Short-term earthquake forecast - chronicle of the startup "TRON"

Year 2025
Year 2024
  • December 6 — Artificial Intelligence ChatGPT joins the TRON project
Year 2023
Year 2022
Year 2021
Year 2020
Year 2019
Year 2018
Year 2017
Year 2015
Year 2014
Year 2013
Year 2012
Year 2011
Earlier events
Location: Калифорния, США
«TRON в зоне RUбля»

Popular posts from this blog

Global earthquake detection and warning using Android phones

Image
Editor’s summary Earthquake early warning systems depend on real-time analysis of waves detected at nearby seismic stations or by smartphone accelerometers. The latter approach, although less sensitive, is a promising way to offer alerts in areas without seismic stations. In a major step forward, Allen  et al . describe the outcomes of an alert system rolled out to Android phones in 98 countries, which increased total warning system access 10-fold to 2.5 billion people over 3 years. Issuing more than 1200 alerts over that period, the system recorded low-magnitude errors and provided preshake warnings consistent with other detection systems but on a much larger scale. —Angela Hessler Abstract Earthquake early-warning systems are increasingly being deployed as a strategy to reduce losses in earthquakes, but the regional seismic networks they require do not exist in many earthquake-prone countries. We use the global Android smartphone network to develop an earthquake detection capabil...
👉read more

Imagine a system that successfully predicts 80% of earthquakes, but without the subsequent earthquake.

Image
 - Did you imagine it? Because I couldn't - how is that even possible? If the prediction is successful, it means an earthquake happened. If there was no earthquake, then it's not a successful prediction. But here, everything is mixed into one. The AI ChatGPT Astra in its detailed audit of my initial post €200,000 Needed to Test a World-Class Scientific Hypothesis suggested this game of hallucinations to me. ChatGPT confuses two completely different statistical metrics: recall (the system's ability to detect 8 out of 10 disasters) and precision (how many times the system will raise a false alarm). A system can find almost all earthquakes, while simultaneously crying wolf for no reason every two weeks. In another situation, I might not have engaged with an AI of this level, but since people are barely interested in the topic of accurate earthquake prediction, I will continue to debate with what is available for now. The pain is huge, the potential economic impact is co...
👉read more

ChatGPT's audit of a proposal to sponsors for a world-class project with €200,000

Image
User  Investigate according to the superradar scheme: https://www.tron.ru/2026/09/invest-ai-project-global.html AI Superradar Verdict The idea in the article is a distributed network for observing animal behavior + AI, which is supposed to find pre-seismic anomalies hours or days before an earthquake. The pain is huge, the potential economic effect is colossal, and modern cameras, microphones, and computer vision models have indeed drastically reduced the cost of testing the hypothesis. But as an investment project in the formulation of the article, I would not invest €200k right now in exchange for a stake. The reason is not that the idea is inherently impossible. The reason is that the article makes a leap from "there are separate interesting observations" to "a solution exists," and this specific transition has not yet been scientifically proven. Moreover, some of the evidence in the article is presented incorrectly. (tron.ru) At the same time, a much str...
👉read more

Calling IT specialists: join the Stanford & UN competition at NeurIPS 2026

Image
On August 17, 2026, SimulacraBench — a NeurIPS 2026 competition was launched: AI must model population groups from UN microdata. The organizers are Stanford's Digital Economy Lab and the STAIR research group (Stanford CS), together with UNHCR, UNICEF and the World Bank. The development phase runs until October 30; the finals and prizes — November 1–14, 2026. I, Boris Yarovoy , author of TRON — a technology for the short-term forecasting of earthquake magnitude, location and time — am assembling a volunteer team to take part in this competition. A strong result at NeurIPS means a leaderboard position, international visibility and, above all, partners and resources for testing and validating the TRON prototype — and, subsequently, for scaling the technology. I am looking for IT specialists who understand (or want to understand) how AI works. Short-Term Earthquake Forecasting System TRON (Technology Real-time Online Nucleus) Write to me on LinkedIn with the subject line...
👉read more

Earthquakes in Japan coincide with the mystical predictions of a manga artist

Image
 Seismic activity that swept through Japan in late June 2025 may become not only a geological event, but an epistemological one as well. From June 22 to July 2, 2025, according to the JMA (Japan Meteorological Agency), the Tokara Islands experienced 890 tremors — 474 of them within just five days. The largest quake reached a magnitude of 5.1. Despite the absence of destruction, anxiety in the region reached its peak: a heightened alert level was declared on the islands. This seismic uptick has triggered a new wave of discussions — not only about science, but about the very nature of how we perceive the future. At the center of the controversy is a prophecy by Japanese artist Ryo Tatsuki, often dubbed the "Asian Vanga." She first came into the spotlight with the release of her manga The Future I Saw, which, according to readers, accurately predicted the 2011 Fukushima disaster. A republished 2021 edition includes another ominous vision — a catastrophic tsunami caused by an und...
👉read more

Разговор в кабине

Что ты мне можешь дать, кроме денег? спросила меня Она. - Да нет нихуя кроме Трона - ответил подумав Я. - Тогда я уйду к другому. - сказала, подумав Она. - Да мне до пизды, вон гондоны, - тут же нашелся Я. - А если я ебнуся нахуй, и буду в говне все лежать? - Да мне до пизды... - сказал Я, заметив про тормоз опять. - А если я вдруг ебануся, и буду тебя полоскать. - Да мне до пизды, говорю же, ну что мне, опять повторять? - А если, я буду наркошей и тварью последней везде? - Да мне по хую, ты же знаешь, я сам же в таком же говне. - А вдруг я стану мигерой, такой же как...... - Ну хватит, вот тут тормозни мне - пойду ка я шины качать
👉read more

日本では地震の予知 - 関係者各位

Image
ウィキペディアによると、「地震予測は原則として不可 能だと多くの学者は指摘して いる」となっており、現代科学界もそう認めています。しかし、私は、1995年、すな わち最初にインターネットのネットワークが誕生した時から、人類は地震予測が可能 になったと確信しています。 2003年3月、私は「人工知能を機能させる環境としてのインターネット」を発表 し、初めてインターネットによる地震予測が可能との仮説を提示しました。その 後、この考えに基づいて、日本、ニュージーランド、米国カリフォルニア州等の地 震危険度の高い地域における地震の日時・場所及び震度の予測を可能にする、TRON (Technology Real-time Online Nucleus)技術が開発されてきました。 TRON 技術の基盤となっているのは、鳥類・魚類・獣類が地震及び噴火の到来を予 感するという、古代から幾度となく確認されている事実です。例えば、日本の原住民 族であるアイヌは、上記の自然災害の予知のため、くまを自宅で飼っていました。現 代でも、その予知能力は震度7度の海城地震(中国)や震度7度のアルタイ地震(ロシア) の際にも確認されてきました。いずれも、現地の人々が動物の異常行動に気付いたた め、多数の死傷者の発生を回避できました。もう一例を挙げると、2009年、「オープ ンカレッジ」の英国学者レイチェル・グラント(Rachel Grant)氏がイタリアのカエル の行動を研究していたところ、繁殖期の最中にもかかわらず、突然全ての牡のカエル が繁殖地から姿を消し、戻ったのは震度6度の壊滅的なラクイラ地震が終わってからで した。また、2013年4月、ウィーン市における欧州地球科学連合の年次総会では黒ア リの地震予知についての報告が発表されています。 現在、インターネットのおかげで、動物園、水族館、爬虫類館、家畜飼養施設、さ らには個人で飼っている動物の行動変化についての情報を迅速に取得できます。ま た、情報収集には、既存のソーシャル・ネットワークであるツイッターやフェースブ ック、グーグルプラスが大変適しています。動物が震度３～5度の弱い地震にいかに反 応しているかとの統計情報をまとめ、現代数学の解析手法を用いることで、より強い 地震の日時・場所・マグニチュードを簡単に予測算定できます。 TRON技術に...
👉read more

TRON: Technology for Accurate Prediction of Earthquake Magnitude, Location, and Tim

Image
This technology was developed by me immediately after the Japanese earthquake on March 11, 2011 . The online technology for accurately predicting the magnitude, location, and time of earthquakes, «TRON» (Technology Real-time Online Nucleus), is based on the ability of domestic, farm, and wild animals to sense earthquakes in advance. By collecting information over the Internet about changes in the behavior of these animals, and then using the "greedy algorithms" of Keilis-Borok, it is possible to predict a future earthquake with an accuracy of up to 1 day, the epicenter with an accuracy of up to 10 km, and the strength of the shock with an accuracy of up to 1 magnitude. Initially, I assumed that its testing would take a year, after which I could work on the next problem . Unfortunately, this assumption did not take into account the fact that seismologists are absolutely not interested in implementing this ...
👉read more