Time To Buy The Dip?

Summary

• Avi’s contrarian core call: sell BTC to buy beaten-down alts. With Bitcoin dominance at 61%, ETH/BTC at 0.028, and “every single alt down 90%,” he argues the moment everyone was buying (every alt up 300-500% on CoinGecko’s monthly gainers) was the wrong entry, and this is the right one: “I’d be shorting BTC, buying some alts” — specifically pairing longs in fundamentally strong names like TAO against a BTC short.
• Jonah’s counter-discipline: “never catch a falling knife in crypto.” Unlike commodities, there’s less supply curtailment at low prices — “tokens still get minted with reckless abandon” — and demand doesn’t necessarily pick up when prices are crashing, so an 80% drawdown can be followed by another 80%. His rule: wait for a token to bottom and start rallying before “dip buying.”
• Avi’s rebuttal is the episode’s sharpest trading point: waiting for the market costs you the premium. Litecoin held as an ETF trade fell 3% with Bitcoin but “Bitcoin bounces one and Litecoin bounces 10” — LTC ended up 30% on the month while BTC fell, and by the time the tape looks safe the idiosyncratic premium is already priced. He now thinks the LTC trade is played out, its premium “juiced.”
• AI coins as cautionary tale: the leader gets disrupted in weeks. Virtuals, “the clear leader like two months ago,” hit $1 and stopped bouncing even when BTC popped 4%, while ARC ripped 40% off the lows on migrating mindshare. Avi took losses (he bought the framework’s token at the highs and held AI16Z too long); the lesson is that “the most dangerous thing that you can do as a trader is to marry your bags” — sell the moment mindshare, not product, is disrupted.
• Solana is the tradeable survivor of the memecoin circus. Jonah is dip-buying — “I nibbled in the 160s and I’m nibbling still at 170” — arguing SOL “survived the meme coin frenzy without getting bowled over even once,” versus 2022-23 when it was “bowled over by a walking app.” Avi’s one worry: a $1.6B unlock, largely unhedged with a ~$60 cost basis, after which he’s bullish; value accrues to L1s that matter — Bitcoin and Solana, “maybe ETH has a play.”
• Avi’s structural trade: short a basket of L2s (Arbitrum, Optimism) against long high-throughput L1s. The appchain thesis “is nonsense” because seamless interoperability was never solved, and “nobody cares about decentralization anymore” — the attack that once capped high-throughput chains is gone (nobody attacks Hyperliquid for centralization). Cosmos and Polkadot are done.
• Both expect the reckoning later: “we haven’t had a good chain blow up since Luna… I think we’re overdue.” Jonah’s theory: decentralization becomes a topic again only after big centralized ecosystems accumulate institutional TVL and someone rugs for “hundreds of millions of dollars and a nice villa.”
• OpenAI’s deep research commoditizes the analyst. Avi got a bank-quality report on the Chinese stock market “in eight minutes”; Jonah’s framing — “the pendulum is shifting from analyzers to doers” — means the edge moves to risk-takers who act, and the two are building “deep research lite for crypto trading.”

Deep dive

1. Everything except Bitcoin looks dead — which is exactly why Avi wants to sell Bitcoin

• Avi’s read of the wreckage: alts have bled for two months and got hit hardest this month while BTC went sideways — Bitcoin dominance 61%, ETH/BTC 0.028, “every single alt down 90%.” Three or four months ago every alt was up 300-500% and the lowest monthly gainer on CoinGecko was +100% — “that’s normally the wrong time to go buy.” Now is the mirror image: “I would be selling BTC to go buy alts here. I actually think that’s probably contrarian right now.”
• The green shoots he’s watching: pockets of outperformance among projects actually shipping — TAO launched a product people are trying to use and is back to its month-ago price even though Bitcoin fell from 102k to 96k over the same stretch.
• Jonah’s pushback — worth keeping: “never catch a falling knife in crypto, especially on a shitcoin,” because even down 80% “it can still sell off another 80%.” Crypto has less supply curtailment at low prices (“tokens still get minted with reckless abandon”) and demand doesn’t necessarily pick up when prices crash in a reflexive, fear-and-greed asset. He got burned trading crypto like commodities early on; now he waits for a bottom plus a rally before “dip buying.”

2. Waiting for the market to turn means missing the premium

• Avi’s rebuttal, via Litecoin: people deferred the ETF trade because “Litecoin is correlated to the market, and if I think the market’s going lower I get to buy Litecoin lower.” What actually happened: “Bitcoin trades down 3% and Litecoin trades down 3%, and then Bitcoin bounces one and Litecoin bounces 10” — LTC up 30% on the month against a down Bitcoin. By the time the market looks healthy, the coin-specific premium is already baked in.
• The honest follow-through when Jonah presses — do you really sell BTC to buy Litecoin at multi-year highs? — is a concession: “the Litecoin trade is kind of played out… a lot of the premium in that trade has been juiced.” The live version of the idea is AI coins long versus a BTC short: “I know that sounds nuts” after six months when it truly was nuts, but the sector is beaten down enough that the pair now works.

3. The AI-coin round trip: leaders get disrupted in two months, so never marry a bag

• The call they got right without monetizing: AI-coin market cap could grow while its composition changed dramatically. Virtuals, “the clear leader like two months ago,” did “so horrifically bad” it hit $1 and didn’t move when Bitcoin bounced 4%; ARC ripped 40% off the lows because “that’s where the mindshare is now.” Avi’s warning: whatever coin you’ve chosen “can be disrupted in a week.”
• Avi’s confession: building their own agent revealed the no-code Virtuals platform “sucks” — its content got “two views per tweet” — so they moved to the Eliza framework (“I bought that on the highs, mea culpa”), which also became obsolete, then to a custom stack. He should have sold Virtuals and AI16Z on every pop as each layer failed; instead he held, and is now selling most of it: “it’s good to just take some losses.”
• Why holding feels rational but isn’t: loss aversion (“you’ve lost money and you want to make it back”), crypto-Twitter tribalism that builds “psychological bubbles of groupthink” around bag holders, and teams that fight mindshare decay with roadmap news, buybacks, and tokenomics changes designed “to psych you out of selling when you should be selling.” Avi’s line: “the most dangerous thing that you can do as a trader is to marry your bags.” Avi adds that the real cost isn’t the loss but the forward opportunities the trapped capital can’t capture.
• Avi’s structural point on why disruption is this fast: what gets disrupted isn’t the product but “the probability of future mindshare” — and mindshare is “fluid, transient, evanescent” in a way business viability isn’t. Avi’s mechanism: everything is open source, so “the four-minute mile is constantly being broken” — any innovation is instantly visible, copiable, and improvable, which is why founders are incentivized to build L1s (harder to disrupt, higher valuations) rather than applications.

4. TRUMP and Milei coins: not grift — but a catastrophic missed opportunity

• Jonah’s report from outside crypto: non-crypto Republicans are asking him “what is this scam that he’s launched?”, while Avi says Milei’s token “has turned into a category five hurricane” in Argentina. Avi’s diagnosis is structural, not moral: leaders with no crypto knowledge rely on “trusted advisers,” and “you put a hundred people in that situation, 99 of them are taking some money for themselves” — he calls them “pretty egregious scammers.”
• The counterfactual Avi wishes had happened: imagine Elon open-sourcing government spending data and launching a coin that earns 1% of all dollars saved — “he’s created a mass army of people incentivized to dig through government data.” Jonah calls it brilliant; his own lament is that politicians “sunk to the level of meming” instead of using blockchain’s actual merit — trustless, transparent databases.
• Jonah refuses the grift framing entirely: memecoins are memes, and buying WIF at the highs because “you were told that the dog would go on the sphere and the dog did not go on the sphere” is being lied to by anons, not by a CFO with fiduciary duties. His casino analogy: doubling down on 19 and blaming the dealer. “You’re buying Trump tokens at 70 billion hoping for what — 700 billion?”
• Avi’s darker take: the episode “humanizes these politicians” — the president doing something “that dumb and that grifty” proves they’re no different from anyone slipping the parking attendant $20 to get their car first. “That’s just how society works.”

5. Solana comes out of the circus battle-tested — buy the dip, respect the unlock

• Jonah inverts the consensus that the memecoin frenzy damaged Solana: builders will note it “survived the meme coin frenzy without getting bowled over even once — a far cry from the days when Solana got bowled over by a walking app back in 2022-23.” With the fat-protocol thesis intact until regulation lets application revenue reach token holders, value accrues to the L1s that matter: Bitcoin and Solana, “maybe ETH has a play in the future.” He’s acting on it: “I nibbled in the 160s and I’m nibbling still at 170.”
• Avi’s one caveat before agreeing: “1.6 billion coming in terms of unlock,” much of it unhedged with buyers in around $60 — “once that supply is cleared, I think it’s backed up.” On blame: “it’s not Solana’s fault that they were such an amazing chain… if Solana sucked, the grift would have just happened on a different chain.” Grift is a feature of every new technology — early PayPal was overrun by fraudsters (read Zero to One).

6. Creator coins are the next iteration — because coins are no longer platform-locked

• Avi’s thesis, disclosed shill included: they’re turning the THX token into a “creator coin” tracking their careers plus an AI agent. Friend.tech and BitClout failed because “you had to be on a specific platform in order for the value to be assigned to you”; now a coin launches anywhere, requires no new user behavior, and bundles community value. “That’s coming… and I’m trying to invest in it right now.”
• Jonah’s extension: imagine Meta letting any creator launch a decentralized token, tradable on Binance, that survives even if Instagram dies — “I think that would blow up.” His portfolio conclusion from the AI parabola: better to hunt early-stage convexity there than “buying the fastest horse in a leading category with a multi-billion dollar valuation and watching it trade down 90% on absolute vapor.” Both repeat the disclaimer: THX trades at $3 million and “definitely don’t buy this coin if you want to make money.”

7. Deep research kills the analyst class — good news for doers

• Avi’s specimen: he asked ChatGPT’s deep research to map the Chinese stock market — top companies, industries, revenues, balance sheets — and “it produced a comprehensive report that I would have expected to get from a bank, or McKinsey, in eight minutes.” Avi, a subscriber using it to research a California business: “insanely good.”
• Jonah’s macro frame: “the pendulum is shifting from analyzers to doers.” Twenty years ago the prestige careers were McKinsey/Bain/BCG — thinkers advising doers; now anyone gets an elaborate, well-researched vision in eight minutes, commoditizing “the analysis paralysis that thought middlemen have profited from for decades.” Crypto traders — risk-takers by definition — are the beneficiaries.
• The hedges, kept as hedged: Jonah notes it “hallucinates sometimes in pretty significant ways”; Avi calls the hallucination worry “a bit overblown — not because it doesn’t happen” but because you always had to cross-reference sources anyway. They’re building “deep research lite for crypto trading,” with a bot update shipping soon.

8. Appchains are dead, decentralization is nobody’s religion — until the next Luna

• Avi’s demolition: “the appchain thing is nonsense” — nobody solved seamless integration between chains. The deeper shift: in 2017 people got mad when projects weren’t decentralized; today they get mad about community token allocation. “Nobody cares about decentralization anymore” — otherwise they’d attack Hyperliquid, “which is not decentralized at all in any meaningful way.” That kills the appchain thesis (which existed to scale while staying plausibly decentralized) and makes him “super bullish on high-throughput L1s.”
• The trade, stated flat: Cosmos is done, Polkadot’s done, “I think the L2s are done — arbitrum and optimism… I would short a basket of those versus long a basket of high-throughput L1s.” Even Aptos is up 15% today — which Avi credits to the reply guy posting “Aptos is looking good here” under every tweet; Avi, having built a reply bot himself, estimates the guy is either never sleeping or paying Twitter thousands a month.
• Jonah’s synthesis on when decentralization comes back: it matters once serious TVL sits on centralized chains — institutions won’t LP size “on a chain that’s centralized,” and he expects the earlier hundred-people-in-a-room test to lead to eventual rugs “on centralized ecosystems… and then decentralization will become a topic.” Avi’s sign-off: “we haven’t had a good chain blow up since Luna.” Jonah: “I think we’re overdue.”