Antonis Ballis

One day the AIs are going to look back on us the same way we look at fossil skeletons on the plains of Africa. An upright ape living in dust with crude language and tools, all set for extinction.

Just finished recording a ~4 hour course on Claude Code for beginners About half is edited already - should be out on YouTube early next week! Subscribe if you haven't already and you'll be the first to know: @aniketapanjwani" target="_blank" rel="nofollow noopener">youtube.com/@aniketapanjwa…

Featured in the latest Digest: China's Rise in Global Research nber.org/digest/202604/…

Is Trump a Secret Agent for BYD? 😉open.substack.com/pub/jundu1/p/i…

Delighted that one of my latest pieces of work here at @AstonBusiness has been picked up by @Investopedia. "Bitcoin is neither a pure safe haven nor a speculative anomaly-it occupies a flexible, contingent position that may shift based on crisis characteristics, investor sentiment, and geographic context." Proud of the work done with my co-authors. Links to both the article and the full paper below. investopedia.com/guess-what-ass… sciencedirect.com/science/articl…

AI agents put in simulations of novel settings offer an alternative for applying theory, requiring minimal or no modifications, from @BenSManning and @johnjhorton nber.org/papers/w34937

HSBC and Standard Chartered securing Hong Kong’s inaugural stablecoin licenses represents a calculated institutionalization of digital credit. By operating under the HKMA’s 2025 regulatory framework, these systemic banks are effectively deploying a narrow-banking model via distributed ledger technology. ​The structural mandate for 100% backing in High-Quality Liquid Assets (HQLA) mitigates traditional maturity transformation risks, while the prohibition on interest serves as a deliberate friction point to prevent aggressive deposit disintermediation. Ultimately, the efficacy of these instruments as a medium of exchange depends on interoperability; without cross-platform integration, they risk becoming "walled gardens" with marginal utility over existing RTGS systems. bloomberg.com/news/articles/…

Άρθρο μας στην Καθημερινή της Κυριακής (01/03). «Είναι οι πληρωμές, ανόητε: Η αθέατη στρατηγική ευαλωτότητα της Ευρώπης». Ευχαριστούμε την @Kathimerini_gr για τη φιλοξενία. kathimerini.gr/economy/564101…

You can see how special the Lognormal distribution is through Maximum Entropy: • Fix mean + variance → Normal • Fix log-mean + log-variance → Lognormal • Fix log-mean → Power law So Lognormal behaves nearly Normal with small σ, but mimics Power Law tails with Large σ #RWRI

The standard approach to portfolio selection first forecasts asset returns and then plugs them into an optimizer, but this separation is deeply problematic, from Yijie Wang, Hao Gao, @camharvey, Yan Liu, and Xinyuan Tao nber.org/papers/w34861

AI valuations can be both rational and fragile: necessary to sustain the investment boom along an optimistic path, yet vulnerable to confidence reversals, from Ricardo J. Caballero nber.org/papers/w34722

This friend of mine was about to sell some of his gold holdings until he saw this.

Another powerful option in Europe’s arsenal: the $8 trillion in US bonds and equities held by European investors (nearly double the rest of the world combined), sitting largely in banks and funds. A coordinated sell-off, even partial, could flood the market, crash Treasury prices, drive yields sharply higher, weaken the dollar significantly, and amplify the pain of any tariffs far beyond trade alone. It’s economic MAD, but it underscores who’s really funding America’s deficits if push comes to shove over Greenland.

Europe has five options for responding to Trump’s Greenland threats. None of them look good theconversation.com/europe-has-fiv… via @ConversationUK @AstonPress

Check out the latest edition of my @FT newsletter. (1/2) Tech dominance is about innovation, and diffusion. That's why China may have the edge in the AI race: as.ft.com/r/869f07ba-83c…

New in @AEAjournals AER: In global tradable shortages (e.g., COVID or 1970s), optimal policy tolerates higher inflation to boost supply & employment, but non-cooperative CBs risk “competitive appreciations,” exporting slump abroad. COVID dodged it via strong jobs; 1980s didn’t (Plaza Accord?). Future trade wars beware. Link: aeaweb.org/articles?id=10…

Silver just hit a new all-time high, up about 40% in the past month and nearly 200% from the 2025 lows. When faith in reserve currencies erodes fast enough for even central banks to quietly rotate into hard assets, the market speaks louder than any press release. Precious metals aren’t speculation here; they’re the oldest form of monetary insurance against policy uncertainty and balance-sheet expansion. Those who still measure wealth primarily in fiat units are discovering in real time what convexity in real assets feels like.

Full paper here: arxiv.org/abs/2510.27688

This looks like a major breakthrough in AI efficiency. Continuous Autoregressive Language Models (CALM) replace next-token prediction with next-vector prediction, compressing multiple tokens into one continuous vector and cutting generation steps dramatically.