So happy we're finally out of stealth. It's been enormous fun so far, and the research is still very early. Engineering Superintelligence is close to a blank canvas: there's no playbook yet, the foundational questions are wide open, and the people who work on them now will help define the field. At the centre of those questions is one of the oldest arguments in Western thought. Parmenides held that reality is unchanging and can be reached by pure logic. Heraclitus held that everything is flux: you can't step into the same river twice. Nietzsche called one ice and the other fire [1]. Jeanne Hersch argued that the whole history of philosophy sorts into their two camps [2], and physics does too [3]. The modern round is Wheeler's "it from bit," the idea that reality is built from information, against Landauer's "information is physical": every bit lives in some substrate, and the substrate gets a vote [3]. @SchmidhuberAI sits on both sides. He has argued for decades that the universe could be the output of a program [4], which is about as it-from-bit as (b)it gets, yet he says there's no AGI without mastery of the real world [5]. These aren't contradictory. One is a claim about what the world is, the other about how you come to know it. That second question is the one Engineering Superintelligence has to answer. How much of a real machine can you learn from simulation, and where does physics overrule you? F1 even rations both sides, capping every team's CFD and wind-tunnel time. Maybe we'll settle the debate one day. For now we're deliberately choosing neither ice nor fire, because the opposition is the point. The second law itself came out of engineering: Sadi Carnot, studying steam engines in his Reflections on the Motive Power of Fire [6], showed that an engine needs two reservoirs, one hot and one cold, and that its work comes from the heat flowing between them. Research works the same way. A lab of only bit people, or only it people, sits at equilibrium. Put both in one lab and you have a gradient. Point them at a hard enough problem and it starts doing work. Bit person or it person, come build ESI with us. apexultra.com/join REFERENCES [1] F. Nietzsche, Die Philosophie im tragischen Zeitalter der Griechen (1873) [2] J. Hersch, as recounted in [3], p. 103; see her L'étonnement philosophique (1981) [3] S. Wolf, Epilogue: Information & Physics, in U. Aeschbacher, A. Hansen, S. Wolf, Invitation to Quantum Informatics, vdf ETH (2020), pp. 97-115. doi:10.3218/3989-4 [4] J. Schmidhuber, Algorithmic Theories of Everything (2000). arXiv:quant-ph/0011122 [5] J. Schmidhuber, xitter.cf/SchmidhuberAI/status/2…, itself citing Sec. 20 of arXiv:2212.11279 [6] S. Carnot, Réflexions sur la puissance motrice du feu et sur les machines propres à développer cette puissance, Bachelier, Paris (1824)