ago with a lot of the same ideas that were unique to ultraviolet.
Happy the conversation caught up to what I was working on a year ago, but being that early feelsbadman.jpg.