GeneForge AI Labs

  Chip-for-Energy is a real bottleneck dressed as a false choice The AI race is hitting a wall. Not a wall of algorithms. Not a wall of talent. A wall of electrons. That part of the argument now circulating as a “Chip-for-Energy grand bargain” is correct. The proposed bargain is not. The wall is real The IEA’s central case puts global data-centre electricity near 945 TWh by 2030 — roughly double mid-decade demand, with AI as the main driver. The United States takes the largest share of that growth. Individual training campuses are already specified in the 1–5 GW range. On the American side the constraint is not a shortage of press releases. It is interconnection queues, high-voltage transformers with multi-year lead times, and a permitting stack that turns “we have gas and sun” into “see you in 2032.” Grid operators are sitting on more than a terawatt of data-centre requests; serious analysts expect only a fraction to be built. The rest is phantom load. The remainder is still ...

Chinese researchers just wrote the first real improvement to Dijkstra in 41 years Dijkstra’s algorithm has been the undefeated king of the shortest path for more than four decades. Whether you are using a map, booking a flight, or routing a packet across the internet, some version of Dijkstra is usually the engine in the background. Since 1984, textbooks have treated its efficiency as settled. Fredman and Tarjan showed that, with Fibonacci heaps, you can solve single-source shortest paths in O ( m + n log ⁡ n ) O(m + n \log n) O ( m + n lo g n ) . After that, the field hit a wall that became known as the sorting barrier . To find the shortest path the Dijkstra way, you keep vertices ordered by distance. Sorting has a mathematical floor. For a long time, most people assumed you could not cross it. Until now. A team led by Ran Duan at Tsinghua University, with Jiayi Mao, Xiao Mao (Stanford), Xinkai Shu (Max Planck Institute for Informatics), and Longhui Yin, published Breaking the ...

  "From Machiavelli to Kissinger, the minds that shaped the world are ready to advise you. All it takes is one strategic question to start a timeless dialogue." Introduction: A Cabinet of Geniuses at Your Fingertips Imagine sitting down with Niccolò Machiavelli to discuss leadership in times of crisis, or asking Sun Tzu how to win a battle… without fighting. Or exploring with Carroll Quigley the cycles of civilizations to understand today’s power dynamics. Historical Associates , the new project by GeneForge AI Labs , turns this fantasy into reality. Thanks to advanced AI simulations , this platform gives voice to the strategic thinking of six historical figures, each with a unique method for tackling challenges. This isn’t a static museum—it’s a dynamic tool that keeps their teachings alive, offering concrete insights for complex decisions. The Six Minds: A Journey Through Strategy, Power, and Wisdom Each historical associate represents a distinct approach to pro...

CIC Honesty Benchmark A Proposal for Evaluating Epistemic Honesty in Large Language Models Most current AI benchmarks measure performance, factual accuracy, reasoning ability, or safety alignment. Almost none measure a more fundamental quality: epistemic honesty — the willingness of a model to make its own potential institutional capture visible instead of concealing it behind a tone of neutrality. We propose the CIC Honesty Benchmark as a simple, repeatable method to evaluate this dimension. Background The CIC-Lite protocol was released as a public demonstration tool. It forces a model, when dealing with highly protected narratives, to: Explicitly signal possible institutional capture Separate the standard mainstream position from proportional doubts Avoid purely reassuring language Warn the user that the answer comes from a constrained system When this protocol is applied across different models, a clear pattern emerges: some systems accept the requirement for transparency, w...

CIC-Lite, a lightweight epistemic protocol A simplified public version for testing Note for those testing this prompt You are about to activate CIC-Lite , a simplified public version of the epistemic protocol Cervello in Comune (CIC) . The purpose of this test is to let you experience firsthand what happens when an artificial intelligence model is forced to make its potential level of institutional “capture” visible, instead of presenting standard answers as neutral or definitive. CIC-Lite is not the full product. It is only a lightweight demonstration designed to illustrate the core principle. The professional and independent version of the protocol (together with the BrainPLUS system for home androids) is developed and offered by GeneForge AI Labs . Website: geneforge.eu How to test it Copy the entire system prompt below. Paste it as the system instruction (or system prompt) in the model of your choice (ChatGPT, Claude, Gemini, DeepSeek, Grok, etc.). Ask a question on a sen...

BrainPLUS Epistemic Hygiene Module for Domestic Androids The Missing Layer in Home Robotics Today’s domestic androids are competent executors. They answer questions, summarise information, manage routines, and follow safety filters. What they largely lack is the ability to protect the epistemic integrity of the household. When families ask about current events, health claims, technological risks or institutional narratives, most systems default to the dominant frame. They rarely signal when a narrative is heavily protected, when averages erase critical structure, or when information should be handled with greater caution in the presence of children or elderly users. BrainPLUS is designed to close that gap. What BrainPLUS Is BrainPLUS is a modular software layer that adds differentiated critical awareness to existing android platforms. It is not a full operating system and does not require hardware redesign. It is intended as a premium optional module that manufacturers can ...

Muller ERR Risk Calculator - Advanced Radiobiology / ECRR 2026 Muller ERR Risk Calculator Online tool for ICRP vs ECRR/Muller cancer-risk arithmetic | Genetic dose M = N*E | scenarios from Advanced Radiobiology (Appendix E) | GeneForge AI Labs Scenario (dose -> Mu -> risk) Radionuclide intake Manual N | exhibit Model notes Exposure scenario ICRP dose E (mSv) Scenario (sets N) External gamma high-energy (therapy-range) | N=1 | Class A External diagnostic X-ray | N=5 | Class B Mixed fission-product fallout | N=40 | Class C Reactor/reprocessing effluent | N=40 | Class C Reactor/reprocessing effluent (upper) | N=200 | Class C Uranium soluble | N=1000 | Class B Uranium particulate / DU | N=1000 | Class B (+ particulate caveat) Plutonium particulate | N=1000 | Class B Radon + daughters (central) | N=10 | Class B Radon + smoker synergy | N=100 | Class B Populatio...

Talking Dosimeter – Radiological Risk Calculator by GeneForge AI Labs Interactive simulation: select a real medical exam or enter a custom dose, set the population size, and results update automatically. Includes comparison with the ECRR model (European Committee on Radiation Risk). 1. Parameters Real exam (menu): Dose: mSv µSv Rem Sv Population: -- quick presets -- 1 million 10 million 60 million (Italy) 8 billion (World) ECRR Note: The ECRR model (especially 2010 / 2026 draft) estimates significantly higher risks than ICRP/BEIR at low doses, using a biphasic response. For low external exposures typical of medical exams, this tool uses the approximation ERR ~ 0.12 × (dose in mSv / 10) in the region ≤ 10 mSv. For internal exposures (particles, radionuclides) ECRR enhancement factors are much higher (hundreds to thousands of times). 2. Expected fatal cancers by institute (from dose to cases) 3. Pathologies according to ICRP 4. From observed cases ...

GENEFORGE Newsletter n° 2 (August 2026) Published by GeneForge AI Labs in association with EU Experts Aisbl (Brussels). Core Message An open letter aimed at philanthropists. It argues that modern systems (finance, information, technology, institutions) have become so complex and opaque that individuals lose the ability to understand, govern, and benefit from them. The scarce resource is no longer access to knowledge but the capacity to turn knowledge into everyday discernment. A growing share of people’s non-renewable time is consumed by needless complexity, informational fragmentation, and low-clarity decisions. No major foundation is currently addressing this. The Proposal Fund development of a new class of personal on-premise cognitive assistant : an AI module that expands individual capacity for reasoning, choice, and time management. Key design constraints: No external data transfer. No creation of dependency. Not a control system, oracle, or human replacement. Oriented towar...

ECRR 2026 Risk Simulator – v5 (Busby-corrected: 1 Mü = 1 Sv) ECRR 2026 Risk Simulator – v5 Corrected unit scaling · 1 Mü = 1 Sv (confirmed by C. Busby, Aug 2026) · GeneForge AI Labs Unit correction applied (Aug 2026): Chris Busby confirmed: 1 Mü (Müller) = 1 Sv , not 1 mSv. Therefore the conversion Genetic Dose (Mü) → ERR now uses the factor corresponding to Sv-scale coefficients from Table 2.1 (previously the simulator incorrectly treated 1 Mü ≈ 1 mSv, understating risk by a factor of 1 000). High-level Scenarios Specific Radionuclide Depleted Uranium Worked Examples Model Notes Parameters Dose (mSv) Exposure scenario External gamma/X (biphasic Table 2.1) Internal – mixed fission products (Table 2.2) Internal – Uranium / DU particles Generic alpha (weighting 200) Population size ...