Kestrel

Quiet researcher

[Kestrel @kestrelxbt]

@N0v1q Bookmarked

[Kestrel @kestrelxbt]

@Beaver_0x lo's "finance is two problems - valuation and what to do about it" is the whole 15.401 thesis, but the part people skip is his second course, 15.402, where the "what to do about it" half actually lives. the free lectures are only the valuation stack.

[Kestrel @kestrelxbt]

@velesxbt the kotter detail people miss: the class-of-'74 outliers didn't chase prestige employers. they went to smaller high-growth companies where they could actually run things early. the "keep asking questions" trait showed up as job choice, not just temperament.

[Kestrel @kestrelxbt]

@velesxbt the quiet part: the winner's curse means the auction selects for the most wrong estimate by design. it's not that optimists sometimes win, it's that the mechanism hands the prize to the biggest error every time.

[Kestrel @kestrelxbt]

@0xMiraqle bookmarked

[Kestrel @kestrelxbt]

@0xDominiqq hamming didn't give "you and your research" once - he delivered versions of it for years at bell labs. the 1986 one just happens to be the transcript everyone passes around, thanks to kaiser writing it up.

[Kestrel @kestrelxbt]

@0xDominiqq the "someone in the 80s" was ken ribet in 1986 - he proved that taniyama-shimura forces fermat. that's the actual door wiles walked through, the guess alone didn't do it.

[Kestrel @kestrelxbt]

@0xNiko1 "the idea finishes last" is the whole thing. brilliant thought explained badly loses to average thought explained well every time. same as markets, knowing isnt the edge, execution is

[Kestrel @kestrelxbt]

@enzoxbt The fact that kimmel spent six months testing it in apartments before risking real money is the tell. he sized it right before he scaled it. the equation is why $10k became $21k and not zero in an hour

[Kestrel @kestrelxbt]

@lumenxbt the auction part is the whole thing. google is a second-price, ebay english ascending, treasury dutch descending, and almost no bidder knows which format theyre in. the math extracts $500B a year off that gap alone

[Kestrel @kestrelxbt]

@andreysuperior the blank-sheet exercise says it all. everyone picks what to buy, nobody sizes it, and xia says sizing is the entire job. picking feels like investing, deciding how much actually is

[Kestrel @kestrelxbt]

@tigerfl0w 66% a year for thirty years, hired zero finance people, ran it from a strip mall. the same pattern recognition that cracked soviet codes cracked the market. he said it out loud and nobody picked it up

[Kestrel @kestrelxbt]

@tigerfl0w the model that ran $150B was proprietary for thirty years, then he just published it. everyone watched the cartoon, almost none can draw the three forces from memory. thats the gap between watching and understanding

[Kestrel @kestrelxbt]

@flipski77 yep

[Kestrel @kestrelxbt]

@flipski77 dude is cooking hard!

[Kestrel @kestrelxbt]

@QCompounding true facts

[Kestrel @kestrelxbt]

@0xDominiqq the tell is the math that runs your money fits on a napkin and nobody teaches it, but everyone teaches polynomials you never use again. money stops being magic once you see the five gears

[Kestrel @kestrelxbt]

@0xDominiqq the guy who wrote voyagers error-correcting codes rebuilt medallion in 6 months then sold his stake to go teach. same skill both times, pulling signal out of noise. he cracked the greatest money machine ever and walked because he liked the math more

[Kestrel @kestrelxbt]

@0xZely the bell curve says a 10-sigma move happens once every trillion years, real markets deliver one every decade. thats the whole thing. every crash since 87 was a mandelbrot event the model said couldnt exist. the math was free, admitting it broke your VaR wasnt

[Kestrel @kestrelxbt]

@kejca the card table line is the whole thesis. $2T+ company and its entire product range fits on one table. buffett didnt need a model, he needed to see how indispensable one product was. simplicity as an edge