HodlCat ∞÷21m

pleb

[Michael Saylor @saylor]

https://t.co/yeG4PgpjEq

[HodlCat ∞÷21m @alphaN1]

@elder_plinius Take care Pliny! We need you

[BitcoinTreasuries.NET @BTCtreasuries]

JUST IN: Strategy’s Michael Saylor just said, "#Bitcoin is going to rally. I think that will catalyze a rally in all the BTC treasury companies and cause a rally in demand for digital credit."

"We're setting up a massive supply shock." 🔥

[HodlCat ∞÷21m @alphaN1]

@m_0_r_g_a_n_ Congrats! Checked X only to see how its going!

[HodlCat ∞÷21m @alphaN1]

@rMe77340969 @peterkovari @panyiszabolcs I respect your opinion. Also as Ukranian im not sure its KGB trained mafia open-air concentration camp. It’s just open air concentration camp. Maybe KGB was involved idk.

Also Ukraine needs money to fight and a weak Putin. Orban stands in the way of both.

[HodlCat ∞÷21m @alphaN1]

@m_0_r_g_a_n_ This is unacceptable!

[HodlCat ∞÷21m @alphaN1]

@StratcomCentre @1basemoney What an douce. Who cares of how other counties call themselves. Souverenity is built on power not on names and propaganda.

[HodlCat ∞÷21m @alphaN1]

@Peter_Quadrel AI

[HodlCat ∞÷21m @alphaN1]

@AiwithTyler UGC

[Michael Saylor @saylor]

Stretch the Orange Dots.

[The Bitcoin Historian @pete_rizzo_]

BREAKING: MICHAEL SAYLOR JUST ANNOUNCED STRATEGY IS GOING TO BUY MORE #BITCOIN 🚀🚀🚀

[Michael Saylor @saylor]

@BorisJohnson Bitcoin is not a Ponzi scheme. A Ponzi requires a central operator promising returns and paying early investors with funds from later ones. Bitcoin has no issuer, no promoter, and no guaranteed return—just an open, decentralized monetary network driven by code and market demand.

[HodlCat ∞÷21m @alphaN1]

@_alisawu ugc

[HodlCat ∞÷21m @alphaN1]

@HermesLux >400m

[HodlCat ∞÷21m @alphaN1]

@m_0_r_g_a_n_ someone decided to be with the god for a lifetime

[HodlCat ∞÷21m @alphaN1]

@wannercashcow PROTECT

[Andrej Karpathy @karpathy]

nanochat now trains GPT-2 capability model in just 2 hours on a single 8XH100 node (down from ~3 hours 1 month ago). Getting a lot closer to ~interactive! A bunch of tuning and features (fp8) went in but the biggest difference was a switch of the dataset from FineWeb-edu to NVIDIA ClimbMix (nice work NVIDIA!). I had tried Olmo, FineWeb, DCLM which all led to regressions, ClimbMix worked really well out of the box (to the point that I am slightly suspicious about about goodharting, though reading the paper it seems ~ok).

In other news, after trying a few approaches for how to set things up, I now have AI Agents iterating on nanochat automatically, so I'll just leave this running for a while, go relax a bit and enjoy the feeling of post-agi :). Visualized here as an example: 110 changes made over the last ~12 hours, bringing the validation loss so far from 0.862415 down to 0.858039 for a d12 model, at no cost to wall clock time. The agent works on a feature branch, tries out ideas, merges them when they work and iterates. Amusingly, over the last ~2 weeks I almost feel like I've iterated more on the "meta-setup" where I optimize and tune the agent flows even more than the nanochat repo directly.

[HodlCat ∞÷21m @alphaN1]

@TheRealPlanC I'm permanent bull, but such posts makes me believe we can go lower. Usually bear ends when almost noone is ready

[Harshil Tomar @Hartdrawss]

https://t.co/IGLH7U9new