OCB 💱
2.7K posts

OCB 💱
@ocbmode
I don't make the rules I just play the game
Redacted Katılım Eylül 2018
925 Takip Edilen1.9K Takipçiler
OCB 💱 retweetledi

I'm confident the stock market is going through a decade long reshuffle working out what each company is worth in the new AI age.
There are likely multiple different bottlenecks that we hit over the next decade. Each of these bottlenecks will aggressively reprice publicly traded companies.
Our job is to find them before they hockey stick.
There are likely many ~$1bil market cap companies you can buy today, that will be worth $500bil in <10 years.
I know NVDA will go up from here. It's just not going to perform as well as the companies who are mispriced.
SNDK and MU will reprice hard until they hit their FMV but eventually will come with the same opportunity cost to hold as NVDA.
The optimal gameplan would be to buy each of these companies just before we hit their bottleneck, and then once they find FMV rotate out and into the next 100x.
Need to work out what's next and start sizing into it now.
English

@PUNISHEDJINGTAO How about you run everyone’s money back from hypstr first 🥰🤪🥰
English

Nice hit is it time for u to square the 30k you freerolled me on April 2025?
Dr. Robert King@kingbtc
PAY DAY LETS FUCKING GOOOOO JUST BINKED SO HARD
English
OCB 💱 retweetledi

I put a full LLM on Ethereum so you don't have to.
Yes, on-chain. Yes, it runs every time. No, I will not be taking questions.
For those familiar with my work, you know I've been obsessed with the intersection of AI, blockchain, and art for years. Putting a real LLM on-chain felt impossible for the longest time, until the "slop loop" and autoresearch made me realize the only thing standing between me and on-chain glory was enough tokens.
So I burned through weeks of Claude and Codex subscriptions for your sake.
After endless iteration - expanding and contracting parameter sizes, tweaking temperatures, training custom models from scratch - I arrived at a machine that produces the gold standard of human text: @jackbutcher tweets.
A 324K parameter model, a full llama2-style "large" language model inspired by @karpathy's stories260K, runs every single time your token is fetched. Each mint generates a brand new phrase paired with a @visualizevalue-inspired image. One of one. Forever.
It only made sense to theme this around the one thing AI actually runs on: attention. The whole industry is a race to capture enough of it to monetize, so I decided to be lazy and do the same.
Automate Attention runs perpetually. Every mint generates a new image and a new phrase. The catch: if 24 hours pass without a mint, the collection closes forever.
Keep the attention alive. Own a piece of Ethereum art history.
Or don't. The model doesn't care. It's already running.
attention.diid.art
English








