Lost

@PeterRO51425 @Mylovanov right so the huge territory thing you stated is wrong

@_l_ost_ @Mylovanov It’s struggling so much isn’t from the huge territory; instead, it is caused by boycotting.

Kasparov: Russia will lose territories after the war. The North Caucasus will likely break away first. China already eyes the Russian Far East and Eastern Siberia — lands that belonged to China until 1860. 1/

@PeterRO51425 @Mylovanov Is that why Russia is struggling so much

@Mylovanov annexing the Russian Far East would weaken China, not strengthen it. Your stance might reflect intention rather than your true belief. Huge territory is a burden, not an advantage; modern warfare has redefined 'strategic depth' since WWII.

@Stefano_Peron @GaryMarcus Is that AI generates

**La fragilità di un mondo monopolizzato dall’IA** Il mondo sta commettendo un errore madornale, e lo sta facendo con l’entusiasmo di chi crede di essere geniale. Ci hanno convinto che il futuro dell’intelligenza artificiale passi per forza attraverso poche mega-aziende che costruiscono modelli sempre più grandi, sempre più costosi, sempre più centralizzati. Pochi colossi che controllano l’infrastruttura, i dati, il calcolo e, di conseguenza, il pensiero stesso. È una stupidità collettiva di proporzioni storiche. La natura non funziona così. Il cervello umano non è un unico enorme processore monolitico: è un sistema distribuito, composto da miliardi di neuroni che formano reti locali, gerarchie, moduli specializzati che interagiscono in modo caotico ma resiliente. La conoscenza umana stessa è **frattale**: si ripete a scale diverse, dal singolo individuo alla comunità, dalla tribù alla civiltà. È ridondante, decentralizzata, capace di sopravvivere alla perdita di intere parti senza collassare. I piccoli modelli di machine learning — specializzati, leggeri, addestrabili localmente, modificabili da chiunque — riproducono proprio questa struttura. Un ecosistema di migliaia, milioni di modelli piccoli e medi, ognuno con il proprio dominio di competenza, capaci di collaborare, evolversi e sostituirsi a vicenda. Questo è l’equivalente artificiale di un ecosistema biologico sano: diversità, ridondanza, antifragilità. Invece stiamo costruendo un sistema monoculturale. Pochi giganti che dominano tutto. Se uno di questi modelli ha un’allucinazione sistemica, se viene compromesso, se i suoi proprietari decidono di censurarlo o orientarlo, l’intero edificio trema. È la stessa fragilità che abbiamo visto nel sistema finanziario nel 2008: troppa concentrazione crea rischi sistemici che nessun “troppo grande per fallire” può davvero mitigare. La concentrazione di ricchezza e di risorse computazionali non sta portando verso nessuna “singolarità”. Quella parola, tra l’altro, è diventata un contenitore vuoto, un termine mistico da conferenze TED che non significa quasi più nulla. Non stiamo andando verso un’intelligenza onnipotente e unificata. Stiamo andando verso un’intelligenza **fragile**, ostaggio di poche entità private che hanno incentivi allineati più al profitto e al controllo che alla vera esplorazione dell’intelligenza. Un mondo con migliaia di modelli piccoli sarebbe più caotico, sì. Più disordinato. Più difficile da controllare. Ma sarebbe anche più vivo, più creativo, più resistente. Sarebbe più umano. Abbiamo dimenticato la lezione più elementare della complessità: i sistemi più robusti non sono quelli più ordinati e centralizzati, ma quelli che somigliano alla natura — pieni di imperfezioni, ridondanze e scale frattali. Stiamo scegliendo l’illusione del controllo assoluto al prezzo della vera intelligenza distribuita. E lo stiamo facendo con la presunzione di chi crede di aver capito il gioco, mentre in realtà sta semplicemente ripetendo gli stessi errori di ogni impero che ha mai creduto di essere eterno. La decentralizzazione non è un’opzione romantica. È l’unica architettura che ha mai funzionato nella storia dell’evoluzione. Ignorarla sarà pagato caro.

Am I right that hyperscaling compute is the biggest bet in history? Any counter examples? It’s way more expensive than the Manhattan Project, the Apollo project, and railways across the US. If it doesn’t yield AGI, it may also be the biggest failed bet in history.

@zweigesh @EllliotttB Should be clear without having too, and him introducing David Deutsch, Kapil Gupta and so forth is no remorse

@_l_ost_ @EllliotttB Really? I’ve never bought a class from him, have you?

Everyone just says stuff, loudly

@zweigesh @EllliotttB He is a grifter

@EllliotttB Naval makes testable hypotheses that can be critiqued and falsified. If proven wrong, the error is corrected. Nothing wrong with that. In fact, it’s how knowledge is created.

@BitPaine @peterktodd 🤡

@peterktodd Trump has not achieved World Peace yet

So I actually spent Easter in a small front line town near Kramatorsk, delivering a van bought with ₿ donations. Literally dozens of armed Russian FPV drones attacked that town during this "ceasefire". Hell, I even spotted one; the soldiers around me kindly shot it down for me.

@DrugGovoruna @maddimogs thats afrikaans in the newspaper

@maddimogs A lot of Slavic vocabulary sounds hilarious to speakers of other Slavic languages. Dutch is the closest major language to English, so you have the same effect to a lesser degree. In 🇺🇦, "незабаром" means "soon", but in 🇷🇺, "не за баром" means "not behind the bar."

there's no fucking way dutch is a real language

@1440000bytes Oh these precious bitcoin experts

@goodalexander @tolstoybb Seems fairly obvious

@tolstoybb it has a right to exist

There’s literally no other reason to fight this war other than for Israel

@Stefano_Peron is this chatgpt

A useful trading question is therefore not: “How many people are talking about this?” But: “How often do I see this exact pattern of chatter when the story later turns out to be false?” If the answer is “very often,” the informational value of the signal is close to zero. Bayes didn’t fail. We just fed it the wrong assumptions. Social media doesn’t convince you with evidence. It convinces you by making noise look statistically meaningful.

Bayes and Noise…. Why X is dangerous in trading One of the most dangerous illusions on X/ twitter is how easily noise gets mistaken for evidence. Bayes’ theorem explains why. Suppose you see a rumor on Twitter/X: “A large fund is blowing up.” You see: • dozens of tweets • viral threads • screenshots • “sources” • market panic

@Egge21M If the actual productivity gains are there spend more on sales and marketing - it doesn’t seem to be a zero sum environment

@_l_ost_ There are very few places that did layoffs in favor of less devs and more AI YET

Wonder why AI development and LLMs are advancing so super fast? US has ~3M devs. The avg. salary is 130K. Building something that can lay-off even 10% of them saves businesses 39 billion every year. Solid business case. Not even including all the other positions it "optimizes"

@MarkBestForex @DanielWhiteson 😂 At least the LLM will tell me its right before its wrong

@_l_ost_ @DanielWhiteson I think its not an LLM problem its something to do with quantum super position but it's in the wrong position in both states. Then when you try to insert it, you collapse the wave function and know for sure you're wrong.

What do these people think "physics" is? Do they imagine it's a finite list of unsolved hard physics problems, the ultimate homework set? Do they think if an LLM solves them, we'll shrug and say "that's it, we have no more curiosity. physics is done" ????

@MarkBestForex @DanielWhiteson If I was running an LLM Lab Id reiterate every year how soon it is for USB to be a solved problem

@DanielWhiteson It means we will finally know why its impossible to put a UBS in correctly on the first try.

@Egge21M @callebtc Doesnt cashu use an database to track double spends ? Wouldn’t that be susceptible to the same accounting bugs ?

Being a bearer asset is one of Cashu's many superpowers. It gives builders a great security net. Imagine a system that pays out revenue to users. With an account system bugs can lead to double spends. With Cashu, this can't happen, as it's a record that needs to be reassigned.

@baconroast1 @tedcruz Yeah right give what he is saying more visibility because they are afraid of it

@tedcruz Fun fact, they talk about Tucker so often because the public knowing the truth terrifies them

Welcome, Ayatollah Carlson….

@EliBenSasson @callebtc Even if AI could do all our current work, human time would be allocated to where its more valuable ?

@callebtc I agree. And I also think that we are still far from AI doing all of our work for us. It can do some tasks, it still requires training and monitoring, it's advancing fast for some tasks, it's moving slowly for other.

The people using AI aren’t working less. They’re working harder than ever. If you think that’s failure, you’re about to get boiled like a lobster.

@callebtc Bet that the russia ukraine war wouldnt happen

tell me something you predicted wrong

@1440000bytes @LynAldenContact @MRBullMktEver Whats the alternative for eCash ?

@_l_ost_ @LynAldenContact @MRBullMktEver

I guess I don't understand why @LynAldenContact is making a point that you can self custody your BTC but not your NVDA stock? The only reason you'd care about self custody of an asset is if you're a criminal and you're afraid your assets will be frozen. Why else would you care?

@1440000bytes @LynAldenContact @MRBullMktEver Do you have a concise critique against fedimint (genuine curiosity)

@LynAldenContact @MRBullMktEver Is that why you support fedimint and cashu?