Over 2,000 compute hours later, they find algorithmic collusion typically breaks. HOWEVER: pair a 32B model with a 14B one and the big model leads, the small one follows, and prices stay 21% above competitive. Capability gaps stabilise cartels rather than disrupt them.

Andrew Perlman (Dean Suffolk Law) has built an archive comparing how 128 US law schools handle AI in coursework. Berkeley bans it by default. Chicago teaches 1L courses device-free while building supervised use into legal writing. Texas dropped take-home exams, etc...

I analyzed the definitions provisions of all 34 EU digital instruments and recorded ~1,400 defined terms. The famous complaint that the rulebook is incoherent? Mostly wrong. This does not mean EU regulation is easy to navigate... My findings: thecompetitionsystem.substack.com/p/definition...

Force a platform to share its data with rivals on equal terms and you may end up with less data for everyone. Hanming Fang and Soo Jin Kim show that weak data neutrality is bypassed through pricing, and strong data neutrality cuts the platform's incentive to refine data in the first place. Equal acc

“A lot of macroeconomics really wastes computing power (...) solving for very complex equilibrium behavior of a sort that goes far beyond what you could ever imagine a real human being is doing...

My July reading suggestions are out. Merger simulations tested at scale, AI agent collusion turning out to be more fragile than feared, backdoors in neural networks that no statistical test can catch, user innovation, and why the AI labs keep hiring philosophers. networklawreview.org/july-2026/

Anti-steering under Article 102: Apple, €40m basic amount, nearly nine years of infringement. Anti-steering under the DMA: Google, €430m, a little over two years. Annualized, the DMA prices the same conduct at roughly 40x the Article 102 rate. Here is my analysis: open.substack.com/pub/thecompe...

All competition agencies now enforce with computational tools. Regulation 1/2003 was drafted for investigators who read documents. I explain (here briefly) where that breaks + three funded PhD positions I am recruiting to work on it. legalblogs.wolterskluwer.com/competition-...

In September 2024 the Court of Justice refused to make self-preferencing unlawful as a general rule. Article 6(5) DMA is that general rule. On Thursday the Commission enforced it against the same company whose litigation produced the refusal. Here is my analysis: open.substack.com/pub/thecompe...

NEW 📻 Peter Howitt won the 2025 Nobel for growth through creative destruction. On #ScalingTheory, he explains why more competition does not always produce more innovation, and why it depends on how close a firm sits to the technological frontier. We also talk about the EC’s draft Merger Guidelines..

NEW: thecompetitionsystem.substack.com/p/what-the-a... September 2025: the Commission refuses to hold Microsoft’s own service to the standard it set for Microsoft’s rivals. July 2026: it imposes that exact rule on Google. As settled law. Both texts are in the piece, quoted and dated. This is reversa

Looking at every European Commission merger decision, the most-cited decisions are not the famous ones. They are market definition anchors in energy, oil, telecoms and retail. Exxon/Mobil sits at the center. The cases we write about do not.

Looking at every European Commission merger decision, the most-cited decisions are not the famous ones. They are market definition anchors in energy, oil, telecoms and retail. Exxon/Mobil sits at the center. The cases we write about do not.

NEW substack. The Commission’s DMA decisions on Google (last week) dress up new law as old practice. I supply the citations it left out. Every antecedent, every contrary ruling, dated and quoted in full. Results? Five documented reversals. thecompetitionsystem.substack.com/p/what-the-a...

EU merger law tells companies they can defend a deal with efficiencies. I read all 57 cases where they tried. Accepted as decisive: zero. The defence has never once saved a merger the European Commission wanted to block. It is a promise the record does not keep. Paper coming.

This paper is an important reminder that “open weights” do not automatically create trust. If model trainers can, in principle, plant statistically undetectable backdoors, then relying on open-source frontier models developed by geopolitical rivals becomes a much harder governance question.

The EU already has digital sovereignty. Just the wrong kind. Across 34 laws, five types written into the rules: ➝ infrastructure ➝ adequacy ➝ jurisdiction ➝ access rules ➝ credentials & coin All of it controls foreign tech. None builds any. Audit 👇 thecompetitionsystem.substack.com/p/the-paperw...

I ran the European Commission’s new draft Merger Guidelines against all the merger decisions ever published. In many regards, the Guidelines go against the case law while pretending otherwise. Paper to follow...

Brussels calls the Tech Sovereignty Package the moment Europe gets serious about digital independence. I mapped 34 EU digital laws. Europe has legislated sovereignty since 2014. The legal stack exists. It governs foreign tech, builds none. thecompetitionsystem.substack.com/p/the-paperw...

Summer = research & running ☀️ Living in Amsterdam for eight years is a gift. Why travel elsewhere in the summer? With the canals, forests, the swimming, we have so much right here. This city is so incredibly hard to beat during the summer time.

Two AI agents. 400 rounds of pricing. No agreement, no messages, and yet the price keeps climbing. One of them even suggests they avoid raising red flags. Tanise Brandão and Carlos Neves (CADE) join Alba Ribera Martínez and me on a new Stanford #ComputationalAntitrust episode.

The Google Android fine was never the story. The story is one sentence that retires the as-efficient competitor test in digital markets. It will be quoted for years. My full audit of C-738/22 P 👇 thecompetitionsystem.substack.com/p/the-quiet-...

I am launching a newsletter. It is called The Competition System. It is an empirical, referenced account of the cases and rules that shape competition in Europe. Articles 101 and 102, the DMA, DSA, the AI Act, MiCA, the Data Act, etc. Every claim is traceable to its source.

The year is half done. I made it to Rome. Now back to Paris. 😅

Wrote a strong Master’s thesis on competition law or competitiveness this year? We are opening the 2nd edition of our award, with ICLE and VU Amsterdam. Three winners present in Amsterdam, travel covered. Deadline 31 August 2026. alti.amsterdam/call-for-sub...

Wrote a strong Master’s thesis on competition law or competitiveness this year? We are opening the 2nd edition of our award, with ICLE and VU Amsterdam. Three winners present in Amsterdam, travel covered. Deadline 31 August 2026. alti.amsterdam/call-for-sub...

1 – Today’s Google Android judgment (C-738/22 P) is the end of a 13-year case. It’s also the start of something bigger. The EU Court of Justice just confirmed the 3rd generation of big tech competition enforcement. And the Commission is already applying it to AI, before the ink is dry. A thread 🧵

Data neutrality does intensify downstream competition, but it kills the platform's incentive to refine and produce the very input the rule is trying to equalize. The DMA’s data-sharing mandate may be equalizing access to a shrinking pie. Fair ≠ welfare-enhancing.

This month’s reading suggestions from @profschrepel.bsky.social are out. Inside: an LLM that predicts case outcomes, the fight over public ownership of the big AI labs, and whether you can tell when AI wrote the article. Plus Pinker on common knowledge. www.networklawreview.org/june-2026/

New #ScalingTheory episode out now, with Steven Pinker. Here, Steven questions whether AI could bring us back together, after social media pulled us apart. ▶️ www.youtube.com/watch?v=ijNj... 🎧 open.spotify.com/episode/36hX... 🍎 podcasts.apple.com/fr/podcast/s...