Crypto traders' self-report: friends are all leaving, what is left to do in this industry?

Author: donn (@tzedonn)

Translation: Deep Tide TechFlow

Deep Tide Guide: There are only 12 projects that earned over $50 million last year, as DeFi is attacked by AI and turns into a fire pit, while on-chain gold miners are still stuck on a battlefield that has already ended—every crypto professional should ask themselves the same question: What else can I do here?

Many of my friends have already left or are considering leaving the crypto industry, so I want to share some broader thoughts on the market and discuss what remains to be done in cryptocurrency.

The core issue is that cryptocurrency is in trouble in three key areas: (i) lack of innovation, with nothing new emerging in the past 2-3 years; (ii) advances in quantum technology threaten Bitcoin’s survival by 2029; (iii) large language models like Claude Mythos increase attack frequency, making DeFi’s risk/reward less attractive.

This raises a question: What else can be done in cryptocurrency?

VCs and liquidity token investors

Due to this innovation drought, the VC industry has been quite quiet, especially in token trading. Every crypto VC will tell you how boring it is unless they’re doing larger Series B+ funding rounds or backing stablecoin payment startups.

A few excellent VCs I’ve interacted with are investing in reverse verticals, such as quantum startups (like Project Eleven, Oratomic) or novel ideas (like Shift Foundation, PostFiat, Ambient).

I think this is normal because we’ve basically figured out what works and what infrastructure is needed, so exciting new things are becoming rarer. We are now in the adoption phase of payments and remittances. The endgame and institutions are already here.

Similarly, for those working in crypto, only roles at stablecoin fintech companies (like Circle, OpenFX, Tempo, Arc, Plasma), trading platforms (like Polymarket, Kalshi, Hyperliquid), or innovative reverse startups (as mentioned above) make sense. Working at an L1 foundation is a decent-paying dead-end job, but it offers no long-term benefit.

The lack of “cool new stuff” in VC means fewer high-quality tokens hitting the market, and VC capital flowing into liquidity markets will also decrease.

Therefore, long-term liquidity token investors, who evaluate tokens based on growth, fundamentals, and value accumulation, now probably have fewer than 10 quality targets, and this number doesn’t seem likely to grow in the short term.

Only 12 token projects generate over $50 million in annual revenue. Of these, only three have a value accumulation score ≥7 (HYPE, PUMP, JUP). Even if you evaluate teams based on “growth” and improving token value accumulation, you might add 5-10 more tokens to the list below (e.g., MORPHO, SYRUP).

The market cap of OTHERS has also fallen from about $450 billion to around $180 billion, while the stock market is experiencing speculative frenzy in high-bandwidth memory, photonics, quantum, peptides, and other fields.

Subjective and systematic traders

You might say that crypto mainly involves narrative trading and momentum trading, so it’s suitable for long/short, narrative trades, or catalyst/news-based subjective traders. That’s my domain and what I’m most familiar with.

In the depths of a bear market, trading catalysts has been more profitable, even though it requires staying alert and reacting quickly. There are many advantages and good trading opportunities here.

That said, there have still been quite a few event-driven trades over the past three months, such as…

Shorting TAO when leaving Templar subnet, dropping from $330 to $260 within 5 hours (April 9) … Going long on TAO when Chamath and Jason Cal called it (success varies)

No trades here, but interestingly, WLD didn’t even rise on Tinder and Zoom partnership announcements (down 8% in two days)

Shorting AAVE during the KelpDAO hack, which occurred around 1730 UTC but was first reported on Twitter about an hour later (~1830 UTC) (April 19)

Shorting AAVE when Marc Zeller left ACI (March 3)

Shorting TRUMP after the dinner announcement (March 12)

Shorting ACX after the token conversion announcement, during its spike to the “conversion price” (March 11)

Shorting DRIFT during exploit, down 40% within an hour (April 1)

Shorting RESOLV during exploit, down about 10% (March 22)

Long ALGO during Google quantum breakthrough (March 31)

Long LDO after buyback announcement, up about 17% over roughly 5 days (March 27)

Long/short plays on pump-and-dump schemes like RAVE, SIREN, STO, PIPPIN, POWER (I prefer not to participate)

Despite this, open contracts since October 10 have decreased by about 60%, and reactions to news are often minimal. You need to selectively choose which news to trade and whether others care about it, as retail trading interest is minimal (usually the slowest to react to headlines), so it’s mainly PvP with other news traders.

I increasingly see subjective traders spending more time on prediction markets and trading stocks/commodities, and Hyperliquid makes this transition much easier.

As for systematic traders and basis traders: with declining trading volume and funding rates, traditional strategies are becoming less profitable. To stay engaged, they trade HIP-3 markets, arbitrage prediction markets, trade Pendle PT/Boros, or arbitrage new perpetual contracts on decentralized exchanges (with limited liquidity).

Yield farmers

Increasing hacks in DeFi have also caused yield farmers (or TVL traders on the institutional side) to go dormant or exit entirely this year, as the last good trades were Plasma and USDai. FlyingTulip might work in a bull cycle, but it didn’t attract much attention at launch.

Typically, “yield trading” makes sense because you can sell rewards for governance tokens (as I emphasize here), but this assumes someone is willing to buy from you. Without liquidity investors/subjective traders adding fuel, the risk/reward becomes hard to justify.

This was true in the past (OHM), now (XPL, ENA), and may be in the future (USDai’s CHIP). The on-chain DeFi yield threshold used to be 15-25% annualized, considering treasury yields around 0% and hacking probabilities of about 10-15%.

OHM was worth it not because of lower risk, but because of much higher returns.

Now, the threshold might be close to 50-60% annualized, as hacking incidents increase (DeFi hacked for $795 million in the first four months of 2026) and Claude Mythos could lead to more hacks and quantum risks. Without anyone buying the released tokens, it’s hard to achieve a reasonable risk/reward.

Most rational farmers have almost completely shifted off-chain, as even traditional fixed-income yields like STRC at 11.5% can offer better risk-adjusted returns (15-20%).

See: Rami poker, CBB, Sisyphus, delucinator, misaka

On-chain gold miners

Last but not least… the infamous on-chain gold miners: those who buy at a $1 million market cap and sell at $100 million. I believe they will still exist because trenches remain the only place capable of pulling out 100x.

On-chain gold miners also feel like WWII soldiers trapped in a cave, unaware that the gold rush era has basically ended.

Not to mention, when our dear President and First Lady launched their coin, meme coins peaked. We’re nearing the end of the euthanasia roller coaster—things no longer rise like before, and there are too many value extractors (“FNF groups,” “LA e-cigarette groups,” continuous carpet factories, and extraction trading fees).

However, I still think this niche won’t completely disappear, as a glimmer of hope keeps people in the trenches. Over the past few months, we’ve seen…

$GAS: GasTown went from $100k market cap to $60 million in 3 days, then back to $1 million in 3 days, now at $500k. (January 15)

$RALPH: RalphWiggum rose from $500K to $55 million in 2 weeks, then dropped to $3 million within 12 hours when developers abandoned it (-93%), now trading at a $50K market cap. (January 21)

$PENGUIN: Nietzchean Penguin hit $170 million in 3 days (January 24), but now trades at $3 million.

$MOLT reached $120 million in one day (January 31), but now trades at $1 million.

$WHITEWHALE hit $200 million (January 10), but now trades at $7 million.

$ASTEROID reached $200 million yesterday (April 19) after Elon Musk tweeted it might become SpaceX’s mascot. A few days later, it could trend toward zero.

While these sometimes do surge, your chance of hitting them is probably <10%, with the maximum potential maybe netting a 10x return (from $10 million to $100 million), because things no longer rise past $100 million. Now, you only have a few hours after the peak to sell, as prices can drop >90% within hours.

So… what am I doing?

I’m maintaining my arbitrage strategy on Polymarket, generating about 15% annualized returns, with a maximum capacity of around $250K. Fully deploying the capital only earns about $3,500 per month. Since Polymarket introduced trading fees, arbitrage opportunities have diminished, and with recent npm package poisoning, I increasingly feel the risk/reward isn’t there anymore (especially after airdrops). My current plan is to kill it after the airdrop.

I continue trading crypto, but not as actively as before. I’ve had ideas about starting an AI hedge fund, but my explorations led me to conclude that AI isn’t yet creative enough for subjective trading (like idea generation). However, if you give it a task or methodology you’ve used before and require some logical reasoning, it’s excellent.

So I’m spending time automating certain “hedge fund analyst” tasks, like automating on-chain reconnaissance I used to do for internal wallets on Polymarket. It’s quite process-driven but requires some logical thinking. That’s a perfect task for Claude.

I’m also increasingly researching AI model fine-tuning, especially from the perspective of crypto and financial data. I’ve been reading beyond crypto, finding topics like AI stacks (social impact and entire stacks of stocks), physical AI (world action models, visual-language models, data issues), and “AI integration” ideas (PE integration but with AI) intellectually stimulating, but I haven’t yet found a sufficiently interesting problem to dedicate my life to solving.

If you have any interesting topics to discuss, feel free to reach out!