Toggle light / dark theme

Get the latest international news and world events from around the world.

Log in for authorized contributors

Qwen: 🚨 China’s AI race just hit another gear

Alibaba has unveiled Qwen4-Max, its new flagship AI model — and claims it beats Anthropic’s Claude Fable 5.1 and Moonshot’s Kimi K3.1 Pro on coding benchmarks. 👀

What’s under the hood:

🧠 1-million-token context window 🎥 Understands text, images AND video 🔌 Works with both OpenAI and Anthropic APIs — easy for developers to switch.

The twist? Unlike most of the Qwen family (the world’s most-downloaded open-weight models), this flagship is closed. Smaller open versions are promised “soon.”

It’s part of a wave of Chinese launches — Tencent’s Hunyuan 4, Moonshot’s Kimi K3.1 Pro — while Alibaba builds the whole stack: AI chips, and a target of 20GW of data-center capacity by 2032. ⚡

⚠️ Reality check: these are Alibaba’s own benchmark claims. Independent verification is still pending.

The gap between US and Chinese AI labs keeps shrinking. Can American leaders hold their edge? 👇

#technews #ArtificialIntelligence #OpenWeights

Archaeologists uncover a witness to the Neolithic revolution

Radiocarbon dating and isotope analysis of an immaculately preserved skeleton from Mesolithic Denmark indicate that the individual lived through the introduction of agriculture and adapted his diet to the change, providing a unique “eyewitness account” of one of humanity’s most significant technological revolutions.

About 6,000 years ago, a biologically male individual died and was buried in a shell midden on the island of Sejerø, Denmark. Nicknamed “Rødhals Man,” he was one of the indigenous foragers who lived on the island, subsisting by fishing, hunting marine mammals and gathering marine resources such as shellfish.

However, this life was not to last, as irreversible change came with the arrival of Neolithic farmers from the south.

Chromatin accessibility trajectories reveal pre-cancerous epigenetic priming and an acid-driven immunotherapy-resistant gastric cancer subtype

Wang et al. build a chromatin accessibility atlas of 650 gastric tissue samples spanning pre-cancer to advanced cancer, revealing a cancer-like epigenetic priming state in metaplasia that licenses tumor initiation. Advanced cancers segregate into three epigenomic subtypes; the pseudo-gastric subtype resists immunotherapy via FOSL2-driven H+/K+ ATPase expression that acidifies the microenvironment and suppresses CD8+ T cells—a phenotype reversed by the acid blocker vonoprazan.

Histone lactylation scarring transmits metabolic and epigenetic memory across the gut-liver axis

Dong et al. show that ulcerative colitis remission leaves persistent LCN2+ macrophages with heightened glycolysis and H3K9la memory. Upon high-fat diet challenge, these cells contribute to gut-to-liver inflammatory transmission and increased susceptibility to steatotic liver disease through a chromatin-remodeling program, thus providing a mechanistic rationale for prior clinical observations.

VAPA-dependent lipid transport at ER-lamellar body contact sites promotes human lung surfactant formation

Grimanelli et al. identify VAPA as a key regulator of surfactant lipid trafficking in alveolar type 2 cells. VAPA interacts with ABCA3 at ER-lamellar body membrane contact sites, linking VAPA dysfunction to impaired lamellar body formation and syndromic lung disease in humans.

Robo-Advisor Performance and the Limits of Performance-Based Adoption Explanations

Abstract

Robo-advisors offer low-cost automated investment services, but their reported investment outcomes do not by themselves explain who uses them. This paper compares ten selected U.S. robo-advisory accounts using a dated Condor Capital Robo Report extract. One-, three-, and five-year trailing results are summarized for total portfolio, equity-sleeve, fixed-income-sleeve, and risk-adjusted returns (Sharpe ratios). The accounts did not show uniformly weak reported outcomes. All ten fixed-income sleeves exceeded the displayed Bloomberg U.S. Aggregate Bond Index reference in the three-and five-year windows, whereas total and equity returns generally fell below the S&P 500 reference. These index comparisons are descriptive: portfolio allocations, risks, tax exposures, and fee conventions are not fully matched. Three-year Sharpe ratios were higher than five-year ratios in the supplied snapshot, but overlapping trailing windows do not establish a temporal decline. Most importantly, the data contain no investor-level adoption observations or direct performance-adoption relationship. The paper therefore does not claim to explain low adoption empirically. Instead, it positions reported performance alongside earlier work on investor trust, digital financial literacy, preferences for human advice, and provider incentives to outline propositions for future testing. The contribution is a bounded performance comparison and a clear separation of observed investment outcomes from conjectured adoption mechanisms.

Robo-Advisors, Artificial Intelligence, Technology Adoption, Behavioral Finance, Retail Finance