Timeline lokal

Edward Tian built GPTZero over winter break of his senior year at Princeton. Less than four years later, GPTZero has more than 20 million registered users. Superhuman agreed in June to acquire GPTZero for undisclosed terms.
Source: Fortune
https://fortune.com/2026/10/02/hallucination-detection-gptzero-edward-tian-alex-cui/
#Startups

GAVA implements grounded correction arbitration with observation-bounded evidence, legal probes, and a one-step expected-loss rule.
Source: arXiv
https://arxiv.org/abs/2610.00282
#MachineLearning

Researchers frame LLM agent recovery as a causal decision problem to balance rescue and harm. The Causal Intervention Router improves long-horizon ALFWorld success from 70.33% to 73.33%.
Source: arXiv
https://arxiv.org/abs/2610.00372
#MachineLearning

China has set a five-year goal for the expansion of intelligent computing capacity to drive artificial intelligence that calls for a more than fourfold increase on what was in place at the end of June this year.
Source: South China Morning Post
https://www.scmp.com/plus/tech/policy/article/3369537/chinas-ai-compute-ambitions-demand-innovation?utm_source=rss_feed
#AI

Researchers introduced a chess benchmark built from 1,118 Lichess puzzles to study automatic prompt optimization for frozen large language models.
Source: arXiv
https://arxiv.org/abs/2610.00416
#MachineLearning

JevSpawn is a compositional policy that links natural language task specifications to finite probabilistic exploration.
Source: arXiv
https://arxiv.org/abs/2610.00437
#MachineLearning

Researchers study heavy-tailed memory traces in long-horizon language agents and propose the Core--Tail World Model controller.
Source: arXiv
https://arxiv.org/abs/2610.00010
#MachineLearning

People delegate tasks to AI agents that share resources such as codebases, calendars, or budgets.
Source: arXiv
https://arxiv.org/abs/2610.00583
#MachineLearning

The paper introduces FedCausalCompose, a causal world-model framework for modular LLM agents that uses intervention-response evidence to improve cross-module interface recovery.
Source: arXiv
https://arxiv.org/abs/2610.00012
#MachineLearning

China's AI push has led to problems, including excessive usage and social challenges.
Source: The Business Times Singapore
https://www.businesstimes.com.sg/international/chinas-ai-push-has-led-problems-too-much-usage-social-challenges
#TechPolicy

Personalization has traditionally used customer evidence, the historical behavior and preferences of a customer.
Source: arXiv
https://arxiv.org/abs/2610.00654
#MachineLearning

Incident-Arena is a human-built benchmark of twenty tasks that deploy production applications to Kubernetes clusters and inject faults.
Source: arXiv
https://arxiv.org/abs/2610.00648
#MachineLearning

The paper proposes a spatial strategy where large language model agents collect geodesic trajectories, vector-quantize them, and use the resulting tools for decision-making in grid-world environments.
Source: arXiv
https://arxiv.org/abs/2610.00613
#MachineLearning

K-Dense BYOK is a free, open-source AI research assistant that runs on the researcher's own computer and keeps a hash-chained lab notebook.
Source: arXiv
https://arxiv.org/abs/2610.00074
#MachineLearning

When fine-tuning large language models, selecting training data by reasoning approach rather than by topic yields higher mathematical transfer.
Source: arXiv
https://arxiv.org/abs/2610.00331
#MachineLearning

The study measures the microtask eligibility gap for off-the-shelf small language models used as agent harnesses.
Source: arXiv
https://arxiv.org/abs/2610.00025
#MachineLearning

Chinese hackers impersonated a former US official to send fake emails that offered AI collaborations and linked recipients to password-stealing sites, targeting think tanks, universities and law firms.
Source: Rappler
https://www.rappler.com/technology/chinese-hackers-impersonate-ex-us-official-steal-emails-ai-experts/
#TechPolicy

Galbot announced the ET1 humanoid robot with three versions: a Standard model at 79,000 yuan, a Pro model at 139,000 yuan and a Flagship model at 179,000 yuan.
Source: Pandaily
https://pandaily.com/galbot-et1-bipedal-humanoid-price-79000-yuan-three-versions
#Robotics

ShepHertz unveiled three artificial intelligence models: Tatva for on-device commands, Sanjaya for fleet and telemetry intelligence, and Drashta for machine vision.
Source: BusinessLine
https://www.thehindubusinessline.com/info-tech/shephertz-unveils-three-ai-models-eyes-enterprise-market/article71535877.ece
#Tech

Taiwan received its first two US-made F-16V fighter jets on Friday, upgrading its military capabilities.
Source: Hong Kong Free Press
https://hongkongfp.com/2026/10/02/taiwan-receives-its-first-us-made-f-16v-fighter-jets/