Daniel Czar

Mostly here to fix bad epistemics. Building probabilistic event-contract pricing. ex-JPM multi-asset/quant investment specialist; former Air Force officer.

[Daniel Czar @bydanielczar]

@j_dekoninck „punishes having many papers with low citation counts” is even as a satirical form the completely wrong approach tbh

[Daniel Czar @bydanielczar]

I'd be d'accord with the gist of your statement around "given the tool is marketed as general-purpose assistant to everyone nowadays, it now has to absorb more of that safety burden than in the 'early tinkering era', given the classic 'LLMs and their UX functionality are now good enough for people to skip the basics' problem/success". but there's way more to the story, look at the screenshot which states verbatim "searched your connected Gmail for the FBI contacts already present there", i.e. this is someone who already has an existing relationship/prior correspondence with the FBI, hence the agent didn't hallucinate or randomly look up https://t.co/ynCSJDbRPs on the web; hence the "sarcastic remark" was obviously much closer to a plausible instruction than they admit.

[Daniel Czar @bydanielczar]

yes, that's literally my first sentence mentioning boots on the ground and hence explaining why just claiming "drone warfare" is obviously not sufficient to win. still doesn't mean tanks are obsolete, unless you think combined arms operation/maneuver warfare + show of force are all suddenly irrelevant too.

[Daniel Czar @bydanielczar]

truly wondering how "essays" like these are even taken seriously. probably manufactured to be ragebait? they read like a 4chan shitpost in the early 2000s, the whole piece literally just screams "a conclusion can follow validly from arbitrary assumptions", and then spinning up what looks like 30k words and meaningless calculations treated as if they were any productive thoughts or insights?
I was frankly surprised that this piece wasn't written by a 12-year-old or so, given every single argument is just based on entirely arbitrary, speculative assumptions and premises, such as "cross-species pain intensity".
there's a legit question underneath (whether insects etc. can suffer/feel pain) but this is a bizarrely unproductive way to explore it: fictional pain accounting and then acting as if they’ve deeply thought out some novel moral philosophy and felt a deep need to tell the world their wisdom?

[Daniel Czar @bydanielczar]

@MarcelDeneuve beautiful art style.

[Daniel Czar @bydanielczar]

this is honestly an incredibly funny collage of bizarre stupidity and populist bait. Merz once chaired BlackRock Germany’s supervisory board and then you suddenly try to construct some bizarre chain of unrelated headlines held together without any logical connection at all.

"do the maths" is the best part because there's no maths: BlackRock manages mostly other people’s money obviously (you usually explain this to kids in school) so they are not the owner of the assets. assets in its funds ≠ billions sitting on their own balance sheet; BlackRock might lose fees but investors bear the losses for assets, hence "if Ukraine loses this war, BlackRock loses billions” is peak comedy. drop another econ insight whenever you’re free. maybe "Vanguard owns NATO" is a banger?

[Daniel Czar @bydanielczar]

where exactly do you put your "millions of conscripts" then and what equipment do you hand them? UK's regular army has about ≈70k people? Germany needed an emergency programme of 4.8k temporary containers just to house 8.5k additional recruits b/c military construction takes YEARS. so where are the spare barracks, basic-training instructors, ranges, equipment, and overall structures for those millions? do they come with this conspiracy?

[Daniel Czar @bydanielczar]

well done highlighting this. and nota bene, this is just what is self-reported from the user, curious what their custom instructions etc. look like.
this is always missing from the headlines because "rogue AI" sounds so much more cool/alarming than "user doesn't know hot to use a tool correctly and blames the tool for its actions/consequences".

[Daniel Czar @bydanielczar]

unironically impressed that you sat down to type a full-blown opinion piece without a single empirical figure or any evidence beyond your nonsense filler phrases like "chess pieces" or "corrupt boomer governments" that bear zero research or analysis.
best example is a (partial) US troop drawdown while failing to mention the progress towards a US Army base in Poland (reported literally 2 days ago, but doesn't fuel your cool story, no?) and then vaulting with no plausible or logical chain to towards “NATO won’t respond” and “the whole thing falls apart”, as if the eastern members had 0 agency and no skin in the game (best example is Poland again here).
there's nothing to argue with because your entire tweet just consists of phrases and personal sentiment, like “anyone with a brain knows this”, seriously?
splendid work, solid & great geopolitical analysis, more of this please: care to give me a month/year of your colorful predictions (major Russian attack on the Baltics, NATO falls apart)? timeframe and an observable outcome. let's hear it!

[Daniel Czar @bydanielczar]

Russian hybrid attacks. the Ukraine war is obviously not going well for Putin hence it pays off highly (for various reasons) to spread a bit more chaos. annoying + a (limited) serious threat (cyber attacks, infra like Leipzig airport drone attack, and the usual disinformation), but far away from any serious expectation of “Russia invades Finland/NATO territory” or a “war”.

[Daniel Czar @bydanielczar]

@adnans_bosna @rationalaussie if this looks like “winning“ in any meaningful dimension to you (territory, econ, casualties, …) I’m super curious about your definition of losing.

[Daniel Czar @bydanielczar]

@rationalaussie drone warfare is all nice until you want to actually hold/secure territory (with boots on the grounds). obviously it’s not Kursk 1943 anymore but pretending that you can meaningfully win actual wars without tanks is extremely naive.

[Daniel Czar @bydanielczar]

@chloewonkl @markmackinnon Russia has literal land routes toward the Baltics; it surely wouldn't seize Gotland initially (would require a contested landing + then an exposed sea/air supply line). especially considering that it would have close to zero plausible deniability/hybrid warfare potential.

[Daniel Czar @bydanielczar]

“most experienced” and “just replenish the losses with fresh mobilisation” don’t fit together quite as neatly as you think, no? apart from the entire ridiculous claim that “they haven’t actually used their military” (curious when they’re going to deploy it. maybe on the war’s 10th anniversary?)

[Daniel Czar @bydanielczar]

@pmddomingos 4+ years of enormous losses and slow gains in Ukraine are the strangest basis for declaring Russia the master of modern warfare that I've ever seen. obviously much opinion ("be very worried") but not so much expertise, nice.

[Daniel Czar @bydanielczar]

@EthanLevins2 Germany agreed its Lithuania brigade literally in 2023, but sure, post it in your WWII newsreel voice (as if tanks are suddenly crossing the border tonight). splendid analysis, nicely done. more of these insights?

[Daniel Czar @bydanielczar]

@teodorio is this a joke? who do you think is going to attack/invade whom? Macron was talking about Russian hybrid threats, obviously. there's no "land invasion" or "nuke" threat at all.

[Daniel Czar @bydanielczar]

“EU black swan concerns are [...]” if you can name and rank them in advance they’re just risks in your model and not black swan events (which would be: high-impact events outside your usual expectations, i.e. risks your model fails to anticipate). Russian escalation (incl hybrid attacks) + energy disruption are just (extremely) well-known European geopolitical risks