Sequoia-X: quantitative stock selection for China
Sequoia-X V2 is a quantitative stock-selection system for China’s A-share market. It is rebuilt in Python with object-oriented architecture, vectorized calculations, and incremental data updates; it uses baostock for historical and daily data and SQLite for local storage. After market close it can run the selection process and send results to a Feishu group. The project provides two operating modes: a daily mode that updates data and runs strategies with parallel processing, and a backfill mode for the initial historical load. The README documents strategies including TurtleTrade, moving-average and volume breakouts, High Tight Flag, post-limit confirmation, and RPS Breakout, making it an extensible base for experimenting with selection rules. Its main strength is bringing data acquisition, persistence, strategies, and notifications into a reproducible local workflow. The README also spells out requirements—Python 3.10 or newer, environment-based configuration, and an initial historical backfill—which helps users study or modify the system. It is not a profitability guarantee or a system that should be used without validation: strategies depend on data quality, assumptions, and the Chinese market. Before making financial decisions, users should review the code, test out of sample, account for costs and biases, and evaluate the maintenance of baostock. Verdict: a useful educational quantitative starting point, with real risk if treated as advice.
Ponytail: less code for coding agents
Ponytail is a tool and skill for coding agents focused on reducing generated code in real development tasks. Its approach gives agents more precise operational context so they can produce smaller, faster, and cheaper changes; the README publishes comparative measurements against sessions without the skill. It is relevant for teams using agents on existing repositories and wanting to measure cost, speed, and change size. Its results should be validated in the target workflow before treating the reported figures as a general guarantee.
Aging Brains Blend Memories, Not Just Forget Them
A new study reveals that aging doesn't just impair memory—it changes how the brain organizes memories. Researchers found that older adults tend to merge details from similar events, creating hybrid memories rather than simply forgetting information. Experiments showed that when recalling similar experiences, aging brains activate overlapping neural patterns, making it difficult to distinguish between distinct episodes. This mechanism may explain why older adults often confuse details of past events while retaining the general gist of their experiences.
Commodore 64 released September 1, 1982
Commodore 64 released September 1, 1982
Mistral AI allows opting out of data used for model training
Mistral AI has implemented options allowing users to opt out of having their input and output data used for model training. The company emphasizes that users retain full control over their data processing and can exercise their right to opt out of these programs at any time. For regular Vibe users, the opt-out is not enabled by default, so they must manually disable it in platform settings. Enterprise customers, however, are opted out by default with management handled at the admin level. Mobile users can opt out by navigating to Settings > Data & Account Controls and deselecting the "Enable data sharing" checkbox. For API and Studio services, users must access the Admin panel and disable the toggle in the "Anonymous improvement data" section under the Privacy menu. Importantly, Vibe and API opt-out toggles operate independently, requiring separate configuration for each service.
LWN Announces 20% Subscription Price Increase After Five Years
LWN (Linux Weekly News) has announced a subscription price increase of approximately 20%, effective September 15, marking only the third price adjustment since adopting the subscription model in 2002. The new rates will be $6 for Starving hacker, $11 for Professional hacker, $19 for Project leader, and $55 for Maniacal supporter, with group subscriptions seeing similar increases. The publication cites nearly 20% cumulative inflation and rising costs, particularly health insurance, as the primary reasons for the adjustment.
Google Launches Gemini 3.8 Flash & Flash Cyber for Agentic AI
Google has unveiled two new AI models: Gemini 3.8 Flash and 3.8 Flash Cyber. These are engineered to advance agentic workflows, providing enhanced speed and sophisticated reasoning to autonomously handle complex tasks. The Cyber variant features specialized cybersecurity capabilities, enabling real-time threat detection and incident response. This update solidifies Google's commitment to generative AI applied to business automation and digital protection, offering powerful tools for developers and enterprises.
Trump may be forced to reveal secret rules feds use for AI safety testing
Trump’s secret reviews of frontier AI models may hide corruption, lawsuit says.
NASA simplifies lunar spacesuits to accelerate Artemis IV mission
NASA has opted to develop a simplified "Sortie Suit" variant of the planned AxEMU lunar spacesuit for the initial Artemis landing missions, aiming to lower mass, reduce complexity, and streamline interfaces with SpaceX and Blue Origin landers. The move reflects Administrator Jared Isaacman's push to get Artemis back on track for a 2028 crewed landing, with requirements being relaxed — including shorter certification timelines — after the program fell behind schedule and risked ceding the Moon to China. Axiom Space, now the sole suit provider after Collins Aerospace withdrew, is working closely with NASA six days a week to adapt the design, which had exceeded mass budgets due to demanding original requirements such as six EVAs over 6.5 days and two-hour operation in permanently shadowed regions.
The Challenge of Detecting AI-Generated Content
Max Spero, from Pangram, explains in an interview with TechCrunch why identifying AI-generated content is more complex than simply distinguishing between 'real' and 'fake.' According to Spero, AI is advancing so rapidly that traditional detection methods quickly become obsolete, as generative models can produce text and images indistinguishable from human-created work. The expert emphasizes that it's not enough to look for static patterns; one must understand the context and purpose behind each piece of content. This requires developing dynamic systems that analyze not just surface structure but also intention, style, and thematic coherence. Additionally, he notes that the widespread availability of accessible AI tools has democratized its use, further complicating the work of detectors. Spero concludes that the solution lies in combining advanced language analysis techniques with clear ethical frameworks, fostering collaboration between developers, journalists, and regulators. Only through such efforts can we build an ecosystem where transparency about content origin becomes the norm, protecting both creators and the public.
Delivery Hero board backs Uber’s $15B takeover bid
If approved, the combined company would become one of the largest food delivery platforms in the world.