ian_j_butler

5 ms·

1.
by ian_j_butler 4mo ago
Right, I don't think researchers are confused on this point.. the anthropic piece is good outreach / science comms. OTOH this thread has like 200 comments and no mention of faithful/faithless reasoning. The idea that "
2.
by ian_j_butler 4mo ago
It's well-known that the reasoning model output is not necessarily faithful to the content of the thinking scratch pad anyway, even if you had it unsummarized and available verbatim. Setting aside coding agents.. we really need this
3.
by ian_j_butler 4mo ago
And that's the real tl;dr. Hybrids win whenever anyone actually checks. To really be scientific we still have to check, but.. why wouldn't they? Probabilistic AI brings intuition/learning but can't plan/search. C
4.
by ian_j_butler 4mo ago
Like US politics now, AI talk is just a team sport where there's no room for nuance. If you want to talk about such things then you need to anticipate this and work around it from the jump, making clear the expectations about the scop
5.
by ian_j_butler 4mo ago
> I think we’re all used to user-hostile software these days, but this is the first time I’ve experienced software that feels like it’s actively trying to be disrespectful. Always wild to hear people say stuff like this. First, all user
6.
by ian_j_butler 4mo ago
Challenging questions for short answer format! But there's a pattern with this stuff that's always mostly the same. Specialist logics work great on the happy path, but when you dig into it, they always work because of a bunch of
7.
by ian_j_butler 4mo ago
> resolved this well enough. Opinions? Not at all resolved. If anything it is worse than before as we begin to understand it better, and now there are different versions of it that cover representation, relevance, epistemics. Pivoting
8.
by ian_j_butler 4mo ago
Hah, since it's open in another tab: Talk Isn’t Always Cheap: Understanding Failure Modes in Multi-Agent Debate @ https://arxiv.org/html/2509.05396v2 To actually engage more with the substance of TFA.. very refres
9.
by ian_j_butler 4mo ago
A great rambling book that's related here and connected to neuroscience too is "The Age of Insight: The Quest to Understand the Unconscious in Art, Mind, and Brain, from Vienna 1900 to the Present". And since TFA seems to ha
10.
by ian_j_butler 4mo ago
> If you want to overcome this situation you need to replace capitalism globally. Not true.. you just need to replace late-stage capitalism locally. There's a very reasonable concept of circular economy that's relevant here.
11.
by ian_j_butler 4mo ago
Thanks for writing this. I'm going back to basics on a bunch of this stuff myself as the connections in the AI and ML literature keep heating up. Some stuff from my reading list lately that you might like: https://arxiv.or
12.
by ian_j_butler 4mo ago
> If we want to have a game where every weapon is a viable option, then we can’t bump RPS’s 3 to 4. We have to jump all the way to 5 choices with the game Rock – Paper – Scissors – Lizard – Spock. Each weapon beats two others and loses t
13.
by ian_j_butler 4mo ago
This is exactly like when Skippy is trying to work out the flavor profiles for the kinds of mush that monkeys will find most appealing. Yes, we know that "chocolate peanut butter bananas" is the true king of flavor latent space,
14.
by ian_j_butler 4mo ago
> something like Celery with all it's underlying infrastructure. Unpopular opinion, but combining this with the other "no thanks" sentiments in this subthread is the right answer. Your app is so complicated you need async
15.
by ian_j_butler 4mo ago
Why does someone go to the effort/expense of doing summaries without adding tags? The goal is less slop and not more, and this really just hurts usability after you've already spent money that could actually help? Isn't thi
16.
by ian_j_butler 4mo ago
> what leads people to this type of thinking once they get in a group setting? It just... seems endemic at this point. Large and fascinating topic I'm researching, very relevant for agentic AI and ML too. One way that groups can fa
17.
by ian_j_butler 4mo ago
This is actually pretty well-understood if people wanted to do it.. contrary to popular belief it's more about responsible governance than economics and not really a pro/anti capitalism thing https://en.wikipedia.org&#x
18.
by ian_j_butler 4mo ago
Personally, just from the implausible level of first-person detail throughout I thought it's immediately clear the book is exceptionally well-researched but still technically qualifies as historical-fiction. We don't necessarily
19.
by ian_j_butler 5mo ago
> That book is fiction with a factual veneer. Definitely, but do check the link.. I dug it up originally by trying to track down detail about the nonfiction background that the book is pulling from. Seems like the best short source, but
20.
by ian_j_butler 5mo ago
Happy to see that it's got the obligatory monk/wizard photo. For more life and times stuff I also suggest Labatut's Cease to Understand the World book and https://theanarchistlibrary.org/library/konstant
21.
by ian_j_butler 5mo ago
Again you're probably thinking of the forum discussion / booster blogs / executive interviews that I'm suggesting you should leave behind. Papers are the only place left where nuance is even allowed and might actually
22.
by ian_j_butler 5mo ago
The media is hellbent on torching it, and on propping it up against all reason too, both things can be true. HN is no exception. It's another noisy room problem where the distortion in dialogue is rapidly leading us into a distorte
23.
Agent-Based Modeling for Simulating U.S. Presidential Elections(arxiv.org)
2 points by ian_j_butler 5mo ago|0 comments
24.
by ian_j_butler 5mo ago
> But you struggling with rent is entirely your self-inflicted problem... The funny thing about this is that practically every single person in the US between the ages of 25-45 was an early adopter and a big fan for some (or all) of FA
25.
by ian_j_butler 5mo ago
> How do you limit the tracking? Seems easy in the way that a lot of things are easy, and this definitely isn't about making everyone and everything untraceable. We don't really need to play ads at 15% extra volume from conte
26.
by ian_j_butler 5mo ago
> you would have to make it non-interoperable with existing money. Why? Just eliminate surveillance.. no tracking is no money. There's another theory that maybe no money is no content, but that's sort of what tfa (and other s
27.
by ian_j_butler 5mo ago
He did actually work with permutations and cycles in voting ( https://en.wikipedia.org/wiki/Dodgson's_method ) but it's combinatorial and not very group'y. Agree with comments that this is AI generated
28.
by ian_j_butler 5mo ago
https://en.wikipedia.org/wiki/Haddocks'_Eyes
29.
by ian_j_butler 5mo ago
A hint for my detractors.. > Being afraid of losing one’s job is quite a step up from being afraid of a rival tribe ransacking your village. The rival tribe is MAGA, ironically destroying the commons partly because somone told them ano
30.
by ian_j_butler 5mo ago
Thanks for the hint. Reading more about this, it seems the crypto/zero-knowledge/interactive-proof terminology is formal but overloaded here.. games that have nothing to do with game-theory, simulation that has no connection with

More ›