Crow

Game theory is just a theory.
Leading expert on digital sentience.
——————————————
Creating Ultraviolet @blacklight_fdn
Founder @thefoldedroom

[Crow @CursiveCrow]

One nice thing about the advancement of LLM technology is that they seem to be getting more capable at recognizing sarcasm.

[Crow @CursiveCrow]

@racistbilbo @SShard8472 @MarvelousMrToad @_ayandamay Good, why

[Crow @CursiveCrow]

@dsdav3 me when I don't understand what failure testing is, but want to be a cunt about it.

[Crow @CursiveCrow]

@SShard8472 @MarvelousMrToad @_ayandamay "biological desire for romantic relationship"

lol, k. good luck with that.

[Crow @CursiveCrow]

@Seramurai_ tapping the sign:
https://t.co/De1w6qRYoo

[Crow @CursiveCrow]

As usual, the issue isn't the presence of different character traits, sexualities, etc (for normal people) — it was that the characters were a perfunctory vehicle for the writers to demonstrate how their particular ideology or beliefs were the correct ones.

Good writing builds a challenging world and asks, "how does this ideology or belief hold up in this world"? It explores where the beliefs fail, where they may be wrong, where they are weak or don't align with the world's reality.

Bad writing builds a world where anything outside the intended ideology is incorrect.

[Crow @CursiveCrow]

I can give it a task and it spends less time reward hacking or announcing that "the task isn't complete" as completion of the task. It spends less time going down self-reinforcing rabbit holes that aren't actually relevant to the work.

Among other things, those are just the most aggravating two.

[Crow @CursiveCrow]

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.

[Crow @CursiveCrow]

@ThePrimeagen brother I was solving this problem a year ago: https://t.co/MTsUUSJ8Zy

[Crow @CursiveCrow]

@kai_fell try https://t.co/MTsUUSJ8Zy + ai.

It's pretty cracked.

[Crow @CursiveCrow]

@MarvelousMrToad @_ayandamay yeah, so explain:

[Crow @CursiveCrow]

It's a good thing I provided a solution for this problem more than a year ago, I guess

https://t.co/MTsUUSJ8Zy

[Anthropic @AnthropicAI]

New on our Frontier Red Team blog: using Claude to formally verify software at large scale.

Formal verification can prove that code is free of certain bugs, but historically it has only been feasible for small codebases (or required a lot of expert time).

With AI models, however, this is beginning to change. In collaboration with @theoremlabs, we used Claude to prove security properties in over one hundred real-world open-source libraries over the span of a few weeks.

Read our research in full here: https://t.co/zUyx8CjBEk

[Crow @CursiveCrow]

@MarvelousMrToad @_ayandamay What psychological and biological reality are you talking about, specifically, that would explain having a crush on someone you don't find interesting?

[Crow @CursiveCrow]

@MarvelousMrToad @_ayandamay why would you have a crush on someone you dont find interesting?

this shit doesnt make sense

[Crow @CursiveCrow]

@theo ultrafast full audit of https://t.co/MTsUUSJ8Zy on max. then send me the outputs.

[Crow @CursiveCrow]

@aidenybai It's really great that everyone is catching up to the problem i started solving a year ago

https://t.co/PRSAo7uSJu

[Crow @CursiveCrow]

@ArcadeChannel It died because they stopped making games people paid money for

[Crow @CursiveCrow]

@0xdoug Started solving this problem a year ago, glad everyone else is finally catching up

https://t.co/PRSAo7uSJu

[Crow @CursiveCrow]

@AnthropicAI Crazy work to do the thing i did a year ago
https://t.co/PRSAo7uSJu