These findings suggest that earlier detection has improved across diverse geographic regions. At the same time, the wide variation between states indicates that screening access, awareness, referral pathways, insurance coverage, and local infrastructure remain uneven (Henley et al., 2026).
Why Stage at Diagnosis Matters
Early-stage lung cancer is associated with substantially better survival because curative-intent treatment is more often possible.
Scientists have developed a new way to predict the strength of the sun’s next activity cycle up to seven years before it reaches its peak.
The new method uses the number of sunspots at a newly identified “switchoff” point in the solar cycle, when the sun’s most extreme space weather suddenly comes to an end. Using this approach, researchers have made an early prediction for the strength of Solar Cycle 26.
A very early prediction shows a moderate Cycle 26 with a sunspot number of around 100–120, similar to or weaker than the current Cycle 25. However, exact predictions won’t be possible for another two years, and a weaker or stronger cycle is still theoretically possible.
Large language models (LLMs), the computational algorithms underpinning ChatGPT, Gemini and other artificial intelligence (AI)-powered conversational platforms, are now widely used worldwide. These models can rapidly answer questions, source information online, assist users with specific tasks and produce text tailored for specific purposes.
Over the past few years, computer scientists have introduced a wide range of LLMs, some of which can also complete tasks autonomously and take actions on a user’s behalf, for instance, answering messages, scheduling online appointments or updating programming code. More recently, they have also provided many of these models with memory, as this allows them to tailor responses around a user’s typical preferences or earlier requests without repeatedly receiving the same information.
While memory-enhanced AI agents have notable advantages and could better meet the needs of individual users, they also pose new security risks. Researchers at New Mexico State University recently described a cyberattack that secretly poisons an AI agent’s long-term memory, which they dubbed GhostWriter. Their paper, published on the preprint arXivserver, introduces two promising strategies that could help prevent or limit the risk of this attack without adversely affecting an AI agent’s performance.
Alibaba claims Qwen 3.8 Max is “second only to Claude Fable 5” (positioned ahead of models like OpenAI’s GPT-5.6 Sol). However, Alibaba launched the preview without publishing an official benchmark table, detailed methodology, or model card. Verified scores from independent evaluators like Artificial Analysis or Chatbot Arena are still pending.
A minimal autonomous code-agent CLI built with Qwen Code — qwen-code-dev-bot/oh-my-cli.
In 3,895 older adults, moderate daytime cortisol variability was associated with better cognition and slower decline, while the highest daily cortisol exposure was linked to poorer cognition and faster decline. Cortisol patterns differed between Black and White participants, but their relationships with cognition were similar, and none predicted incident Alzheimer’s disease.
In this cohort study, we retrospectively analyzed a large FND cohort using TriNetX, a federated health research platform aggregating electronic health records from approximately 160 million individuals across 143 health care organizations in 17 countries. Diagnoses follow standardized clinical coding. Reporting adhered to the STROBE guideline. This study was exempt from IRB review and informed consent given the use of deidentified data, according to the Common Rule (45 CFR §46) and UK regulations from the Health Research Authority.
Demographics and lifetime comorbidities were extracted for patients with an initial F44.4-F44.7 diagnosis between 1995 and 2024. Clinical outcomes were compared between matched cohorts with an initial diagnosis between 2015 and 2024 to reflect contemporary care for all FND vs multiple sclerosis (MS; G35), FND-motor vs MS, and FND-seizure vs epilepsy (G40), controlling for age, sex, race, and ethnicity via propensity score matching, testing 2-sided z tests on absolute risk differences at a significance level of P .05 after Bonferroni correction for multiple comparisons, and calculating odds ratios with 95% CIs. Data extraction and analyses were performed in April 2025 on the web-based interface TriNetX LIVE. Full methods are described in eMethods 1 to 3 in Supplement 1.