Senior PowerPoint Engineer

@depression2019 Who the hell was buying AMD of all things at the top

Absolutely insane shit happening over at Robinhood Social today

@kapoano @HML_Compounder @strictrope ya. if you're the type of person who is like "40% is fine" you won't hit 25x. But at least you don't go to 0.

@ryxcommar @HML_Compounder @strictrope All of it??

Bottom? 😭😭

@Merridew__ im long a bit of vol should i long more

@MariGO2thepolls "nothing in particular" is such a cowardly answer

@chrisalbon I don't think those things are like, directly causal.

One of my most defining experiences was 15 years ago watching two ppl with roughly same followers. One HAD to tweet the zinger, HAD to argue with everyone. The other mostly stayed positive and built. The former ended up having a middling career. The latter a monster career.

@HML_Compounder @kapoano @strictrope this was me when I sold MU in November 2025. 😭

@kapoano @ryxcommar @strictrope Well to be fair, people that sell when they’re up 25X also would’ve sold some when up 3, 5, 8, 12X too…

@jiggy_quant @DeepDishEnjoyer My original point and my continued point here is this does not require "utility" at all, and that it's not a useful framing of the Kelly criterion to think of it first and foremost as being derived from or having to do with a utility function. That's the real first principles.

@ryxcommar @DeepDishEnjoyer Try to think of it from first principles, each bet is a new bet so they aren’t structured together, they are independent, but your post bet utility is higher so u can bet more on the previous trade.

the nerdy zoomer that thinks kelly is too risk averse needs to be studied

There are a lot of people who wanted to be part of "the next big tech thing" who don't know how to discern what the next big tech thing will be so they went with crypto. Now it's clear that AI is the actual next big tech thing so they went to that. It doesn't hurt that both have a certain level of consumer ease of access compared to things like "quantum computing" or "the cloud."

I don’t think AI is a scam but it’s weird how it replaced the exact mindshare of a previous scam (crypto)

@jiggy_quant @DeepDishEnjoyer you dont "value you capital logarithmically", this arises purely from how the bets are structured

If u start reading my tweets and thinking instead of trying to show ur smart you will see these 2 statements are the same. You say yourself ur rule doesn’t apply for single trials, ofc each trade you do is a single iteration, so you have to trade ir according to the post trade utility, which is only logarithmic because you value your capital logarithmically because it allows you to put more size on future trials/ trading opportunities you believe might come.

@CL207 Parabolic with a negative coefficient

ok maybe i shoudlve sold everything when they made an index called the parabolic 7

@LinkofSunshine cmon basil you really that broke

(Sponsored post)

Ohio is slightly choosing “grandma gets thrown boiling water at her” over “everyone gets a puppy”

@flex_luther13 This did NOT work out well for me in the end

We're saying different things. At risk of repeating myself. It *doesn't* come from utility. All that's required for my utility function is that it is strictly increasing in wealth Kelly betting is asymptotically optimal, even if U(wealth) = wealth. This would not be true in single trials; in that case, use utility framing, point me in the direction of Jensen's inequality, whatever. But the motivation is maximizing a local objective (i.e. one trial) which is asymptotically strictly optimal (i.e. over many trials), and if you use this framing it doesn't require utility functions whatsoever.

@ryxcommar @DeepDishEnjoyer I mean I guess that is true, but it does come from your utility as well, guess we are saying similar things in different words. In practice you do have a lot lower utility for losing a ton because that means u can bet less in future opportunities provided u expect them to exist

@jiggy_quant @DeepDishEnjoyer On that note fractional kelly is just a form of regularization. (You could also laplace smooth your odds, different mechanism same effective result.) Which is to say naive kelly is unregularized, which is also to say it's the *most* aggressive you should ever be, re: p4's point.

Kelly criterion maximizes the growth rate of wealth under specific assumptions about what bets you're allowed to make (unquantized, unbounded, accurate odds). Maximizing growth rate also asymptotically maximizes wealth over many independent wagers. That's fundamentally where log(wealth) wealth comes from, not utility concerns; Kelly criterion is the FOC of the objective which maximizes wealth asymptotically over many trials. On a single trial and with variance introduced, you can think more in utility framing, I guess, but it's wrong to think of utility functions as the motive or where it comes from.

@alisha_hg that sounds a lil sad but on the other hand there are some upsides, like you'll never be truly heartbroken

@jiggy_quant @DeepDishEnjoyer I do not think it makes sense to think of log(wealth) in the kelly criterion derivation as first and foremost a utility function

Idk, pretty low confidence statement, but I do think logarithmic utility feels a bit aggressive? Do we truly have the same utility from going from 10M to 100M as from 1 to 10? I think there are a lot of situations where it makes sense (for example optimized bankroll for long term EV) but not clear to me that there aren’t q lot of examples especially in personal life where it’s a bit too aggressive?

@alisha_hg not even like a crush?

@FournesMaxime @PauseAI @sebkrier this is not a call to violence, it's a throwaway shit post using a common meme format

@PauseAI condemns incitations to violence against everyone, even as a joke, and especially deplores people in positions of influence inviting violence. We think the behaviour of Google Deepmind's Head of Frontier Policy @sebkrier pictured below inviting violence against AI safety advocates is unacceptable and we hope that measures will be taken by Google Deepmind to prevent such a situation from ever happening again.