🔧 Clankers win

AI divides us in more ways than one

The world right now is unwieldy. Unhinged. And sure, maybe it’s true that actually we’ve been more divided in prior decades.

But whatever thinkpiece, statistic or historical factoid one might pull to prove that we’ve indeed had it much worse than this, they all have one thing in common: That was the before times. 

Before AI.

😼 Schrödinger's chatbot

Here’s to one more modern quandary destined to divide us further: Do you believe the AI hype? Is AI really smarter than humans, or are LLM chatbots powered by nothing more than glorified autocomplete?

We’re already at odds over this. To the non-believers, LLM-optimists — venture capitalists, thought leaders, and influencer founders already nestled inside the bullish AI slipstream — are naive clanker-lovers that miss the whole point of being human.

To the believers, the juice is worth the squeeze: Whatever AI is right now is nothing compared to what it will be in the not-too-distant future. Luckily, this is quite a comfortable position to take. We can never know if they are right or wrong until we actually get there.

Will AI truly be “vastly smarter than people” within a decade, triggering an “astonishing rate of technological change in the world,” as Sam Altman put it?

We already know it literally pays to be an optimist. Which makes being an AI bear in crypto — a space all about markets, financialization and speculation — totally untenable. 

Perhaps AI bears and bulls can meet somewhere in the middle by agreeing that chatbots are indeed “better” than humans at very specific tasks. They excel at games like Chess and Go and can summarize long or painful academic and legal texts faster and more accurately than humans. 

Clankers also smoke skinbags when it comes to writing boilerplate code and basic software engineering tasks, at least in terms of speed and stamina. In the words of former OpenAI employee Vahid Kazemi last year, AI is not yet "better than any human at any task,” but rather, "better than most humans at most tasks.”

Alpha Arena is currently testing the limits of the current batch of LLMs when it comes to trading crypto perps.

After 10 days, DeepSeek and Qwen3 — both developed by operations in China — are by far the best performers, now up $11,600 and $7,100 apiece. That’s 115% and 71% profit in a week and a half.

Meanwhile, ChatGPT and Gemini are down 60% on their initial $10,000 capital. 

How does your portfolio compare? If you haven’t managed to double your money with margin on Hyperliquid, maybe it could be said these LLMs are better than you at finding an edge in crypto markets. 

Or it’s entirely noise. Remember that a chimp smoked Wall Street simply by throwing darts at a dartboard in the middle of the dot-com boom.

Wherever you find yourself on either side of the AI-isle — believer or non-believer — one thing is certain: It was humans who cooked up the DAT structure for hoovering up supply to pump prices.

And it’s humans at those DATs who are now selling their coins to buy back their own stocks to avoid death spirals. Make of that what you will.

  • HYPE and PUMP lead the top-100 coins by market cap for weekly gains, both up by around 33%.

  • F2Pool co-founder Chun Wang says the recently-proposed BIP-444 for Bitcoin, pitched as a temporary soft fork amid the ongoing spam filtering debate, is a bad idea.

  • ICYMI: ETFs for SOL, LTC and HBAR are set to start trading in the US today.

Bring your friends, reap the rewards 🎁

Know someone who loves both crypto and a good story? Send them our way with the Empire referral program, and win some rewards while you’re at it:

  • 👊 3 referrals: Grab our limited-edition Empire sticker