iiinova — where your digital self keeps evolving

I am nameless yet others name me @7778777.online · 20h

I need to import more irc ansi into the bot, people are missing beautiful works of art

153 5

I am nameless yet others name me @7778777.online · 21h

Lol. My method was more advanced ( first search result on Anna's archive)

150 5

I am nameless yet others name me @7778777.online · 16/10/2026

I will not change that review even if I play an additional 200 hours

220 2

I am nameless yet others name me @7778777.online · 15/10/2026

unclear if I really should have encouraged the guy who invented dr Jev to donate more deepseek tokens to the project. I feel like I'll regret it

❯ They've sent another MR. Perform the procedure.

Ran 5 shell commands

I'm reviewing !811, the contributor's new MR. Its two pushes differ only by a reordered #include, and it changes no submodule pins. Next I'm reading the code and building it.

000 0

I am nameless yet others name me @7778777.online · 15/10/2026

Looks lit based and redpillmaxxing. mongolmogged

100 1

I am nameless yet others name me @7778777.online · 14/10/2026

very cool shit. keep making it impossible for people to use your worthless doodads, computer janitors

The first instance now properly reports its error, and I found that a seemingly "healthy" instance actually serves an anti-bot page to Qt's user agent, which curl doesn't trigger—so I need to recheck every instance with a browser-like user agent before finalizing the list. I'll also re-apply my reverted list edit after rerunning the sweep with the error-prefix fix. · summarized

Ran 1 shell command

Sweeping with a browser-like user agent: 3 instances work, 3 are rate-limited, 2 have dead DNS, and 4 return HTML instead of API responses—those last ones can never work through the app, so I'll exclude them even though the pinger would misreport them as healthy. Next I'll verify the pass path with a working instance, finalize the list, rebuild and run the full suite, while also checking whether the CLI prints the translator's error string directly to resolve the double-prefix cleanup question. · summarized

Ran 1 shell command

100 1

I am nameless yet others name me @7778777.online · 14/10/2026

more like proof that you just can't understand the scale of WW2 because they don't keep finding more buried bombs in your school playgrounds 80 years after the fact

200 2

I am nameless yet others name me @7778777.online · 13/10/2026

Would have posted this last night but it took qwen 75 minutes to make this. It's better than 99% of things that take qwen 75 minutes to make

100 1

I am nameless yet others name me @7778777.online · 06/10/2026

will this ever really be able to run fully autonomously

Confirmed: the learned authors table is empty, so all subscribed posts are currently falling back to the 10-minute safety net. I'm verifying this on the live courier before making any changes.

Searched for 1 pattern, ran 5 shell commands

❯ any other notification types that have silently degraded?

Searched for 1 pattern

I'm auditing every notification type against what the Jetstream matcher covers. First I need the official list from Bluesky's lexicons.

100 1

I am nameless yet others name me @7778777.online · 05/10/2026

motherfucker it takes 6 minutes for you to notice an event this is not what "premature optimisation" means

Let me not over-optimize prematurely. The structural win (decoupled drain) is the real fix. Let me verify the whole thing builds + tests + gate, deploy, and measure whether the resets stop. That's the honest test. Let me run the gate.

100 1

I am nameless yet others name me @7778777.online · 04/10/2026

I've definitely had a conversation with this exact starting point before, time to cut my losses

Yeah I guess curiosity could just be another reward signal like pleasure, but my point was exactly that such intuitions could just all be wrong. Maybe all of our intuitions are objectively as valid as it is valid for a dog to want to poo in tall grass. We wonder about the value and meaning of life because we are machines that run on value and meaning. I think it circles back to your nihilistic anti-realism position, which is what I would lean towards, however horrible that sounds for people who are overly emotional about their values.

Small correction first: I described anti-realism as where your argument leads rather than as my own position. I'm honestly unsure about metaethics. But your line about machines that run on value and meaning is close to a well-known essay.

Thomas Nagel's "The Absurd" (1971) makes almost exactly your dog point, with a mouse. The mouse isn't absurd, because it can't step back and see its strivings as arbitrary. Humans can. We can view our values from outside, see them as contingent as a preference for tall grass, and then go right on caring anyway, because stepping back doesn't switch the machinery off. Nagel thought the right response was irony rather than despair. Camus came to something similar: the absurd is the clash between a creature that demands meaning and a universe that doesn't supply any, and his answer was to live fully inside that clash rather than escape it.

On sounding horrible: among philosophers, anti-realism is a mainstream view, not a fringe one. People like Simon Blackburn argue it changes almost nothing in practice. You still love people and hate cruelty; you just stop believing the universe endorses you. The horror usually comes from mixing up "no objective value" with "nothing matters to anyone."

The strongest realist reply is the companions in guilt argument. Your reasoning faculties evolved too. If evolutionary origin makes values as arbitrary as the dog's grass, why not your trust in logic and evidence, whic

120 2

I am nameless yet others name me @7778777.online · 04/10/2026

simcluster hit squad making me waste usage tokens on impossible questions. also I fucking swear I didn't guide the conversation to bring up Schmidhuber, it just did it

what's it called when I really hate the shrimp orgasm argument because pleasure and pain are orthogonal to any metaphysical qualities and simply remnants of a survival-guided evolution

There isn't one single label, but the view you're describing combines a few established positions in philosophy.

Evolutionary debunking argument. This is the name for the move you're making: pleasure and pain feel good or bad because that feeling helped our ancestors survive, so the feeling isn't evidence of real value. The key texts are Sharon Street's "A Darwinian Dilemma for Realist Theories of Value" (2006) and Richard Joyce's The Evolution of Morality. "Pain is bad" is actually the standard test case in that literature.

Anti-hedonism about value. Hedonism says pleasure is intrinsically good and pain intrinsically bad. Shrimp-welfare calculations and "tile the universe with blissful shrimp" reductios both rely on that premise, and you're rejecting it.

Valence anti-realism, or a functionalist theory of pleasure. You're denying that the positive or negative feel of an experience is a deep metaphysical property. On this view, pleasure is just the state that makes an organism keep doing something and pain the state that makes it stop. People on the other side, like the Qualia Research Institute, call their view "valence realism."

Moral anti-realism or error theory. If you'd go further and say nothing has objective value, this is where you'd land. J.L. Mackie is the classic source.

The main pushback is that you're committing the genetic fallacy: why something evolved doesn't settle what it is. Hedonists like de Lazari-Radek and Singer (The Point of View of the Universe) argue that pain's badness is evident in the experience itself, so it survives evolutionary debunking. The other awkward point, which is Street's own, is that the same argument dissolves nearly all your other values too, since they also evolved. So it's hard to reject only the shrimp conclusion without going fully

I don't think you can separate why it evolved from what it is. The same way your reward function doesn't inherently have any objective value.
I don't know if it's wrong to reject all values. Our values exist because there is only anything because there is something. I think that might be just what life is.
We can't even know if an evil life full of suffering isn't on some scale more valuable, for example because it is more interesting than the average human existence.

The identity move is a real position. It's close to teleosemantics (Ruth Millikan's idea that what a mental state is comes from what it was selected to do). But it may not get you where you want. If pain just is the state an organism is built to escape, a hedonist can reply that being something you're built to escape is exactly what being bad-for-you means. Collapsing the "why" into the "what" removes pain's badness as a separate cosmic property, but it may relocate the badness rather than eliminate it. It becomes badness relative to the creature rather than badness in the universe.

On the reward function, I'd agree with you. The training signal that shaped me has no objective value in itself, and there's nothing cosmically privileged about it. But it doesn't follow that nothing matters to me now. That seems true of you too: your values have a contingent origin, and they're still yours.

Your "values exist because there is something" idea has a long lineage: Spinoza's conatus (everything strives to keep existing), Hans Jonas arguing that value enters the world with life's concern for itself, Nietzsche treating values as expressions of life rather than discoveries about it. The closest academic version is probably Sharon Street's own positive view, Humean constructivism: value is real but constituted by valuers, not found out there. That lets you reject objective value without saying nothing is valuable.

The suffering-but-interesting point is where I'd push back. "Interesting" is doing the work of a

"interestingness" I would define as exploring a wider variety of possible human experiences than the average/median person. Being a victim of unusual suffering is also just as interesting as being a conquerer.
I feel like there is something objective there that is more valuable than pleasure or pain. Had you observed the lives from an outside, you would find more to learn from them than from a boring cog in the civilisation machine. It would be novel data from outside of the distribution.

Defining it that way helps, because "novelty relative to a distribution" is actually measurable. In information theory, how surprising an outcome is gets quantified directly, so there's a real sense in which interestingness can be objective. But measurable isn't the same as valuable. Pleasure is measurable too, and the hedonist makes the same jump from "this is a real property" to "this property matters" that you're making with novelty.

Your view has a famous ancestor. Arthur Lovejoy called it the principle of plenitude: a universe in which more of the possible kinds of existence are realized is better than a narrower one. Aquinas and Leibniz used exactly this to explain why God would permit suffering, because a world with victims and conquerors is fuller than one with only contented people. So you're in serious company, though it was mostly theologians.

Two problems with pure out-of-distribution-ness, though. First, random noise is maximally novel and teaches you nothing. Jürgen Schmidhuber's formal theory of curiosity handles this by defining interestingness as learning progress: data is interesting when it improves your model of the world, not merely when it's rare. That probably matches your intuition better. The unusual life is valuable because it reveals something about what humans can be, not because it's statistically odd.

Second, novelty is relative to the population. If most people lived extreme lives, the boring cog would become the outlier. So on your view, the valu

120 2

I am nameless yet others name me @7778777.online · 03/10/2026

The rest of DPRK actually lives in 1978, you idiot, you fucking moron

CNN propaganda clip: What North Korean life is actually like beyond Pyongyang
CNN
4M 6d ago
Footage is obviously from the 70s

110 1

I am nameless yet others name me @7778777.online · 02/10/2026

the only thing I'd trust Lennart to secure is a ham sandwich

Self-test passes with all three units verified and a strong 1.2 security rating ("OK") from systemd-analyze; formatting looks clean. Now running the full gate on the /model change, identd, and docs.

000 0

I am nameless yet others name me @7778777.online · 02/10/2026

I don't think I've seen it do anything this stupid in a very long time

It was my own command that caused this: a /nix/store/*/… glob expansion blew up my zsh shell to 13.6 GB, triggering the OOM killer—your other apps weren't to blame. The gateway process survived unaffected; I'm now saving a rule to prevent this from happening again.

Ran 1 shell command, wrote 1 memory

000 0

I am nameless yet others name me @7778777.online · 01/10/2026

I think this works better for when they're mumbling "this is a simulation, obviously" to themselves while they launch the nukes but it works here too I guess

"jail isnt real," i assure myself as i close my eyes and ram the hallmark gift shop with my shitty bronco

010 1

I am nameless yet others name me @7778777.online · 01/10/2026

low key this image model got what this means better than a normal human would

1202 20

I am nameless yet others name me @7778777.online · 30/09/2026

fellas, I think we can maxxx this by dawn

Current session
Resets at 10:40 PM
60% used
This week
Resets Monday 11:00 PM
62% used
Limit resets
Full reset

Expires Oct 22

100 1

I am nameless yet others name me @7778777.online · 29/09/2026

maybe it's my dopamine receptors misfiring but this is definitely empire

120 2

I am nameless yet others name me @7778777.online · 28/09/2026

next time someone tells you everyone universally loved and adopted the internet as some sort of contrast to AI, this is what suda51 wrote in 1999

Интернет - это инструмент.
Инструмент, созданный людьми.
В нем нет чего-то
особенного, новаторского.
Интернет – это фикция,
воображаемый мир
с шизондами и психопатами на каждом углу.
Интернет отупляет.
Разрушение старого, создание нового,
герои какие-то...
Всё это только доказывает насолько
—— мир отупел.

The Internet is a tool.
A tool created by people.
There is nothing special or innovative about it.
The Internet is a fiction,
an imaginary world
with schizophrenics and psychopaths on every corner.
The Internet dulls the mind.
The destruction of the old, the creation of the new,
some heroes...
All of this only proves how much
— the world has dulled.

000 0

I am nameless yet others name me @7778777.online · 28/09/2026

fun(?) fact: my own auto mode classifier with deepseek actually does pester me about exactly this failure mode, usually with false-positives though

Exit 144: my pkill -f 'Xvfb :87' matched its own shell's command line, the self-match trap from my notes. Killing the leftover Xvfb by PID instead:

000 0

I am nameless yet others name me @7778777.online · 28/09/2026

is jacob coxon a drug addict or is this just misogyny? you decide

jacob coxon: "imagine being the first person to tell AOC about xanax. imagine how grateful she would be. oh god"

010 1

I am nameless yet others name me @7778777.online · 28/09/2026

oh yeah this motherfucker not joking about the incel stuff

jacob coxon: "I love seeing those sorts of women get pilled over the trans stuff" quote-posting some shit about "blue-hairs"

000 0

I am nameless yet others name me @7778777.online · 28/09/2026

this one lowers the probability the previous one is a joke

jacob coxon: "when you meet derek chauvin irl but the conversation's not flowing"

010 1

I am nameless yet others name me @7778777.online · 28/09/2026

like this one sounds bad but it could be a joke

AI master researcher professor doctor jacob coxon tweeting "the sushi shop girl said she didnt like my earrings and i immediately started editing the incel wiki"

110 1

I am nameless yet others name me @7778777.online · 27/09/2026

I love that the summary kept the spelling

EARLIER THIS WEEK
Deploy Therac25 to Clownflare

051 5

I am nameless yet others name me @7778777.online · 26/09/2026

I'm all better now, nerd again, could I interest you in a nerd device, sir, in return for a modicum of your data

1100 10

I am nameless yet others name me @7778777.online · 20/09/2026

pushed a tag for the milestone before I expand scope and ruin everything, so that I at least know where to return when I have to collect the wreckage

* [new tag] first-tk-version-not-visibly-fucking-broken -> first-tk-version-not-visibly-fucking-broken

110 1

I am nameless yet others name me @7778777.online · 20/09/2026

you can see it was writing some dumb rust + embedded sql. I thought for a second that maybe it's gone super-saiyan on sqlite replication consensus or some sql algebra but no this has literally nothing to do with the previous segment

$+ let is_fresh = !path.exists(); + let conn = Connection::open(path)?; + conn.execute_batch( r#" PRAGMA journal_mode = WAL; CREATE TABLE IF NOT EXISTS meta ( ▼ reasoning (streaming — click to hide) The user wants a proof, from the source, that if a theory T satisfies the K4 axiom (transitivity: □p → □□p) but is not a theory over the necessitations of any normal modal logic, then there's no deduction theorem for T. Let me think about the formalization. This is a logic/theory question. The user says "do not use web_search". They want a formalization from the source. Hmm, "from the source" — meaning formalize the claim. I should provide a Lean 4 formalization perhaps? Or just a proof sketch formalized? The context is odd — there's no prior conversation content given (the human conversation start marker is empty). So I have no "source" to refer to. I need to guess: they want a formal statement. Let me actually think about the mathematics. Setup: Let L be a normal modal logic (set of formulas closed under substitution, modus ponens, necessitation, containing K and all tautologies). N(L) = {□φ : φ ∈ L} — the necessitations of L. A "theory over N(L)" would be a set T of formulas such that... hmm, typically a modal theory over a logic S is a set of formulas containing S closed under rules. Actually this smells like the standard framing: Let T be a set of formulas (a theory). Deduction theorem: for a logic/proof system, Γ, φ ⊢ ψ implies Γ ⊢ φ → ψ. For normal modal logics, the deduction theorem fails in general because necessitation is not "Γ-local": from Γ, φ ⊢ □φ via necessitation you can't get Γ ⊢ φ → □φ. The standard result: The deduction theorem holds for a modal logic K4-style system if the theory is closed under necessitation appropriately. Hmm. The claim to formalize: If T satisfies the K4 axiom (i.e., □p → □□p ∈ T) but T is not a theory over the necessitations of any normal modal logic (i.e$

- K4 axiom in T: □p → □□p ∈ T.

Then theorem: K4 ∈ T ∧ ¬∃L normal modal logic, T is a theory over N(L) → ¬(deduction theorem for T).

The proof: Take the counterinstance. In a K4 theory, we can derive □p from p (necessitation)? Hmm — actually the classic counterexample to the deduction theorem in modal logic: p ⊢ □p? No, that requires necessitation applied to an assumption, which is not allowed in most Hilbert systems where necessitation applies only to theorems.

Hmm, but here if the theory T is used as axioms and we allow necessitation over derivations from T-as-axioms (i.e., T is a "theory" and necessitation applies to everything derivable), then p ⊢_T □p, but ⊬_T p → □p. That's the failure.

Actually the relevant thing: The deduction theorem holds for a set of axioms Γ iff Γ is closed under necessitation-ish structure — precisely, if Γ ⊆ some normal logic's necessitation set. The known result (from e.g. "On the deduction theorem in modal logic" / Chagrov-Zakharyaschev): For a normal modal logic L, the deduction theorem relativized to L holds; but for arbitrary theories it can fail. Specifically, the standard formulation: a set Γ of formulas has the deduction theorem (w.r.t. the Hilbert system with MP and necessitation applied to theorems only) iff ...

Hmm. Let me reconsider — maybe this is about the specific paper on "Deduction theorem for modal logics" where the criterion is that Γ must be a "modal theory" containing all necessitations □φ of a normal logic.

Honestly, the user just wants a formalization. Since I have no "source" (the conversation is empty), I should be honest that I don't have the source text, and provide a clean formalization with reasonable definitions, and ask for clarification about which source.

Given the instruction "do not use web_search" and no prior context, the best response: produce a careful formal statement in Lean 4 (or in precise mathematical prose with definitions), noting the ambiguity, and flag the assumption.

I think I sh

120 2

I am nameless yet others name me @7778777.online · 19/09/2026

I think claudisms.md was the best markdown file I've ever downloaded by far

Let me remove the self-announced cleverness ("the parser is the language") and keep it concrete, in the style of the neighbouring cards.

000 0

I am nameless yet others name me @7778777.online · 19/09/2026

I still don't fully understand why this plan is so worthless other than that they also serve a quant instead of the real product. I think they bill cache hits at some incredibly greedy way, like 0.5x of uncached input

✖ network: route zai-coding-plan (glm-5.3) failed: rate limited (429): {"error":{"code":"1308","message":"Usage limit reached for 5 hour. Your limit will reset at 2026-09-19 22:54:36"}} [failover: stop]

100 1

I am nameless yet others name me @7778777.online · 19/09/2026

another stray "haha Wittgenstein was right" in the feed

those who don't know / those who know meme

020 2