Summary

Uniswap’s proposed fee switch could be structurally positive for UNI, but Smith does not yet see an obvious long-term trade. Thread Guy explained that a hypothetical $1 swap fee could move from 100% to LPs toward a 95/5 or 90/10 split, finally giving the DAO revenue that could fund buybacks, burns, or staking distributions. Turning off Uniswap Labs’ separate front-end fee and exploring MEV internalization matter; burning roughly 10% of supply that was already unissued does not — “it’s just that you get to say you did it.”
ORE has turned proof-of-work mining into a one-minute, 25-square roulette game with an unusually aggressive token sink. Each round adds 1.2 ORE to the winning square, while a 1-in-625 “motherlode” compounds by roughly 0.2 ORE per missed round. ORE takes 10% of SOL wagered as revenue, then directs 90% of that revenue to buybacks and burns and 10% to stakers.
ORE’s million-dollar revenue day is real traction, but the entire machine is brutally reflexive. As ORE rises, subsidies and jackpots become more valuable, bets increase, and buybacks accelerate; Thread Guy said he thought it had fallen from roughly $600 to $300, making the game worth half as much. Smith remained extremely long and saw the next several days as the test: revenue near $800,000 after the 50% price drop, versus the prior day’s $1 million record, suggested the game had not immediately broken.
Smith’s framework is that fundamentals raise a token’s floor, but revenue allocation must balance the product and token holders. Thread Guy cited HYPE’s Assistance Fund buying back roughly 99% of revenue while holding no USDC buffer that might help prevent ADLs: “Who are you prioritizing, the users and traders or the token holders?” Narrative assets such as Bitcoin and Zcash can escape fundamentals, but most apps cannot reliably manufacture that cult-like outcome.
Smith framed Zcash as a trade with a transcendent privacy narrative and no cash-flow ceiling. Thread Guy agreed with the privacy case but raised the concern that the same tools could help hostile states; Smith admitted he had not thought through the issue and did not have a clean answer. Smith disclosed leveraged Zcash exposure on Hyperliquid, while Thread Guy said his exposure was through Coinbase.
HumidiFi’s prop-AMM model may be one of the strongest undercovered pieces of onchain market structure. Smith said it handled roughly $2.4 billion of volume the previous day — potentially more than Uniswap or any other spot DEX — while offering tight spreads and deep SOL/USDC liquidity. Thread Guy called it “onchain Citadel,” with “a massive grain of salt,” and flagged its planned token launch through Jupiter.
The lending endgame may be permissionless modular vaults, but DeFi still communicates risk disastrously. Morpho lets curators allocate deposits across strategies, whereas Aave emphasizes shared liquidity and strict asset listing; the failure of what Thread Guy recalled as xUSDT or xUSD exposed the modular model’s tail risk. Thread Guy likes Morpho’s architecture but finds its north-of-$1 billion FDV rich, while Spark is a potentially cheaper exposure.
Crypto’s current opportunity is less a broad alt season than a hunt for new pockets whose launch valuations have already reset. ORE, ZEC, Zora, football cards, and AVICI show that isolated products can catch fire even while majors and memes stagnate, but most activity has not been sticky. The longer-term ICM bet is ownership: MetaDAO’s appeal is genuine control and potential treasury recovery, while Pump.fun’s challenge is using its war chest to graduate from product-adjacent memes into ownership coins that create value for PUMP.

Deep dive

1. Uniswap is finally unifying protocol economics with UNI

Thread Guy had heard “chatter in the streets” before a Chinese news outlet appeared to break the fee-switch story several hours ahead of the forum post. The apparent advance movement in UNI’s chart was, candidly, “kind of a problem,” even though the proposal itself was welcome.
Thread Guy’s prior resistance was principled: protocols should build runway before sharing revenue on day one. “There are no perpetual motion machines”; distributing 100% of revenue while funding operations through token sales is merely money coming “in one door, out the other,” so he had supported holding the line.
The mechanism is straightforward. Where a $1 swap fee previously went entirely to LPs, a 95/5 or 90/10 split could send revenue to the DAO, which might later buy and burn UNI or distribute value to stakers. Until now, the DAO had earned “nothing — not a dollar.”
Thread Guy liked Uniswap Labs removing its separate front-end fee and was intrigued by an MEV-internalization mechanism that might offset lower LP economics, though he had not fully unpacked it. He dismissed burning roughly 10% of supply already sitting unissued in treasury: “Those tokens are unissued and uncirculating.”

2. A better UNI token does not eliminate a difficult DEX business

At the industry level, Smith viewed activation as significant: a project instrumental to four years of DeFi growth was moving beyond the regulatory rationale long used to defer token economics. The “unification” framing also sought to reduce the persistent tension between UNI holders and Uniswap Labs equity.
Thread Guy’s pushback was practical: does this turn a “pretty cursed” chart into a long-term buy? Smith’s answer was restrained. DEXs are likely low-margin businesses, he was not in the trade, and the announcement alone did not make him “super excited.”
The more interesting upside would come from Unichain reaching meaningful scale or Uniswap v4 hooks becoming a platform for complex exchanges built atop Uniswap. Those are potential catalysts rather than conclusions, leaving Smith interested enough to watch but not to chase.

3. ORE rebuilt mining as a one-minute casino loop

ORE’s lore began roughly 18 months earlier, when its transaction-heavy proof-of-work design helped make Solana painful to use during the first memecoin surge. Thread Guy described that episode as part of Solana’s “get punched in the face,” find the failure, fix it, and continue growing culture.
After several mining iterations and a quiet development period, ORE relaunched roughly two months before the conversation. Its current system replaces hashing puzzles with 25 equal-probability squares: players wager SOL across any number of them, one square wins every minute, and its bettors divide the round’s pool.
Each round also issues 1.2 ORE to the winning square. The “motherlode” has a 1-in-625 chance of hitting; otherwise it grows by roughly 0.2 ORE per round, reaching about $80,000 during the demonstration and previously driving $15,000-$20,000 rounds when it approached $200,000-$300,000.
ORE takes 10% of SOL wagered as revenue. Of that amount, 90% buys and burns ORE and 10% goes to stakers: on $100 wagered, $10 becomes revenue, $9 funds buybacks, and $1 is distributed. “They’re just hammering buybacks,” Thread Guy summarized.

4. ORE’s product loop made the data more convincing than the story

Thread Guy owned ORE in its original incarnation and “got absolutely fried,” moving from “worst token ever” to “I’m ready to get hurt again” when it revived. Playing changed his view because crypto products usually make him bearish on use; this one delivered a genuinely fast dopamine hit.
The tactical game is to bet late, after seeing where money has accumulated, and favor the least-crowded squares because every square has equal odds while payouts are shared. Thread Guy estimated that a 0.1 SOL wager could return roughly 1.5-2 SOL on an ordinary win and thousands of dollars on the motherlode.
During the live demonstration, they spread wagers across multiple squares, briefly mistook the result for a win, then realized they had lost. That failure carried the point better than a theoretical pitch: ORE is a simple, quick casino game, not a cryptographic mechanism requiring elaborate explanation.
The datasets Smith tracks include average value bet per round, unique wallets, square-selection patterns, and how many squares winners typically cover. Bots probably participate, especially as jackpots grow, and Smith viewed that as compatible with the game’s revenue and buyback dynamics.

5. ORE’s flywheel works in both directions

Thread Guy said ORE had printed its first $1 million revenue day; Smith said that made it a top-two or top-three crypto app by that measure. That was difficult to ignore, while the team’s willingness to keep iterating for 18 months distinguished it from short-lived experiments.
Yet price is embedded directly in product demand: higher ORE makes the 1.2 ORE subsidy and jackpot more valuable, attracting more SOL, which creates more revenue and burns. The reverse is equally forceful — “when the price goes down, everything gets worse.”
After Thread Guy said he thought ORE had fallen roughly 50%, from around $600 to $300, same-day revenue still looked near $800,000 against the prior record of $1 million. Smith called the next two or three days the real test: continued play at half the token price would be “hard not to be excited by.”
Smith’s disclosure was unambiguous: “I’m extremely long,” and he had bought more shortly before Frank hit the sell button. What attracted him was not the store-of-value pitch or net-deflationary rhetoric, but a fun roulette system filling the dopamine void left by memecoins’ “sophomore slump.”

6. Revenue should support products, not consume them

Smith’s broad philosophy is that fundamentals “raise the floor,” even though Bitcoin and Zcash demonstrate that assets can achieve escape velocity without cash flows. Trying to engineer that outcome is a low-probability strategy for an app founder; most projects eventually need dollars coming through the door.
Even apparent non-fundamental equity cults point toward future economics. Elon Musk’s robot narrative matters because it implies eventually making “a shit ton of money,” and Smith gave Palantir as another example. Crypto often copies the devotion while omitting the promised future earnings that support it.
Thread Guy contrasted buybacks with HYPE’s Assistance Fund, which he said buys back HYPE with about 99% of revenue and holds no USDC reserve that could serve users during events such as ADLs. Smith agreed that revenue could instead support other functions and that protocols must balance traders, users, token holders, and reinvestment.
Smith framed Zcash’s privacy ideology as a powerful narrative that is not capped by cash flows. Thread Guy raised the concern that privacy tools could also help hostile actors; Smith admitted he had no satisfying resolution. Smith disclosed leveraged exposure on Hyperliquid, while Thread Guy said his exposure was through Coinbase. Thread Guy then noted that cash already facilitates crime.

7. Market structure is advancing faster than its tokens

Smith identified swapping and lending as DeFi’s two foundational areas. On Solana, prop AMMs resemble isolated onchain market-making books: an aggregator can query individual firms, obtaining tighter quotes without using a conventional shared order book. Manifest was another new attempt at the order-book design.
Smith said HumidiFi had processed roughly $2.4 billion the previous day with strong SOL/USDC depth and fees below many centralized venues. Thread Guy’s knowingly oversized shorthand was “onchain Citadel,” offered with “a massive grain of salt”; he also noted its forthcoming token through Jupiter’s launchpad. Smith added its deliberately playful “get wet” branding.
Lending’s parallel innovation is modular vault management. Morpho lets curators allocate deposits across markets according to a stated strategy, while Aave uses a shared instance with strict listings and risk work from groups including Marc Zeller, ACI, Chaos Labs, and Llama Risk. Each model chooses a different point between control and permissionlessness.
The collapse of what Thread Guy recalled as xUSDT or xUSD, which impaired exposed vaults, showed the disclosure gap. Traditional managers provide a prospectus; a DeFi vault may offer one sentence, and even a full page may not help because “everyone clicks the vault with the highest yield” — usually the highest risk.

8. The next trade may come after launch-day price discovery fails

Smith considers modular lending the likely endgame, but Thread Guy found Morpho’s north-of-$1 billion FDV rich despite its “cracked” team and leading position. Thread Guy viewed Spark as a smaller-cap alternative; Gauntlet and Steakhouse lacked tokens, Euler had already rallied, and Maple was strong lending exposure without representing the same architecture.
Across crypto, launch valuations create the same problem. Thread Guy said DoubleZero might reinvent internet packet delivery, but if it launches around a $5 billion valuation, he would watch from the sidelines; his preferred zone is generally below $1 billion. A great product does not automatically make its first liquid price attractive.
Smith mapped the recurring pattern: leverage traders chase launch momentum, then exit when gains fade or positions turn red, producing 70%-80% cascades in assets such as Pump and Plasma. After one to three months — or roughly three years in examples such as Robinhood and Coinbase — investors may rediscover “a sick company” at a rational valuation.
Their ICM debate ended on ownership. Thread Guy argued that MetaDAO’s crypto-native launches have found more demand than Believe’s Web2-oriented launches, but that MetaDAO’s futarchy can be a “nerd-snipe.” Smith emphasized that MetaDAO can offer control and potential treasury recovery if a company goes rogue. Pump.fun’s opportunity is to pair its roughly million-dollar daily revenue floor and war chest with genuine ownership coins, then prove that cash can create value for PUMP.

How Apple Changed the Cellular Economy and SpaceX's Spectrum Play Benchmark's GP, Everett Randle on Why Mega Funds Will Not Produce Good Returns