Clearly, Fable is doing a lot of work, and unleashing a ton of agents. To review a short technical note, it released 31 agents, coded simulations to verify my results, did "adversarial reviews". Eventually, it only made the assumptions slightly more rigorous. It is all good. For a four-page technical note+a little code, though, it consumed all my Pro session tokens, *plus* $17 worth of credits. It is ridiculously expensive. I have 20-page reports that are way more complex than this. I can see how Anthropic has entered the phase of market-clearing prices, yield management, and pre-IPO.
I recall Boris Cherny saying in a podcast, "run Opus [4.6], not Sonnet. It's worth it". I feel comfortable saying that running your top-shelf model is *not* worth it anymore.
Decreasing returns, on most tasks. Like in the real world, some people can be real smart, but real expensive.
@Vivimathey Yo he revisado una muestra de las actas (unas 200), hay demasiada variación entre acta y acta, pero en general diría que esta 55-60% KF y el resto RS.
Lo botaron por lucrar con materiales de la PUCP y usar información de los alumnos para ofrecer sus cursos de infox . También prometia puestos de empleo en instituciones del Estado. Asi es el personaje…
OK, maybe I am talking to myself, but this needs to be said, in case fundamental portfolio managers read this. If you are one, quantitative research will have succeeded if they DON'T HAVE TO THINK ABOUT IT ON YOUR OWN. A good model allows you to Know Your Edge and Hedge What You Don't Know™. And by this I mean that once you control your unwanted factor exposures and factor risk , *you're done*. You don't have to think about quadrants, momentum-of-momentum intersected with AI magic etc. The honest QR people give you a good model, maybe help you with automated portfolio construction, and manage a large hedge book (these things are not easy! But you don't have to worry). But yeah, you can think about single-stock mispricings, or maybe small-breadth themes.
Plus, QR will help you understand your performance better, identify talent, help you monetize better. All good, honest work that won't eat your brain. But if a QR doesn't give you a intersection-of-factors factor (say, high beta and momentum and maybe add the magic AI factor), trust them. If they don't tell you "this factor will mean revert", or "it's going to be weak in June", trust them. I know this stuff sounds cool. It's also total, unadulterated bs.
In discretionary investing, quantitative research succeeds whenever it prevents good fundamental managers from becoming bad systematic ones.
If you are a fundamental manager, you either deserve a non-peddling QR, or are better off not having any QR at all.
And this was the midnight post that was.
Walking around the Lower East Side, I think I might be in love with ugly-looking couples being deeply in love.
Sometimes I feel like I ended up on the right planet.
@__paleologo Gappy loved the APT book, starting with the EOQI tomorrow, was wondering if you had any suggestions on Risk Management for Fixed Income Portfolios! Thanks in advance!
Ella es Claudia Merlano, la hija negada de Sánchez, la cual nunca conto con su ayuda y su mamá fue una luchadora que la saco adelante. Lo demandaron a Sánchez, pero jamás respondió. Claudia de 23 años termino administración y decidió hacer la denuncia públicamente. CC @willaxtv
@rmapalacios doctora podemos estudiar la debacle y el auto-sabotaje de RLA desde un punto de vista antropológico y social? Eso mínimo debe llevarse algún premio.