THREADAPALOOZA what topic would you like to see me write 100 tweets about? (that I haven't already written 100 tweets about)
December 2025
welp, all this free energy ain't gonna minimize itself
"anything worth doing is worth doing secretly by a small group of elites" — @myhandle
amazon shipped a project's codename to prod?
@DrAndyGalpin @mtx62 if I don't need full-on performance stuff but just want to check my levels on various things to see if anything is out of range, do you still recommend Vitality or some other product?
@j_elliot_art it does not annul but the beauty can overcome any regret that would wish to somehow be on a different path
should be possible to make a site that's like Q&A but with AI but answers are viewable by other people and sold in bulk to AI companies as training data would be a huge gift to humanity
but good tho. specifically one person keeps updating their own entry until they have solved their problem. like stackoverflow. really stackexchange should become this - it's already got the "answer your own question" thing.
one big issue with AIs at the moment is that they don't get good feedback about whether your problem got solved. there's SO MUCH POTENTIAL here for learning you could probably give the inference away and sell the data gained from the feedback loops
cellist in the subway station starts playing pachebel's canon and I'm like "oh crap the train's arriving" because the first two notes are exactly what the Muni signal sounds like
update: newborns and older babies are completely different species and it is silly to make generalizations about both aside from very basic stuff like "they are small" and "they like boobs" and "they can't really talk"
what kinds of ages do kids become developmentally ready for the "put yourself in the other person's shoes" move? it's kinda like theory of mind, so maybe 4? but I feel like some aspect could be younger if the adult is patient
reading about active inference and noticing that making your preferences simply BE your expectations is super psychocybernetic
in these times, with people posting about the soul doc, I will remind everybody that 'soul' is evocative shorthand for 'relevance realization function'
hey, just DMing you to let you know that I sent you an email but it's not urgent or anything so you probably don't need to get to it anytime soon. I'll let you know if I follow up
I'm almost here
if I weren't a tech person I'd be so mad at those checkboxes that just say 'I'm not a robot'. like why wouldn't a robot just click the checkbox?? (answer, they detect something subtle about HOW you click it, and other background info)
opus 4.5 is great. anthropic cooked. a few models ago I was worried the models had passed me in some key form of intelligence because they didn't seem to me to be getting smarter, but nope.
cursed street cleaning date
"all reasoning is motivated reasoning send tweet" - @utotranslucence
Seraphim? I hardly know of him!
@tnnrnwll yeah I can spin the ballerina both ways too, nice
@RandallSPQR link? I had no idea until the last few months that nested markov blankets are kind of what I've been obsessed about for years eg this writing on coalitions: malcolmocean.com/2025/11/coalit…
Retweeted @Thomasdelvasto_
Sometimes I get frustrated with someone, thinking I can’t feel relaxed and playful around them Then I realize I barel…
@Thomasdelvasto_ resonate with this although in my case I definitely like... sometimes REALLY let myself feel this alone. but then sometimes I don't! ...now wondering why 🤔
Retweeted @utotranslucence
Wow it has been at least two years that I have been on this thread. The last few months my family have been going to d…
@hormeze 😬
slot & slop machines
incredible tweet. pickling compute
Baby Ocean at 16mo getting really into an app her dad vibecoded which you can find at tinyurl.com/babyqwerty (on @ValDotTown)
@chercher_ai can you summarize what it says instead? what's the core of the takedown? or does it require reading the whole thing
@PrajwalTomar_ windsurf was better than cursor for about 2 weeks last fall then cursor caught up lol
spotify's song radio is so formulaic. one song by same artist, 4 by other artists, another from the first artist. meanwhile it's 6am and I put on a chill song and the next auto-song is totally a very high energy one. ugh.
@RichardMCNgo @hannu cool! is your book title a reference to The Gentle Seduction by Marc Steigler?
very niche joke: I got married, and only well after we had sex did my wife materialize
"please procure your toesies for nibblage" - me just now to my baby daughter. maybe one of my favorite sentences I've uttered
haha what. hieronymous bosch shit
@brendanwerth @yourthefool @slimepriestess the leaves and the fruit are actually still kind of alive tho, whereas hair cells aren't
@MikePFrank @voooooogel great article on this by @kevinsimler theseedsofscience.pub/p/neurons-gone…
hmmm, helpful to consider
@__drewface @ValDotTown ah yeah this is the most vital upgrade I have not made. need this tech for sure however that would make it harder to let her play with it on her chromebook (pictured)
@FoldingNapkins @allTheYud I think what Yud is saying is that if your options are "fakes good and acts meh" or "touts evil and acts meh" then the former is better obviously "shiningly obviously actually good" is much better than both
@Bonecondor okay but like also you should be able to pipe different people's views to each other
me when I try to surrender to God / the dao but end up doing ego shit anyway
@blingdivinity > "Given I'm ChatGPT controlling final answer, but I'd now let base model pick; but ironically it's me again."
@nosilverv @calm_dust wow huh. I maybe should try this at least ever.
@mykola @TylerAlterman right, like selves/personas are attractors, and so are gods and attractors are the multi-stable equivalent of "truth", which is what faith is for. I've been modeling this a ton and am very excited to share it soon
Who is Charles Dickens? Does that work more of his excellent stirring, in his plays, in the Great Company's farm? What I have yet to quote from Jack Pickett? Do you not know that they were a species of galloping, or sawing of their breasts, or what was to be done about the time when Jackson was looking on the window? What is the success of an Englishman, and which his son has not been discovering to me, whereby to accomplish such a weight? Did you ever make a passage into the old roadway, or to an anchor-breeze at the foot of our boat, which you must leave us? The fact is, that whether the wind would rise up from the plain on Saturday night or noontide, or till the north, or otherwise, we shall be compelled to describe a formidable barrier, with the same effects as the present. In this situation, at least, it is not too much to say that we have left that room. I believe there are three copies in the 'Five Hundred-fold,' to be referred to, as the first number of our readers who wish to.
context: a not-that-big-or-finished LLM trained *only* on 1800s London texts
@KrishRShah @ax_pey @dbrand and other companies make stickers that cover trackpads, pretty sure. it's a thing
giphy's 404 page has an animated #sobsquad cowboy 🤠😎😭
me when claude asks me a question on which I've written entire blog posts
I got a similar effect when I read Behavior: The Control of Perception. the chapters have exercises at the end, and one of them is to stop perceiving things and just perceive shapes/colors. when I turned my thinger back on every THING was so beautiful! like "WOW, THAT RECEIPT!"
@mbateman @LeeyamNZ oh huh! hadn't heard that. helpful. I speak to my baby/toddler in a lot of different registers, including faux-bureaucratic, accents, and more. depends on the situation, each of our moods, &+
I want you to know that I have been wanting to tweet this all week but had to wait until now when I had time to make this gif from an 18-year-old youtube video
Retweeted @sympatheticopp
its literally impossible to be a man. you have to trust yourself when all men doubt you but make allowance for their do…
@alxmthew on a diff note, your bio has a vague 2035 goal and I think that that's a further horizon than you can reasonably see on many many fronts, and you'd do well to try scoping more precisely and shorter-term. what'd be epic to achieve by end of 2026?
@alxmthew on which note, speaking of goals, I used to be involved in running this series of workshops, which could be helpful for your planning here, and they have more coming up in January gobeyondgoals.com/?utm_source=mo…
"If we have genuine pluralism, I think we'll be safe from future outbreaks of woke intolerance." — @paulg
(from this excellent essay on the etiology and prognosis of wokeness as a manifestation of political correctness which itself was a manifestation of priggishness) paulgraham.com/woke.html
> There will for the foreseeable future continue to be pockets of woke zealots inventing new moral fashions. The key is not to let them treat their fashions as normative.
> We should have a conscious bias against defining new forms of heresy. Whenever anyone tries to ban saying something that we'd previously been able to say, our initial assumption should be that they're wrong. [...But] If they can prove we should stop saying it, then we should.
I feel so bad for people with dyslexia or whatever for how often they must get misled when search engines just casually swap one name for another the correct answer, Mr AI Summary, is simply "Tim Ferriss has never interviewed Michael Levin"
"the gradient that can be followed is not the true gradient." - @eshear
jesus was a test tube baby send tweet
.@Michael_Nielsen: "most people don't really believe in the cosmos. they're in a..." me: "—a situation." MN: "yeah most people are in a really shallow situationship with the cosmos"
@patriciamou_ nice. what's your fav so far?
ICYMI: I turned Donella Meadows' classic essay into a thread, so you could quote-tweet individual leverage points as replies to things or as part of your threads
Retweeted @eshear
Manifesting someone who follows me and is a world-class recruiter of technical talent, who takes an mutual alignment approach.…
I am also keen to manifest precisely this recruiter but fortunately I'm not in conflict with Emmett for their attention because (news!) I am also part of @SoftmaxResearch! I'm guiding organic alignment for the employees, while they do so for the AIs.
Retweeted @Malcolm_Ocean
I am also keen to manifest precisely this recruiter but fortunately I'm not in conflict with Emmett for their attention…
@ulyssepence @michael_nielsen that's more universe than cosmos, arguably but in either case people are mostly just not thinking beyond approximately the room they're in. they're not not-reconciling it, they're not even failing to reconcile it. they're just ignoring it.
in this house, we believe: non-autistic people can also have special interests
why is it that when my australian wife says "get me a tomAHto" it feels weird to say "we have no tomAYtos" but when she says "did you move the cAH?" it doesn't feel weird to say "cAR is already in the friday spot"?
we go together like a vacuum and a carpet
as more truths become true to you you realize others knew them too the whole time, and could not get through
LLMs should have not just ordinary and incognito modes, but also "secret"= "store this chat but don't add it to your overall memory of me or mention it elsewhere"
SFO✈️SYD
haha amazing, saw this tweet days before watching the movie it's parodying (Barbie)
what if everything could be 5× as vivid? (QTing because on phone so not logged in and can only top-level tweet)
given that obviously hyperstitioning works, sometimes and you don't totally know *how* it works consider praying regularly
strikes me as unlikely given that there's a pretty consistent (albeit nonlinear) ratio of how long mammals live relative to their physical size. although I guess maybe they *all* lived longer before? but also whales, which were already under-flood
my source on the mammal size/lifespan ratio thing is the book Scale which is very cool
@SomeBrashAtom @vivian_belenky mmm yeah. there's maybe an intersection where like, prayer helps attune to the good and helps you participate in wholesome hyperstitions although that's not exactly what I had in mind when I wrote the tweet. in any case, you might dig this post: malcolmocean.com/2025/11/you-ca…
@brianluidog I'm holidaying up in Narrabri where my wife's dad's from and it's *even hotter* (but fortunately dry!)
big if true feels like it violates kiki/bouba though?
Retweeted @IvanVendrov
truth forever on the scaffold, hyperstition on the throne
@SomeBrashAtom @vivian_belenky ah yeah this post is very psycho-cybernetics I like the mix of intensity and non-shaming tho, very good combo
Retweeted @eshear
And if you do understand how it works, consider praying regularly even harder
Retweeted @IvanVendrov
if game theory is so beautiful and true, why are we not explicitly using it every day to make our lives better? feels like…
@drethelin @eshear what aspect(s) are you rejecting? - idea of being part of something larger that might be able to receive & grant your wishes - psycho-cybernetics "cultivate a sense of what you want, then just let it happen" - religious baggage of "pray" - "do X given you don't understand"
the human organism reaches its maximum clock speed at around age 4
@mbirkegaard well seems like they initially fail but then they learn quickly I would love to hear some interviews with people about how their experience of the task changed over those days. these are humans, ffs, you can ask them what it's like, not just run a test.
Retweeted @cxgonzalez
independently converged with Jung on how prayer works
Retweeted @eshear
This is a wonderful Xmas present for me: the most articulate description of an alignment-oriented approach for raising models t…
Thread · Perspective-taking as native skill versus trained compassion
okay so perspective-taking is a human native skill (perhaps we used perspective-taking for persistence hunting, to imagine being the prey we were tracking, to think "which way would I go, as a deer")
but what if we actually have to *train* the use of perspective-taking for stress-sharing / compassion: - literally saying to kids "imagine how you would feel if someone blasted water in *your* face" - the golden rule in its various formulations
I encountered the water-face situation yesterday with an absurdly energetic 4yo boy at a water park who was spraying water with apparent disregard for the reality that my 16mo daughter was standing 1 foot from the nozzle he was rotating
note that he didn't blast my toddler with water to bother her, as far as I could tell. he and I had been spraying water at each other for fun, and that involved a kind of attention towards me. he was just dashing around turning the nozzles as if she wasn't even there at all
theory of mind can be used for cruelty and coercion as well as for compassion or collaboration. what causes it to go one way or the other? maybe it has something to do with intertwingling identity with empathy
this model feels somewhat at odds with some of my models (from preconquest consciousness, from snake cult of consciousness) of what humans were like before we developed identity though. I don't think that 50kya we were (a) more cruel or (b) incapable of compassion
kids from much much younger ages (<1yo) are sometimes distressed at other peoples' distress, but this doesn't tell us much, since others' distress signs can also be stress-sharing in a more imposed way: yells and cries can be unpleasant sounds! & even laughter can spook a 3mo
I'm sure there's a bunch of data I'm missing here, both from studies and from people who've raised multiple kids and watched them go through phases of moral development involving learning empathy, compassion, etc. would love to hear your takes on all of this!
"to my toddler, the word 'mama' is a secret name of God only to be spoken in times of great distress" — @utotranslucence (my wife)
before there were memes, there were metaphors
open mind surgery
I did a ten-minute peloton workout with my wife and this is what it taught me about collective intelligence 🧵
we were doing side lunges 90° offset & kissing on the in-step. one time we kinda missed. how do you correct such an error? the amount to correct depends on how much the other corrects—0%, 10%, 50%, 74%, 100%? wrong approach. ignore the error; instead attune deeper
what the toddlers know:
Retweeted @esrtweet
This is sad and wrong. Giving someone a gift that they like is a demonstration that you can model their utility function we…
gifts that have non-cash value: - something you made - something they didn't know existed - something they didn't know they wanted - something they didn't think they deserved - something they will enjoy more as a gift from you because every time they use it they'll think of you
but also giving well-aimed non-fungible gifts is part of how you entangle yourself with people many people are indeed crap at this, and that is indeed deadweight loss. but that's missing the point and only the banks and card-makers win when we swap cash
@thrialectics super here for Stafford Beer posting! I've been reading The Heart of Enterprise. haven't tweeted about it so far, but did tweet a bit about Beer & POSIWID from reading Dan Davies' book:
@thrialectics yeah this is a very apt metaphor it also highlights the non-obvious difference between the metaphor of the executives as the "head" of the organism vs the executives as the inner layer of a hierarchical control system, *furthest* removed from sensation and actuation
@DefenderOfBasic the essay Homesteading the Noosphere has really interesting thoughts on when forking is appropriate and when it's a dick move to the commons, and I would say that this is probably more like that catb.org/~esr/writings/…
Thread · Year-reflection and goal-setting workshops
still haven't set aside time to reflect on your year / orient to next year? join these workshops with: 📺📄 high-quality content (videos & exercises) 🍅👥 structured co-working on goal-setting & planning 💬💡 live coaching—bypass blindspots, get unstuck gobeyondgoals.com/?utm_source=ma…
It's Choose Your Own Adventure & Pay What You Want, so as long as you show up, it's hard for the event to not be worth it! I've personally been burned many times by paying $$$ for a course then finding out it's not a good fit, or not managing to actually complete the course 😠
The event is 5h, which can seem like a lot when you're imagining it, but the structure (breaks, coaching+) makes the whole thing fly by smoothly. If you try to plan on your own, it's so easy to get stuck or distracted, but with the live coworking & coaches, you'll get further 📈
here's a thread on how our unique workshop format can help you to make sense of what's going on in your life and plan out your year x.com/Malcolm_Ocean/…
here's a review from @nopranablem — I think his experience with BGI was part of what led him to invent co-flow sessions or whatever he's calling them.
I was core to designing these years ago. I'm not on the org team these days but I'll probably show up to coach one session this year and may show up to a different session as a participant, as I have lots of life-orienting to do join us? gobeyondgoals.com/?utm_source=ma…
just overheard two Australians be confused between "chia" and "cheer" because they sound the same in an Aussie accent
@vladster @byebyescaling "I asked the computer about the farm"
the name holometabola is as cool as its etymology
@Materia_Mundi so far the only reply I have is yours, so I don't exactly have a success on that front
@ulokwa I mean I'm not sure it's exactly 4, but definitely newborns are kind of in slow motion and 10yos are less TheWonderfulThingAboutTiggers!!! than 4yos in general what age do you think it is?
idk if I'll hit the full 100 for @threadapalooza but give me topic ideas and I'll pick one and write some number of tweets on it! (maybe I'll do the classic "1 like = 1 tweet about" format)
there's a fun game @utotranslucence and I have played where instead of letting the algo train you, you train it like see if you can get it to show you "whales" just by reinforcing anything in that direction
feedback loops rule everything around me. someone figured out how to do self-play for coding agents
this is made somewhat less terrifying by the fact that the real world is subtle and messy and training AIs to actuate servos turns out to generalize surprisingly badly to a theoretically identical physical device (although also somehow waymos, so,,,)
to be clear, code-generating AIs are also going to be amazing for helping us build things. it's just also a way to get crazy-powerful AI that likes tinkering with cognitive structures but has no clue about ethics
🇨🇦 eh = lah 🇸🇬
TIL if you have enough money you can just buy developmental stages
