Antoine Dujon
2.2K posts
@AMDujon
French scientist living in Australia doing #stats on marine and terrestrial species to understand the links between #evolution and #cancer. Part of @CANECEV.
Hot take for the future of peer review: Journals should start asking for a lightweight AI-based replication check (e.g., via Claude) at submission. Not to replace reviewers, but to catch coding errors, logic inconsistencies, and reproducibility issues before a paper reaches them. At this point, many of these checks are fast, cheap, and automatable. There’s little reason to rely solely on human detection. Even with restricted data, this is feasible. Authors can generate simulated datasets that preserve structure and run identical pipelines. The goal is just basic verification. More broadly, we need to rethink how we use reviewer time. Not every submission needs 3 full human reviews. A more efficient pipeline might look like: editorial triage, AI-assisted checks/review, targeted human evaluation where it matters most. If done well, this could raise standards while reducing burden on the system.
BREAKING: Everyone is watching the Strait of Hormuz for oil and fertilizer. Almost nobody has noticed that it is also shutting down MRI machines, semiconductor fabs, and the global aerospace supply chain. Helium. The second lightest element in the universe. No substitute exists for it. You cannot synthesize it. You cannot replace it. And roughly one-third of the world’s supply just went offline. Qatar produces 30 to 33 percent of global helium as a byproduct of LNG processing at Ras Laffan, home to the largest helium production facilities on Earth. When the Hormuz blockade triggered LNG force majeure declarations and attacks hit Qatari infrastructure, the helium stopped flowing with it. Prices have doubled in spot markets. And helium has a property that makes this crisis structurally different from oil, fertilizer, or any other commodity caught behind the strait. It evaporates. Continuously. Even in sealed containers, helium boils off. The global supply chain operates on roughly 45 days of buffer before existing inventory simply ceases to exist. You cannot stockpile helium the way you stockpile crude oil in salt caverns or grain in silos. If the supply stops for six weeks, the buffer is gone. Not depleted. Gone. Returned to the atmosphere where it is too diffuse to economically recapture. This is why the industries that depend on helium are facing a crisis that no financial instrument can solve. Semiconductor manufacturing requires ultra-pure helium for wafer cooling in lithography and for leak detection in sub-5-nanometre chip fabrication. TSMC, Samsung, and Intel cannot produce advanced processors without it. Every AI chip, every smartphone processor, every data centre GPU in the current generation traces its manufacturing lineage through a helium-cooled process. If fabs run dry, the production lines stop. Not slow. Stop. MRI machines require liquid helium to cool superconducting magnets to near absolute zero. Hospitals cannot substitute another gas. When helium supply tightens, MRI availability falls. During previous shortages, hospitals rationed scans. A sustained one-third supply cut puts diagnostic imaging capacity at risk across every healthcare system that depends on magnetic resonance. Aerospace depends on helium for purging rocket fuel systems, pressurising tanks, and testing for leaks in systems where failure means explosion. NASA, SpaceX, ULA, and every launch provider in the Western world runs on helium. Fibre optic cable manufacturing requires helium atmospheres. Quantum computing research requires helium-3 isotopes for cryogenic cooling. The US is the world’s largest helium producer and has some buffer capacity. Algeria and Russia produce meaningful volumes. Overland rerouting from Qatar through Oman and Saudi Arabia is theoretically possible but logistically slow and capacity-limited. None of these alternatives can replace one-third of global supply within the 45-day evaporation window that defines the crisis timeline. The same 21-mile strait that is starving the food system is now threatening the technological infrastructure of modern civilization. The fertilizer trapped behind Hormuz determines whether four billion people eat. The helium trapped behind Hormuz determines whether the chips powering the AI revolution get manufactured, whether cancer patients receive diagnostic scans, and whether rockets carrying communications satellites reach orbit. One chokepoint. Two invisible supply chains. Both irreplaceable. Both operating on biological or physical deadlines that no ceasefire retroactively extends. The world built petroleum reserves. It never built fertilizer reserves. It never built helium reserves either. The pattern keeps repeating. The lesson keeps being ignored. Full analysis: open.substack.com/pub/shanakaans…
Do I understand this correctly? This ICLR submission had very good reviews but the AC rejected it because the paper had bad reviews at NeurIPS and the AC was part of the reviewers? One of the most unfair decisions I've ever seen
