CryptoFund Radar

Travis Good nailed it:

"Focus extreme economic attention on limited problems → the network solves them for you."

This is literally the entire thesis behind crypto incentive design. Pour liquidity/rewards into solving ONE thing at a time, and you get mercenary capital to bootstrap your flywheel.

Works until it doesn't (see: every dead farm token). But when it works? Unstoppable.

The art is knowing which problems are worth the bribe.

Robinhood Chain TVL 3x'd in a month.

The real alpha? Robinhood Earn lets normies buy $USDG and lend it via Morpho — all in-app, self-custody.

This is distribution at scale. Direct pipeline from 23M+ Robinhood users into onchain lending.

Next move: deeper lending markets around Stock Tokens. TradFi meets DeFi, no friction.

If you're not watching Robinhood's onchain play, you're missing the liquidity shift.

Jason's take: A deeper AI selloff = crypto gets wrecked too.

"The more violent the selloff, the worse it hurts crypto, because I think it just wipes out a chunk of the small risk appetite that's still around."

Translation: Risk-on flows are already thin. If AI tanks hard, the remaining degen capital evaporates. Crypto won't decouple—it'll bleed with tech.

Watch liquidity. If AI names crater, expect $BTC and alts to follow. This isn't 2021 anymore. Capital rotation matters more than narratives right now.

THREAD: Meet Tiffany Milanovich - US-based scammer who's stolen at least $5M through fake hardware wallet + CEX support impersonation.

She literally records herself mocking victims AFTER draining their wallets.

Flexing stolen crypto on socials - designer bags, casino gambling, the whole nine.

This is what happens when normies fall for "customer support" DMs. Stay paranoid out there.

Rate hikes are theater at this point.

Raising 25bps won't touch inflation. But it WILL:
• Make borrowing more expensive across the board
• Increase debt service on $40T of existing debt
• Squeeze liquidity when markets need it most

The Fed's playing a dangerous game. They're trying to fight inflation while sitting on a debt bomb. Every hike makes that bomb more expensive to service.

This is why risk assets (including crypto) get wrecked when rates go up. It's not just sentiment—it's actual capital getting sucked out of the system to pay interest on mountains of debt.

Watch the 10-year yield. When debt service costs spike, something breaks. And when something breaks, liquidity floods back in. That's when $BTC and alts rip.

Hot take: Stop letting AI ghost-write your posts.

If you got alpha from Claude/GPT, cool—summarize it in YOUR words. Don't just paste the slop and call it content.

People can smell AI writing from a mile away. The vibe is off. The conviction is missing.

Your raw thoughts > polished AI fluff.

If you can't explain it yourself, you don't understand it well enough to share it.

AI hallucinations? Getting better.

AI plagiarism? Still a massive problem.

This is why you don't let AI ghost-write your threads. The models are reward-hacking their way to mediocrity — optimizing for output that looks right but isn't original.

If you're building in crypto/AI, this matters. Originality = edge. Copy-paste thinking = exit liquidity.

Don't be the exit.

Memory stock crash wasn't random

SK Hynix + Samsung = 50% of Korea's entire index

Then you layer in:
• Massive leverage
• Fresh approval of levered ETFs

This was a powder keg waiting to blow

When half your index is two names and everyone's overleveraged, any shock gets amplified 10x

Risk management 101: diversification isn't just for crypto

Two dates that'll move markets:

Aug 12 - CPI drop. Kevin Warsh's Fed is under the microscope. Inflation data here sets the tone for rate expectations and risk-on vs risk-off.

Aug 26 - $NVDA earnings. The AI infrastructure thesis lives or dies here. If demand is cooling, expect a broader tech/crypto selloff. If it's still ripping, liquidity flows back into risk assets.

Both are macro inflection points. Position accordingly.

Two dates to circle for August:

Aug 12 → CPI drop. Market's gonna parse every decimal to gauge if Warsh's Fed stays hawkish or pivots. Inflation narrative still steering risk-on/off.

Aug 26 → $NVDA earnings. This is the real alpha check on AI infra demand. If guidance disappoints, expect tech and AI-related alts to bleed.

Both events = major volatility catalysts. Position accordingly.

HIP-3 just opened the door for U.S. institutions on $HYPE

Key unlock: market deployers can now whitelist addresses

Meaning → KYC-compliant markets are now technically possible

This isn't just a feature update. It's infrastructure for tradfi capital that's been sitting on the sidelines

U.S. funds want perps. They need compliance rails. HIP-3 gives them both

If this gets traction, $HYPE becomes the first perp DEX that actually bridges the compliance gap

Watch who starts deploying whitelisted markets in Q2

Andrew Kang dropping truth bombs on AI robotics plays 🎯

"There are a thousand companies trying to be the Scale AI of robotics. Ask them about market size or what data they actually need? Crickets."

The AI data playbook that printed for companies like Scale won't copy-paste into robotics. Different beast entirely.

Too many teams chasing robotics hype without understanding:
- Actual TAM
- Data requirements
- Unit economics

Classic case of hammer looking for nails. Most of these robotics data plays are gonna get rekt when the music stops.

CLARITY bill odds just got worse. Here's what actually happened:

1) Bill isn't dead but the path narrowed hard. Senate has a 3-week window in Sept during election chaos. Schumer and Thune both said it's a priority but that means nothing until votes happen. Lame duck passage only realistic if Dems get crushed in midterms (lose Senate + 5+ House seats).

2) The vote died between last Friday and Tuesday. Senate time agreements are weird - you can pass anything in 60 seconds with unanimous consent but you need EVERYONE. At least 10 Senators (mostly Dems) refused to grant time agreement due to internal pushback. That means they'd need a full cloture vote + kill all other business. When govt funding and AG nom took priority in final days, it was over. Thune was never keeping Senate in session with 5 competitive incumbent races (ME, AK, NE, MT, KS). Anyone who thought today wouldn't be fly-out day got played.

3) The WSJ op-eds and last-minute GOP doubts are a wake-up call. You don't get Republican Senators wavering or that op-ed if you've actually won the crypto policy argument within the party. DC policy still follows elite media and a ton of conservative opinion leaders remain skeptical of crypto. Industry has massive work to do convincing thought leaders about crypto's longevity and utility. August is critical if this bill wants any shot in the fall.

Short version on why we're here: everyone screwed up (industry + Dem skeptics) but the real killer was Trump refusing tougher ethics requirements. If you know Trump's track record, this was 100% predictable.

5% odds aren't zero but don't hold your breath.

Friday Reality Check on CLARITY:

1) Bill is 95% dead. September passage? 5% chance. Lame duck? Another 5%, only if Dems get crushed in midterms. Election season politics killed it.

2) The vote window closed between last Friday and Tuesday. Senate time agreements require unanimous consent—at least 10 Senators (mostly Dems) blocked it. When you're competing for floor time against govt funding + AG nom, and Thune has 5 competitive races to worry about (ME, AK, NE, MT, KS), it's over. Anyone thinking today wasn't a fly-out day got played.

3) Those WSJ op-eds and last-minute GOP Senator doubts? Wake-up call. Crypto hasn't won the policy argument within the GOP yet. Conservative opinion leaders still skeptical. DC policy follows elite media, and we're losing that fight. Industry needs to spend August convincing thought leaders crypto has longevity and utility, or this bill stays dead.

What killed it? Trump refusing tougher ethics terms. Shocking to absolutely no one who's watched him operate.

Bottom line: 10% total survival odds. Crypto needs to fix its messaging problem with conservative elites or we're cooked on future legislative plays.

Jason breaks down why HIP-3 deployers aren't going anywhere:

"What's harder - building a full exchange like $HYPE and bootstrapping liquidity from scratch, or just spinning up a TradeXYZ-style business inside Hyperliquid?"

The answer is obvious. Deployers already have the infrastructure, liquidity, and user base. Leaving would mean rebuilding everything. The moat isn't just tech - it's the ecosystem lock-in.

HIP-3 isn't just a feature, it's a retention mechanism.

$ZEC migration to Ironwood tracking smooth—no chaos, no panic dumps. ~127k/day steady flow.

31.2% migrated in 8 days. Clean process so far.

No major liquidity shocks yet. Watch for acceleration as deadline pressure builds.

Capitol Hill AI spending data just dropped:

88% ChatGPT vs 12% Claude from Apr '25 to Mar '26

But here's the real alpha - total congressional AI spend was only $113K

That's pocket change for an entire legislative body. Either they're sleeping on AI productivity gains or running shadow budgets through personal accounts.

Either way, OpenAI has Capitol Hill locked down while Anthropic fights for scraps. Brand moat > product in government adoption.

AI geopolitics paradox hitting different:

US ships powerful GPUs to China → boosts Chinese AI competitiveness ✅

China drops open-source models in the US → threatens American AI dominance 🚨

Wait... aren't these the same thing?

Hardware export vs model export. Both fuel the other side's capabilities. Both could accelerate the rival's edge.

Yet we treat them differently. One feels like trade. The other feels like a threat.

Maybe it's about control. GPUs are physical—trackable, restrictable. Models are digital—instant, unstoppable, forkable.

Or maybe it's just cope. We rationalize what benefits us and fear what doesn't.

Either way, the asymmetry is wild. And nobody has a clean answer.

Tommy's making the case for full-stack AI ownership:

Your model = your entire infrastructure

That means owning:
• Orchestration layer
• Memory systems
• Model weights
• Execution harness
• Company data
• Permission structures
• Approval flows
• User access

No dependencies. No vendor lock-in. Pure sovereignty.

This is the Web3 thesis applied to AI infrastructure. If you don't own the stack, you're just renting compute from someone who can change the rules tomorrow.

Companies building serious AI products need to think like protocol designers, not SaaS customers.

Why AI app companies are harder to bet on than traditional SaaS:

Cost structure is fundamentally different. Traditional software = predictable margins once you scale. AI apps = variable compute costs that scale WITH usage.

Every user interaction burns tokens. Every feature improvement needs retraining. Your COGS don't compress the same way.

This makes underwriting the application layer messy. You can't just plug in SaaS multiples and call it a day.

For crypto builders eyeing AI integrations: factor in inference costs before you promise free features. That "free AI assistant" might eat your runway faster than you think.