codexarcanum

Hah! I win my shitty personal bet that he wouldn't make it a week without getting back to fucking golf! God how I hate that bloated, predictable shitbag

And now, it's time for 12 straight hours of the same catchy 30 second clip from this song you overheard in the grocery store!

I haven't gotten to the finance part yet but his discussion on the state of AI currently is excellent. I'm going to post something later about my experiments with DS-R1 in offline form and playing with "the big model" through their apps. The differences are pretty interesting and the big model is quite impressive.

Are you making any investment decisions today? I was thinking it might be a good day/week to "buy the dip" assuming NVidia actually is still a valuable company and will likely experience some overcorrection in the opposite direction. It's been jittering all day but the downward trend continues for the moment.

Edit:

Wanted to add that I also read this today and he has some different insights to DS as well. In particular, he thinks DS is being honest in their costs, and does some analysis on how their optimizations make their claims of computing on a lot of H800 chips plausible.

He also talks about how DS-R1 uses a somewhat different chain-of-thought/reasoning model than ChatGPT o1. They used RL (no HF!) to have DS "teach itself" how to reason. To me, that's maybe the most alarming bit. I'd been confident this approach wouldn't get to AGI but with these techniques... I dunno, maybe we're much closer to bootstrapping and the AGI revolution than expected? Like...within 2 or 3 years maybe?

Ah, the OP also mentions this:

With R1, DeepSeek essentially cracked one of the holy grails of AI: getting models to reason step-by-step without relying on massive supervised datasets. Their DeepSeek-R1-Zero experiment showed something remarkable: using pure reinforcement learning with carefully crafted reward functions, they managed to get models to develop sophisticated reasoning capabilities completely autonomously. This wasn't just about solving problems— the model organically learned to generate long chains of thought, self-verify its work, and allocate more computation time to harder problems.

An impeccable classic as well!

This movie?

I don't know why but the "fwip!" on pulling out the gun and then the carjacker's shocked look are sending me today!

Thanks for the clarification!

(I'll have to check the icons on my computer. I mainly use voyager on android and inline custom emoji support needs some work still.)

No problem! It only stood out to me because it'd be an amazing coincidence if 52! fit snuggly into 2^8. Being a nerd, I have many powers of 2 memorized!

With a little playing around, it seems like 57! is the last whole number that fits (254.5 bits needed) so you could add a couple of jokers in or even use a tarot deck's minor arcana (56 cards) and have room with a 256-bit wide BigInt, with a few bits to spare!

Seems like a great call!

Hey, there are only 20 paying users anyway, right? So were no vouchers used this time? And does a person vouching apply per vote or to the person (like is it a proxy vote or a renewable vote coupon?)

I got 226 bits? log2(52!) right?

I've always found this factoid pretty dubious. First off it's statistical so there are no absolute truths or facts, only tendencies.

But do this experiment for yourself, I just did a few times: take all the kings and aces out. Order them (I used Spade, Club, Heart, Diamond, Aces first then kings) and put them on top of the deck. Now do 7 riffles and look through the deck.

What I found is that the 8 chosen cards are still weighted towards the top of the deck (a little less than half moved to the bottom half, but none to the bottom 1/3rd). The suits got shuffled a little but all the spades and most clubs were still in the top 1/3rd of the deck. Most of the aces and kings stayed together in close pairs or triplets, within 5 cards of each other.

So no, 7 riffles isn't realistically good enough for fair play at actual tables. You need to mix in some cuts and pool shuffles to break up the structure of the deck and properly distribute the cards.

Fantastic album, like others have said, it was on heavy rotation in my teens, sounds track to many late night game sessions.

"Smack my Bitch Up" being a hugely popular and influential song, is also part of a long running lost media hunt for me. In the late 00's there was a youtube video made by a 3d animator as a demo reel. It was an extended anime fight between a yellow and black sentai/robot and a bunch of other robots. SMBU was the soundtrack to the video of course, lots of bits were synced to it.

That was back in the glorious pre-copyright-bot times, and since it was a demo for an unreleased project using a hugely popular song by someone who's name I've long forgotten, I can't find it! Probably scrubbed off the net by now, the creator moved on to being a cog in the Marvel movie machine.