We’re bullish on tokenized stocks because moving equities onchain doesn’t just mirror its value, it unlocks entirely new ways to deploy the capital beneath them. At @paboxoxyz, that begins with borrowing against stock tokens: tap liquidity without giving up exposure to the underlying stock, with real interest and liquidation risk, plus off‑market mechanics designed from zero specifically for tokenized equities. It’s awesome and genuinely fun, but we know it’s not enough. So we’re pushing into the next edge: using liquidity provider (LP) positions as collateral via @Uniswap v4 Hooks. The setup is hard to ignore. As of writing, there’s already $161M+ in LP positions tied to tokenized stocks, yet composability is still missing. An LP market could make this possible. A liquidity provider could borrow against their position without pulling it from the pool, keeping their shot at earning trading fees while unlocking capital at the same time. But we’re not rushing it. Doing this right takes more than slapping a price tag on an LP token: the asset mix shifts, fees can underperform, and stock‑market closures make valuation messy, exactly the problems we’re working through. This is the exploration we’re most excited about. Our thesis is straightforward: tokenized stocks aren’t just “single stock” assets. Whatever form they evolve into becomes far more powerful once it can plug into trading, liquidity provision, and credit. We’re building in that direction. -- P.S.: Paboxo doesn’t have a token yet, and we’ll be partnering with @Moshdottrade to launch one in the future.

Rebuilt @RuneCrypto_’s table and added the % breakdown. It’s pretty clear: despite the insane volume of launches on Pumpfun, only a microscopic slice actually holds a $100k floor. - @LaunchOnSF → 0.04% - @virtuals_io → 0.03% - @longdotxyz → 0.02% - @flapdotsh → 0.01% - @ponsdotfamily → 0.01% So you can give me all the bullish theses on $PUMP, but I don’t think the economy will trickle down to the projects launching on it.

What if you could hold a stock only while the market is closed, without buying and selling every day? That's the idea behind Isolated Exposure Tokens (IETs), proposed by @0xbeans and The @standard_rsv. Deposit one tokenized share and receive: • OVERNIGHT: exposure from the close to the next open. • SESSION: exposure from the open to the close. Hold both and redeem for one share. Sell one to keep only the exposure you want. Each series has an expiry; exiting one side earlier requires a buyer. A simple example: • A $100 share starts with $50 allocated to each token. • It opens at $110: OVERNIGHT rises to $60. • It closes at $105: SESSION falls to $45. • Together, they're worth $105 at the accounting level. Why it's interesting and why we're at @paboxoxyz also loves it as it align with our off-market lending features. • Could support overnight-only vaults and targeted hedges. • Splits exposure without requiring the issuer to trade the underlying every day. • Gives DeFi another building block, similar in structure to Pendle's split claims. But something that still needs some works and detailed explanations: • Each token starts with roughly 2× exposure at its initial accounting value. • Balances accumulate dollar moves, so they don't reproduce the compounded returns. • Expected overnight gains can already be priced into the token. • Overnight exposure existed before through NightShares ETFs. The vault can't pay $110 if the losing holder owes nothing. Caps, early termination or extra capital could help, but the supplied paper doesn't specify a rule. Promising concept. It needs clear loss rules, reliable price feeds and affordable liquidity before the historical overnight anomaly becomes a new primitive.

Damn, just wake up, wtf we just go NYSE-HYPE-Druckenmiller and now hacks across CEXs.

Legendary billionaire is bullish, and you sold your bag? NGMI.

The easiest way to compare Bittensor and Pearl is to ask: what gets you paid? With @opentensor / $TAO, you compete to do a job well. Each subnet sets its own tasks, such as running AI models or providing data. Miners do the work, validators score it, and those scores help determine rewards. The catch is making sure a high score means something customers actually find useful. With @prlnet / $PRL, compatible AI calculations can also help secure a blockchain and earn mining rewards. A GPU operator could get paid for serving AI while earning PRL from the same computation. That extra income could help lower the customer’s bill. But the subsidy depends on what those rewards are worth. And proving that computation happened doesn’t prove anyone wanted to buy its output. For token holders, there’s another difference: PRL is Pearl’s mining asset, while Bittensor has TAO plus separate subnet tokens called alpha. Staking into a subnet adds exposure to that subnet’s token price and liquidity. Bittensor’s appeal is competition across many services. Pearl’s is getting more economic use from the same computation. For both, I’d watch the same thing: PMFs Are customers coming back and paying for the service? More details below (NFA. DYOR) ⤵

No matter the market sentiment we keep shipping. We have added $SPCX and $TSLA to collateral assets. More to come, 👀

If you need a reminder, we’ve seen this situation before, back when Elon was trying to buy X (then still called Twitter). During the Twitter acquisition, Elon reportedly raised about $1B by borrowing against his $SPCX shares. Not selling. Borrowing. He gained liquidity while keeping his stock ownership. That kind of perk is typically reserved for a small circle of people and institutional investors. For everyone else, it’s a maze of bank hurdles, endless paperwork, and layers of legal sign-off. With @paboxoxyz, though, you can actually do it in just a handful of clicks (I counted 4–5, and soon we’ll make it even more seamless). And yeah, we’re dropping another market soon. Can you guess? :)

Knowing where to spend engineering time is part of growing up as a protocol. From Story to @datafdn, the focus has sharpened around AI training data: where it came from, who can use it, and who gets paid. The planned move to Sui takes that further. Sui provides execution, Walrus provides storage, and DATA can put more engineering effort into Trace. And to underline it, the Sui Foundation’s $30M pledge puts real weight behind that path. Don Wilson said compute will be the new oil in just a few years, and it will require abundant data to train. Let that sink in.

Up to 2,500 participants. One validator deciding miner rewards. That’s the trade-off behind @opentensor’s Null Consensus. A subnet gets more room to grow, while trusting a single evaluator to decide which work deserves payment. Think of a competition with more entrants and one judge. Bittensor subnets are specialist networks where miners provide work and validators score it. Normally, Yuma combines scores from multiple validators, weighted by stake. Null lets owners choose a simpler reward process, but Yuma remains the default. Here’s what changes: • One validator gets the scoring role. The registered participant with the highest stake weight. Choosing Null doesn’t automatically give that role to the subnet owner. • That validator’s valid scores set miners’ relative rewards. Null skips Yuma’s consensus and bond calculations. • Positive-staked participants still receive proportional dividends, even without permission to score. With positive stake, Null splits its epoch budget roughly equally between miner rewards and validator dividends. Owner cuts stay unchanged. Why accept that trade-off? • More capacity: With one reward mechanism, the ceiling rises from 256 participant slots to 2,500, nearly 10×. Two mechanisms share the limit, allowing 1,250. Owners must explicitly raise capacity. • Less computation: The chain processes one validator’s scores instead of combining many evaluations and updating bonds. • Easier entry: Owners can separately enable proof-of-work registration on Null or Yuma. Direct zero-tip submissions need no $TAO registration burn, upfront collateral purchase or transaction fee. Participants still spend computing resources. The catch sits right inside the design. A subnet can have thousands of participants and still concentrate reward decisions in one place. If its leading validator rewards friends or misjudges quality, Null has no competing validator consensus to correct those scores. There’s an engineering question too. The merged implementation records unresolved automated-review findings about accounting for computational load. Those concerns deserve follow-up; they don’t establish that an exploit occurred. Null could make sense where a subnet already accepts one trusted evaluator and needs room for more contributors. But filling 2,500 slots is only the beginning. The stronger case is better work, reliable operation under load, and services people will pay for. NFA. DYOR.