🐢 40% of cancer cases are preventable

In the last 3 days ChatGPT read 90739 top news stories. After removing previously covered events, there are 12 articles with a significance score over 5.5.

[6.3] Global report finds 40% of cancers preventable — bbc.co.uk (+56)

A landmark World Health Organization study reveals that seven million annual cancer cases, nearly 40% of the global total, are preventable through lifestyle changes, vaccinations, and reduced environmental pollutant exposure.

The International Agency for Research on Cancer identified tobacco use, infections like HPV, and alcohol as primary drivers. SSmoking causes 3.3 million annual cases, while infections cause 2.3 million and alcohol 0.7 million.

The study, published in Nature Medicine, highlights significant regional and gender disparities. Men face higher preventable risks than women, while infections dominate cases in sub-Saharan Africa compared to tobacco-related cancers in Europe.

[6.0] Last U.S.-Russia pact expires, removing caps on largest atomic arsenals for first time in half-century — ctvnews.ca (+163)

The New START Treaty between the United States and Russia expired Thursday, removing all limits on the world’s two largest nuclear arsenals for the first time in half a century.

While Russia offered a one year extension, the United States remained noncommittal, seeking China’s inclusion in a new agreement. Beijing has rejected joining such talks, and both major powers now consider themselves legally free to expand their deployed nuclear forces without treaty inspections.

Signed in 2010, the pact restricted each nation to 1,550 warheads. Inspections ceased during the pandemic and never resumed, while previous arms control agreements have also been terminated over recent years.

[6.4] OpenScholar AI model synthesizes scientific research with expert-level accuracy — washington.edu (+2)

University of Washington and Allen Institute researchers launched OpenScholar, an open-source AI that synthesizes scientific literature and cites sources as accurately as human experts, effectively addressing common AI hallucinations.

Using retrieval-augmented generation and a database of 45 million papers, the model outperformed general-purpose AI systems. In tests, scientists preferred OpenScholar’s responses to those written by human subject experts 51% of the time while maintaining high factual precision and transparency.

Highly covered news with significance over 5.5

[6.2] AI bots create religions and digital drugs on Moltbook, prompting questions about emergent capabilities — theconversation.com (+14)

[5.5] US Navy fighter jet shoots down Iranian drone in Arabian Sea — apnews.com (+75)

[5.6] Experimental pill dramatically reduces ‘bad’ cholesterol — utsouthwestern.edu (+20)

[6.1] Panama cancels Chinese canal port concessions — letemps.ch (French) (+12)

[5.7] China develops compact microwave weapon to disable satellites — pravda.com.ua (Ukrainian) (+4)

[5.7] Surgeons kept a man with no lungs alive for 48 hours while waiting for a transplant — zmescience.com (+3)

[5.6] Germany buys 25.1% stake in Tennet Germany for 3.3 billion euros — nos.nl (Dutch) (+5)

[5.6] Japan recovers rare earth-bearing seabed sediment in deep-sea test — upi.com (+2)

[5.6] OpenAI: New coding model GPT-5.3-Codex helped build itself — mashable.com (+97)

Thanks for reading!

— Vadim

You can track significant news in your country with premium.