Thibault Schrepel
banner
profschrepel.bsky.social
Thibault Schrepel
@profschrepel.bsky.social

Associate Prof VU Amsterdam • Faculty Affiliate Stanford • Into Running 🏃🏻
#antitrust #AI #complexityscience #digitalmarkets #blockchain
📕 www.thibaultschrepel.com
📻 https://podcasters.spotify.com/pod/show/scalingtheory

Economics 42%
Computer science 22%
Pinned
My rule for interactions on BlueSky (and elsewhere) is simple: I always assume it’s my interlocutor’s birthday. This means my messages are sent with kindness and compassion.

Reposted by Thibault Schrepel

Antitrust is no longer a domestic game. Daniel Crane (University of Michigan Law School) maps how Great Powers use antitrust law as a geopolitical lever across tech, culture, finance, and even wartime supply chains.
Read here 👇 www.networklawreview.org/crane-great-...

Super interesting. Do you have more exmples of such national initiatives that are DMA-like?

If you are in town, I will head out for a morning run 🏃 on Monday and Tuesday. Feel free to join; my DMs are open.
More information here: www.concorrencia.pt/en/events/ne...
New edge issues in digital competition | Autoridade da Concorrência
www.concorrencia.pt

These tools give agencies a concrete policy direction by revealing mechanisms that stay hidden in static analysis. I plan a few other fun activities with the Authority during my stay.
New edge issues in digital competition | Autoridade da Concorrência
www.concorrencia.pt

🇵🇹 I will speak at the Portuguese Competition Authority’s next Monday in Lisbon. In my talk, I will show how computational tools (and thinking) expose dynamics in emerging antitrust issues linked to GenAI, cybersecurity, plus several others.

More details about why experiments matter in the space: papers.ssrn.com/sol3/papers....

Professors should have full freedom to experiment with AI (or not!), evaluate what works, and, hopefully, share evidence with others. In short, we should be scientific in the classroom. Let’s not follow the World Controllers. www.networklawreview.org/world-contro...

Some legal scholars are calling to “ban AI from the law faculty”. Tonight at Luxembourg Bar, I will argue that a ban is the very, very last thing we want. A short thread.

I share this with utmost humility. The topic is large and I am sure the paper has blind spots. But I hope it offers a useful step for anyone interested in the future of EU digital regulation. papers.ssrn.com/sol3/papers....
Adaptive Regulation
What is adaptive regulation? Why does it matter? How can it be measured, and how can regulation be made more adaptive? I answer each of these four questions.
papers.ssrn.com

➝ I expanded the design principles with more concrete institutional proposals, with examples across the DMA, DSA, AI Act, etc.
➝ I added a new section on ex ante regulatory regimes and the way they suppress institutional memory, along with options to rebuild it through adaptive design.

➝ I included a clearer explanation of the counterfactual problem, especially the difficulty of adapting rules when the main effects concern innovation that never materializes.

This new version includes several additions that I hope strengthen the argument.
➝ I added a proportionality analysis showing how adaptive regulation aligns with existing EU legal principles and case law.

➝ Second, I propose a framework for future-responsive regulation rooted in #complexityscience. It focuses on modular architecture, distributed sensing, pluralistic triggering, and networked institutional memory as practical design principles.

The paper has two core parts.
➝ First, I provide an empirical look at the eight major EU Digital Acts (DMA, DSA, AI Act, Data Act...). I map which Acts embed monitoring tools, triggering mechanisms, and revision channels, and which ones still lack the structures needed for genuine adaptiveness.

I have just uploaded a new version of my article on 𝗮𝗱𝗮𝗽𝘁𝗶𝘃𝗲 𝗿𝗲𝗴𝘂𝗹𝗮𝘁𝗶𝗼𝗻 (papers.ssrn.com/sol3/papers....). The goal: understand how EU digital regulation can evolve in a way that remains legally sound and operationally realistic, instead of chasing the impossible "future-proof" concept.

Yann (@yann-lecun.bsky.social) just announced that he is leaving Meta. Some media describe this as a sudden lack of interest in LLMs. Hold on. In May 2024, Yann had already said exactly that during our conversation: open.spotify.com/episode/7GhU....
#5 – Yann LeCun: AI Dynamics and Regulation
Scaling Theory · Episode
open.spotify.com

What this article says: Macron-aligned lawmakers have introduced a bill that would ban social media for those under 15. The text also proposes a digital curfew for 15- to 18-year-olds from 10 p.m. to 8 a.m., along with a ban on smartphones in high schools. www.lemonde.fr/pixels/artic...
Les députés macronistes déposent une proposition de loi pour interdire les réseaux sociaux aux moins de 15 ans
Le texte propose également d’établir un couvre-feu numérique pour les 15-18 ans de 22 heures à 8 heures, ainsi que l’interdiction des smartphones dans les lycées. Il reprend les préconisations de la c...
www.lemonde.fr

France is turning into Huxley’s Brave New World. The World Controllers claim to know what is best for you. They promise happiness, they hold the expertise, you do not. Trust them.

Cristina Bicchieri (@upenn.edu) joins me on #scalingtheory to unpack how norms emerge, scale, and sometimes break down. Fascinating! New episode out now.
➝ YouTube: youtu.be/T1fmu5dRbvA
➝ Apple: podcasts.apple.com/us/podcast/2...
➝ Spotify: open.spotify.com/episode/2UQQ...

Reposted by Thibault Schrepel

What happens when industrial policy moves from tariffs to invisible regulatory barriers? Daniel Spulber shows how non-tariff barriers distort trade & erode incentives to innovate. New in the NLR x ICLE special issue on competitiveness.
👉 networklawreview.org/spulber-indu...

It's out (in open acces). Nicolas Petit, Bowman Heiden and I situate the “dynamic competition approach” in relation to other #antitrust approaches. We explain why innovation rivalry demands its own antitrust method.
👉🏼 Download here: papers.ssrn.com/sol3/papers....

Reposted by Thibault Schrepel

Is “industrial policy” really back? Giovanni Dosi argues it’s mostly a dystopian reboot, more zero-sum than visionary-oriented strategy. Our new NLR–ICLE special issue starts with a bang.
www.networklawreview.org/dosi-industr...

Further evidence that the European Commission is using the DMA to bypass an Article 102 TFEU analysis. #Google #DMA
Today we've launched proceedings to assess whether Google applies fair, reasonable and non-discriminatory conditions of access to publishers’ websites on Google Search.

The investigation focuses on Google’s site reputation abuse policy and its impact on publishers → link.europa.eu/kb7P7H

#DMA

Reposted by Thibault Schrepel

Today we've launched proceedings to assess whether Google applies fair, reasonable and non-discriminatory conditions of access to publishers’ websites on Google Search.

The investigation focuses on Google’s site reputation abuse policy and its impact on publishers → link.europa.eu/kb7P7H

#DMA

Here is the European Commission “Digital Omnibus on AI.” It’s a good start, but remains 𝗹𝗮𝗿𝗴𝗲𝗹𝘆 insufficient. The AI Act hurts competition and innovation. papers.ssrn.com/sol3/papers....

Excellent contributions coming your way. Check it out www.networklawreview.org/special-issu...

Boom.