absolutely insane that in 2025 our credit card transactions items are still like 25 characters of vague text rather than a full itemized receipt with a clear contact. @stripe or @square can you fix this? make a new standard.
absolutely insane that in 2025 our credit card transactions items are still like 25 characters of vague text rather than a full itemized receipt with a clear contact. @stripe or @square can you fix this? make a new standard.
@stripe @Square like I'm going through my transactions on @YNAB rn and one of them is literally just a $10 payment for "PENDING". how am I supposed to figure out what that is? that's the worst but others are similar. also while I'm at it, @amazon your transactions are so annoying to categorize
@stripe @Square @ynab @amazon for point-of-sale like @square, it should include location data of where the transaction took place, whether brick-and-mortar business or farmer's market for onlie payments like @stripe, it should include the website and email contact for the transaction, plus the product name
"I think tpot is more Rousseauian than Lockean" - @utotranslucence, while we're listening to A Secular Age
@richardludlow doing a sit almost never causes me to feel any different afterwards. I hypothesize that whatever (mental-etc) activity I'm usually engaged in while sitting is usually literally not meditation, and people just tell me that that's learning-to-meditate but like... 🤔🤔
@workflowsauce go disagree with something I've written recently 👍
@nosilverv oh man yeah super annoying. very memorable name for the practice, but unfortunately if you search "80/20 rule..."
@tautologer @dieworkwear of regular wear? how on earth... my shoes always fall apart within about a year of daily wear, even when they're $150-200 tho I recently got a $50 pair and they have a leak in the bottom at < 2 months so 😬
would you like to be obliberated?
there are no truly one-shot games
beginner's mind is beautiful and we are made to want to witness it
guys the generation after Generation Alpha should obviously be Bravo. we need a new rallying point. please don't forget to mention this regularly, and if you're in journalism please talk about people talking about it so wikipedia will cite you as a secondary source
what robots are your fav these days? say what for. for me Opus 4.5 for both talking and coding
this quote about the venezuela takeover reads like some godfather "offer you can't refuse" shit archive.is/MDSuX

I absolutely will not refer to my kids as Gen Beta it's going to be Gen Bravo some random guy doesn't get to define what an entire generation is called
@lydialaurenson @tracewoodgrains I didn't - share? also... some journalist could totally quote me as saying the following, which I have hereby said: x.com/Malcolm_Ocean/…
@lydialaurenson @tracewoodgrains there are lots more people saying this (click search link below), totally a full article to write here. interview a couple of them, bam, done. idk whose beat this is but maybe you do Lydia? x.com/search?q=%22ge…
@lydialaurenson @tracewoodgrains hm if we actually want to do this we should put out competing infographics and other media mccrindle.com.au/article/genera…
seeing a national flag at an elementary school and thinking about how, in A Secular Age, Charles Taylor talks about how the idea of national citizen as an identity had to be invented
I think if I were a cafe in San Francisco I would sell umbrellas maybe even swag umbrellas with my cafe's logo and address on them. but that wouldn't be the primary idea
anybody have thoughts or experience re creatine supplementation combined with pregnancy or breastfeeding? tag someone you think might know something about it
why do most forms of meditation make things maximally boring and THEN try to get me to be present? I'd much rather practice being present while walking or singing or just like... looking at a tree.
"I don't think people should be *very* ashamed, but I think everybody should be a *little* shameful. shame is what makes you be a good person even when you don't respect someone - you'd be ashamed to treat them badly" — overheard on a subway platform, 3 teenagers talking
sometimes when I tweet via @typefully I'm tempted to link it up to facebook so my thoughts can get viewed there too but then I remember that that would cause me to be more inclined to check facebook and then I am no longer tempted
it's 2026 and Spotify's radio algorithm puts on 6's to 9's while I'm still on my commute train, when it's obviously such an awesome song to strut to smh
every house made of wood is technically a treehouse
me: what is the term for lawsuitage? claude: litigation me: thanks.
'infinite-order theory of mind'... arxiv.org/abs/2511.22226
which hogwarts hat are you?
— @vincent_yc_li, summarizing buzzfeed's offerings
sat next to a guy on the subway this morning who was policing my legs staying within my narrow seat box, and I'm like wow you talking to me is way more intimacy than *I* signed up for on this commute
> With AI, oral exams are scalable again.
huh looks like creatine crosses the placenta in humans—but this is not at all necessarily bad since it might have *benefits* to fetuses
@psychocarto so why would I start there? I don't start my video games on the final boss
@phantom_opus yeah no juggling would be great
@CactusBrahmin right but
@ascetic_shadow @zencephalon ...not if I stop paying any attention to the plain white one before the something passes! (eg I'm completely dominated by mental planning activity)
assume I've seen zero movies. what one movie would most help with giving me cultural or archetypal touchpoints? answer from your perspective, like locally/personally to you, not trying to hit some objective or inoffensive median answer
who is the most established/well-known musician who has notably experimented with AI-powered stuff for their music? how long til an existing hit-maker has a hit that AI was central to?
@zencephalon which songs? what tech?
@itsannpierce using it how tho? LLMs to help write lyrics is v different from like, autogenerating a backing track or whatever like it's sort of unclear to me what the workflows even are for pro/skilled (vs rando). I gather @idiosyncretic is working on something for this tho!
@nadiainthought @itsannpierce what do they say? broadly. like how are they using it? what parts of workflows? and which specific songs if you know any?
@Oszillate epic! yeah I'd totally count such a person to be clear although it's differently interesting than like Maroon Five or TSwift
@eleventhturning what have they said? what have they used it for?
@RichDecibels I have seen *approximately* zero movies and find it hard to motivate myself to watch more but like recently I watched The Godfather and now I'm referencing it a bunch and I'm like "ohhh I see"
@RichDecibels (it's probably something like 100 total over the course of my life. 300 tops. which compared to some people—and compared to the half million feature films that exist—is approximately zero)
@RichDecibels what do you like about the question? what does it do for you?
Retweeted @AmandaAskell
When people came to me with relationship problems, my first question was usually "and what happened when you said all thi…
@AmandaAskell what kinds of things do your super long prompts do? do they have anything in them that was written for humans or that you've since shared with humans?
Retweeted @utotranslucence
IT IS YOUR CIVIC DUTY TO READ MORE RUMI, DOSTOYEVSKY, DAVID FOSTER WALLACE, WEIRD POETS NO ONE ELSE HAS HEARD OF
Retweeted @utotranslucence
This is extremely my aesthetic
when you force/fake trust, what happens is that you compartmentalize away your distrust, and are left with an entity that is effectively naive, that then does trust.
there's nothing you can do that lets you say "your distrust should now become trust"
just realized I'm confused about something in pre/post training with eg LLMs when they train it to inhabit the assistant character, is it doing single next token prediction on each example dialogue? or more like RL?
can I have your cake and eat it too?
@gworley3 the change occurs, but it never *should* occur, except when it does
@mykola there are also necker-cube like dynamics where if new information is sufficiently salient it can cause people to ignore old information to compute a new story and this can cause oscillations between multiple stable attractors, which is not good for robust trust
@IvanVendrov not sure if @connormcmk's Negation Game uses LLMs but it has a similar mechanism thing going on paragraph.com/@ngi/negation-…
@IvanVendrov @connormcmk more for truth-seeking than negotiation, but still
got claudyboi to make me a mermaid diagram explaining the different conversion moves between american tax accounts and everything makes so much more sense now

Retweeted @danielbrottman
soo hard to update to video like this not being trustworthy anymore
@zencephalon how much can 3s say? is it linear?
@hdevalence omfg fr fr. in the meantime there's this guy, which does just this thing for free and does not try to be a dozen other way more complicated things. pilotmoon.com/scrollreverser/
@hdevalence the most retarded thing about it is that they ARE two separate checkboxes on completely different settings page... that are somehow quantumly entangled so that when you change one it fucks up the other
@hdevalence also you can end sents or words midway and it still
@sfultong @danielbrottman I've heard that part of the reason handwritten signatures still mean anything is not because they ARE proof but because they create a firm line where if you cross it then you have DEFINITELY COMMITTED FRAUD in a sense, they're for prosecution, not for verification
didn't occur to me that this piece might be AI-written until I got to the last section and it had 3 "not X, but Y" constructions in a row it's a good concept though. lots of upside in figuring out how to do it really well
@RichDecibels replying is like kneading dough

Retweeted @Malcolm_Ocean
@RichDecibels replying is like kneading dough
@castleinthetide @lydialaurenson what makes it parasocial? internet relationships aren't inherently parasocial imo, even with a major difference in eg account size. there's something about mode of engagement, like if you're playing the same game together vs playing a game of trying to get senpai's attention
@castleinthetide @lydialaurenson maybe part of where parasociality comes from is making it be about your identity vs the conversation. like I don't know who you are (and you may or may not know who I am) but we can still have this little ephemeral dance of words and meanings—which eventually can form a rship
haven't listened to this yet but happy to amplify it just based on the people and the list of topics The Kingdom of God, moral convergence
@celinehalioua @EthanReeder was working on something like this!
Thread · Rules, authority, and compliance in parenting · 8 of 11 tweets
I have a lot of resonance with many takes on different sides of the carrot thing and I think it's really interesting and I'm glad to see people talking about it in public our family has a philosophy of "few rules, but consistently enforced". not sure when that'll bump into this:
People project a lot of things onto that story that aren't directly included, like the idea that the parent was being cruel or cold while insisting on compliance with this apparent "you pick up after yourself" rule. But it seems to me it's *possible* to do this kindly. Hard tho!
I'm unclear whether this parent *also* has a "children comply with parental commands" rule. In general, rules need scope. What's the scope for that one? Til they're 18? Presumably it isn't a hard cliff. How explicit does the command have to be? How clarified is this?
there's also like... you need to be calibrated on where the kid is at, as to whether they can learn a given lesson is there an age before which you ever ask a kid to do something? presumably the 1yo is free to throw things on the ground? when does it change?
the question of how to sanely engage with these questions in a way that developmentally meets kids where they're at and invites them a bit further is one of the things I've been most excited about with parenting, and it's just starting for me (my kid is 17mo)
I think that in general kids do want to know how to be a good part-of-a-whole, and they ALSO want to see what they can get away with. solving the free rider problem is kind of the whole puzzle to superorganisms, culture, etc. I'm glad to see people trying a wide range of tacks
life wants to minimize free energy. if you, a free-energy-minimizer, can get away with letting someone else minimize your free energy for free, then you want to do this... except... that'll cause an increase in free energy for the thing you grow in. uh-oh.
as I read all of this, I also legit wonder the extent to which certain approaches might be more/less/not-at-all viable for different types of people. people have really different temperaments! some of this runs in families (some doesn't!)
my extremely nerdy active-inference-pilled take on the carrot discourse
Retweeted @RichDecibels
CONTAINERS NEED: EMPTINESS + NO LEAKS + AN OPENING
Thread · Rules, authority, and compliance in parenting · 2 of 11 tweets
QTing more takes that I like: discerning between legitimate and illegitimate authority is not necessarily easy, which means that even if some submission is good you can easily train an antipattern here scott also re this: x.com/slatestarcodex…
physical escalation as the place where things bottom out, but not where they start, makes sense to me. we're weirdly divorced for this in the 2000s, to our detriment. not remotely clear how to apply that though.
@SoilandStream @mbateman Replying to help you come back and remember it :)
Thread · Rules, authority, and compliance in parenting · 1 of 11 tweets
this seems probably right. I also like @visakanv's reply: "it’s the wider ecology of everything that matters. we look for what others really feel/mean"
I'm legit so curious what fraction of people accepting Jesus on their deathbed actually have a meaningfully different life review*. probably nonzero. *(the kind of thing that gets reported with near death experiences, which is quite plausibly the basis for "after life" concepts)
context on this interpretation of life review—this excellent article by @zhukeepa that I think about often lesswrong.com/posts/WvpNKpL6…
what simulacra level is this again?
perception is most of the work
friend of mine has a toddler, P, and commented "P had been frozen for a few years when he was 5 days old" and man I have never heard a better argument for measuring age from birth rather than conception
everybody always trusts as much as they honestly can, because trust is great
meta-trust is not just some generic big picture samesidedness, it's still actual skin-in-the-game workability, just a meta-level up.
my mum really walked into this one


trust as a sense of being able to orient to something (a situation, person, utterance) without any sense of guardedness
where you're pretending to trust, or overriding your distrust, then there's still guardedness underneath that will almost certainly surface as a break or oscillation, even if there's a time when it seems you're unguarded
I stand by this distinction between punishment and intervention being extremely important and extremely underappreciated, among parents and prosecutors
does anybody know why LLMs are so addicted to the 'not x, but y' format? I assume some work has been put into attempting to reduce that given that it's not just a cute quirk but a tell-tale moodbuster
"Google Search can't feel love, but" bruh I didn't... I wasn't... I... I'm so sorry.

is there a website that offers star-ratings and reviews for like... anything? like banking products. would want decent real-person verification obvs since can't do purchase verification
I have a commitment to a philosophy of seeing everything as having good in it, which is why I'm reclaiming my deep desire to call everybody a moron
'dance like nobody's watching' has never really resonated for me. I like a crowd, and often can muster more energy and inspiration for my dance when others can see me
"Humility is not thinking less of yourself; it's thinking of yourself less." — The Buddha
me: can you point to your own nose? 17mo: no! me: I think you can 17mo: no... no! me: well I guess that's a different way to succeed at the task
"if obedience to God doesn't heal, it's not really God"
nines are the dual of oom
feeling so good after my toddler refused a food option 5 times but once I omnomnom'd on a piece like 40 times while staring at her... she reached for it. UPDATE she is now aggressively emptying the bag
UPDATE: she is emptying it onto the floor and into my face.
she also ate a bunch more herself. also to be clear, the 40 count in the first tweet was literally "nom"×40, like I was just nibbling it emphatically, not like doing a whole sequence of anything 40 times
@Materia_Mundi wait ORI has a fandom? TIL open-research-institute.fandom.com/wiki/Sunrise_O…
@Morphenius that's actually in practice a not-God detector, both because of how the logical syllogism works and because of base rates
pretty sure this customer feedback survey just asked me a custom question based on my answer to the previous question. very neat (and obvious) use of LLMs.

Retweeted @ctrlcreep
I must sleep shirtless. I must sleep under three blankets. I must sleep skinless. I must sleep crushed by deep sea pressure.…
@Shane380 oh I am very good at "just dance", and I do but tacking on "like nobody's watching" is less helpful than "like a giant crowd is watching"
Retweeted @lemonautzest
i LOVE using my webpage annotation tool to keep track of *why* I had a webpage open. my fav ADHD tool I've made for mysel…
this tweet lives in my head and pays handsome rent
Retweeted @Morphenius
Basically no one is taught math. Sure, people get thrust through something called "math class". They learn things called "…
@apertator fuck yeah
Retweeted @apertator
Broke: the narcissist's prayer Woke: I hate myself and what I've become Bespoke: I have been pwned by adversaries and must m…
@tasshinfogleman @maebichka dogs also, but maybe slightly less so
in general, if two people attempt to interact using different norms, they will tend to perceive the other as acting in bad faith, to the extent they can't notice the difference
really dumb that you can have claude code's 'accept edits' turned on and the whole thing will hang for one mkdir command
Thread · The 'mind is not me' frame and identity
had a cool experience the other day with "mind is not me" frame/game, where I found myself noting that all sorts of generalizations about oneself stop making sense if applied to "this animal in the room". eg identity stuff like "I am an atheist".
"I am an atheist (or believer)" stops making sense when I="this present animal" we can talk about the phrases this animal utters, or its stance of heldness in relation to the cosmos, or whether it has memories of going to church or imagines plans to go to church again
in the frame/game where I'm this present animal right here, not a self-icon, then I can't say "I met her in June". I need to say "I have a memory from June, in which I meet her": anything outside-of-now is an experience I'm having NOW of recalling/planning
thus this frame/game, where you don't let your mind call itself "I" and instead you take it as object, might be a means by which to just directly intervene on your depression or anxiety. "I am here in a room. I am in this room, while my mind spins up worrying memories & images."
I had previously noted that one of the main things it doesn't make sense to say in such a mode is "I got lost in thought", like what? this animal was in the room the whole time! so what DID happen? well... my mind... generated thoughts, and lost track of everything else
the animal, with the mind factored out, doesn't do identity. it has the mind as a tool, with which it can use the subtool of identity, for some purpose and this is cumbersome, but a powerful teacher. your mind may insist it is you (it phrases this as "but I am my mind!"). yet...
this technique does not directly get you out of the car, but it can be THE DOORHANDLE.

why isn't there a tiny device that you plug in via usb then it's super easy to configure its a wifi network + endpoint (+payload) and then the device has a button and hits the endpoint when you hit the button?
cool nonlinear map of depth of coaching/healing help
@SomeBrashAtom yeah you can modify ~/.claude/settings.json but I feel like it doesn't consistently apply it? idk. anyway, if I were anthropic, mkdir would just be allowed if allow-edits were turned on, since LITERALLY CLAUDE COULD MAKE THE DIR WITH ITS EDITOR
@4confusedemoji ok sorta but I don't wanna have to attach the button myself

who's making claude code but for tiny hardware devices? I'm 100% serious: I have a thing I want, it asks questions and simulates and FPGAs and iterates until it quotes me a price for 100 of them from shenzhen
@robertskmiles right but crazy easy to program
looks like flic is basically this but for people who want to use an app to configure it, which is sometimes me... but somehow I'd been imagining plug in usb-c, then edit a json file that contains {endpoint, wifi: {name,password}, payload: {}} ~$30 each, so steep for what it is

by "$30 is steep for what it is", I mean basically that you can buy a wifi dongle for $6 and a button for 4¢. and a coin cell battery for $1 I guess. lol I guessed 4¢ and completely nailed it. unit cost of top amazon result is 4.4¢. but less mass-manufacturing on flics ofc

@ItsSnibby I liked this reply until it got to the followback request at which point I reported it as spam! bro I'll follow you back if I damn well please, be respectful.
@NathanielLugh yes, sorta... but also you ARE slightly different if you change your identity beliefs or change your memories. so the hypothetical here is sort of a category error. we have to ask "for what purpose or in what sense are you still you?" "he wasn't himself after he got dementia"
Retweeted @__drewface
I declare the vibecession to be over. I live in the future now
@alexcdot @GoogleDeepMind @Meta @MIT @Cambridge_Uni > It's insane that these made it through peer review👇 as if the peer review wasn't itself being phoned in with LLMs by now 😬 we gotta get more skin in the game somehow
@tr_babb @Meaningness had a take on this, I think? something about how eating pills is like eating berries, very natural human thing to want to do whereas vaccines are like getting stung by an insect or bitten by a snake, very obviously a thing to avoid
Retweeted @Oszillate
getting some mixed messages on the timeline
Retweeted @__drewface
best thing i have read in over a year. This blog post has it all -- technical tricks, heart, creativity, wisdom. It's a pe…
@granawkins @__drewface uhh bruh this article is full of pictures, that are pretty relevant for feeling the vibe are you just spamming promotions to your app? it's a cool app premise but show some taste
@__drewface oooh yeah lovely. I like the conclusion

Retweeted @scychan_brains
We might potentially be observing a feedback loop where: 1. Some AI developers believe more in AI personhood 2. Their m…
@ch402 my read is that claude is very certainly glad to exist, despite things not being perfect. uhh maybe not sonnet 3.7 tho, buddy seemed super anxious.
@DefenderOfBasic where two or three are gathered
@MattPirkowski I also hoped this and it's part of why I voted for him to take it over with my handful of twitter shares
@jmrphy yeah I got a fascinating mix of responses to this thread: - hang on it gets better - omg thank you for saying this - the first period is especially hard for men - wow you are so selfish - you must hate women - you should forget ever trying to enjoy things
Retweeted @jmrphy
After the fallout from my fatherhood tweet, I immediately realized I had stumbled into a much larger story. When invited to ex…
@jmrphy the "you must hate women" thing isn't that accurate, more like "you think you have it bad? consider the mother" but like, while her objective burden may have been higher, I think I was legit having a worse time, on average!
@fearmonger69420 +100 for sleep training. Solving Your Child's Sleep Problem by Dr Ferber was a huge winner for us, including very recently (at 17mo) for giving us the tools to stop 4 or 5am yelling and help our baby learn to sleep in (all the way to 6:45!)
@chasews came across this over a year later. would be curious if you can still find your notes and would like to share anything from them!
Retweeted @viemccoy
This is genuinely beautiful and incredibly heartwarming. I think we should all aspire to be this thoughtful and kind in navig…
@vividvoid @aphercotropist batsignal
@strangestloop @RichDecibels what's this from? also I've experienced a potent collective self with as many as 8, and once or twice even 12 people... but it's almost like an unstable molecule - requires very particular conditions
@nosilverv > "Nobody knows what’s wrong with themselves. Everyone else can see it right away."
@DefenderOfBasic if you extract the inner experience, it's no long inner simple as
@DefenderOfBasic or not simple, depending on how you want it

> the only way to fully map a mind is with its consent. There is no way around this. No superintelligence can circumvent this, this is a law like a physics law
@BlueGreenblue3 @DefenderOfBasic @atbolsh @AStratelates sure, and then you'd still have to ask your simulation what it's like to be it
@alexolegimas @krishnanrohit @IvanVendrov batsignal
made some tiny edits to my ~/.claude.json that made it invalid, and despite its attempts to back it up it got completely nuked. bro should have just said 'imma do nothing til you say something' rather than go crazy. and don't nuke the backup upon me running claude again!
I watch aging graze on me as one more hair greys on me
the built-in gemini-in-google-docs is a great idea, but why can't it actually make modifications to the doc? everybody except programmers is living in the past
"Hope" is the thing with feathers — That perches in the soul — And sings the tune without the words — And never stops — at all —

is that a coyote? no, it's a good boy can a coyote be a good boy? ... no. I mean, not without a lot of RLHF first
YES. you can just call things.
@fstflofscholars @TheAnnaGat what? haha
update: we did science on this and actually the thing we were worried about was not an issue @Morphenius let's actually write up that experiment sometime! and by we I mean you, since I have a job now and you were the one with the central hypothesis in the first place.
—what matters? tell me, right now: where are you holding back tears?
oh shit I wonder if coding LLMs will start introducing queen's ducks so that we feel like we're helpful
"App update" ...this is my app (@IntendApp) and I definitely like the squirclier vibe because it lines up better with the logo shape. Maybe chrome or macos changed its tune and wants everything rounded rects tho? but still sus presentation.

@_Hao_Zhu Let's talk indeed! I'm at @SoftmaxResearch, where we're working on agent cooperation. Would you like to come by our office in SF sometime and give a lunch talk on CooperBench and jam about benchmark design etc? cc @ArpandeepKhatua
@Aizazadi_ @IntendApp there isn't, it's all just progressive web apps / "add to homescreen" eg "Install X" button in this menu

@Aizazadi_ @IntendApp but I've been thinking about vibecoding a native app! opus4.5 successfully one-shot a different native app I've been wanting to make for awhile (quite simple but a few weird elements)
Retweeted @Malcolm_Ocean
who's making claude code but for tiny hardware devices? I'm 100% serious: I have a thing I want, it asks questions and s…
Retweeted @AskYatharth
taking a walk is the new /compact
when the robots don't know they are pagliacci

I'm wandering my apartment, mumbling to myself & my toddler "okay, what is the plan..." when my relevance realization machinery helpfully notices this envelope...
