Stripe
Sea Expands Credit Book Past $10B, Integrates Stripe in Regional Push
Sea is embedding its SPayLater and ShopeePay products into the Stripe Business Network across Southeast Asia to drive growth across underbanked markets.
Stripe
Sea is embedding its SPayLater and ShopeePay products into the Stripe Business Network across Southeast Asia to drive growth across underbanked markets.
OpenAI · 45.3K views
OpenAI expects to build an automated AI researcher by March 2028, driving an accelerating feedback loop that will require decentralized safety shields and structural economic adjustments like universal basic compute.
OpenAI · 12.2K views
Combining artificial intelligence models with open genomic data allows terminal patients to bypass traditional clinical trial bottlenecks and design targeted multi-drug therapies.
OpenAI · 12.3K views
Ashe Magalhaes argues that automated AI agents and personal taste will replace traditional long-hour software engineering as models master production code.
Sequoia Capital · 132.3K views
Oak Lab co-founders Rich Sutton and Khurram Javed argue that static LLMs hit a wall on finite internet data, advocating instead for continual learning algorithms that continuously update model parameters from real-world experience.
Sequoia Capital · 8.8K views
By linking fragmented municipal data instead of deploying new surveillance hardware, Peregrine enables long-horizon AI agents to resolve cold cases and track local threats while preserving data sovereignty.
Sequoia Capital · 45.6K views
Reinforcement learning environments built by human domain experts are replacing crowdsourced labeling as the primary competitive moat for frontier AI models.
Sequoia Capital · 33.8K views
AI agents fail in domain performance because platforms discard trillions of interaction tokens instead of converting user edits and execution traces into fine-tuning signals.
Sequoia Capital · 74.2K views
Building a custom agent harness is necessary when tasks diverge from standard model training distributions, though core components like file-editing primitives must still align with provider-specific reinforcement learning patterns.
张小珺商业访谈录 · 2.8K views
基础设施公司因同质化供给无法维持高额利润,AI原生应用时代的真正巨头必然诞生于下一阶段的新平台。
视野环球财经 · 4.3K views
美联储褐皮书揭示数据中心建设已取代传统消费成为支撑美国GDP与就业的核心支柱,AI资本开支呈现“只能成功不能失败”的不可逆趋势。
Sequoia Capital · 27.2K views
Application startups can match top-tier artificial intelligence research on a budget by pairing open-weight models with domain-guided synthetic datasets and multi-provider serving infrastructure.
Sequoia Capital · 38.6K views
Anthropic is betting its AI ecosystem strategy on open protocols like MCP and modular execution harnesses, arguing that long-term enterprise adoption depends on standardized interoperability rather than closed vertical stacks.
Sequoia Capital · 25.2K views
Fully burdened unit economics for autonomous flights have fallen below human automobile delivery as the drone network expands toward 1 million daily orders.
新官场 · 166K views
分析认为倪岳峰去向大概率走下坡路并提早无缘2027年高层格局,反映出最高层对地方经济治理的极度不满与官场派系制衡。
新官场 · 312.3K views
刘鹤通过危机比较研究为最高层提供经济理论支撑,其操盘的供给侧改革与贸易谈判深刻塑造了近年来中国经济与大国博弈格局。
Sequoia Capital · 56.4K views
Startups and incumbents must transition from frontier APIs to proprietary post-trained models after reaching product-market fit to prevent catastrophic inference costs and build defensible moats against UI cloning.
Sequoia Capital · 19.4K views
Clay Co-Founder Kareem Amin proves that defying traditional software norms—charging for usage over seats and building for technical power users—drives faster enterprise growth than conventional founder playbooks.
Sequoia Capital · 31.8K views
Chai Discovery relies on massive compute scaling and streamlined architecture to collapse biological design timelines from nine months to nine days.
Sequoia Capital · 52.6K views
Betting the company on a high-stakes federal lawsuit proved to be the single decision that allowed Kalshi to survive regulatory blockade and dominate American event contracts.
Sequoia Capital · 33.6K views
Standard transformer architectures have hit a scaling ceiling because they lack test-time adaptation, requiring an automated research lab focused on custom GPU kernels to discover a self-improving alternative.
Sequoia Capital · 60.5K views
Open-weight models combined with domain-specific fine-tuning now outperform closed frontier APIs while lowering long-term compute costs and platform risk.
Sequoia Capital · 41.9K views
Matan Grinberg argues open-weight models will capture most software development traffic as enterprise engineering shifts to asynchronous 'dark factories.'
硅谷101播客 · 13.6K views
由于真实世界缺乏遥操作数据集,机器人行业将在2026年下半年摒弃纯黑盒端到端大模型,重新回归上下分层架构与软硬件协同路线。