Xiaomi's New Product Launch in November 2026 is recapped in a single video, covering the Xiaomi 18 Pro Series, new wearables, and the latest Mijia ecosystem products.
Matt Pocock held a roughly 46-minute live conversation with Pi author Mario Zechner and Armin Ronacher, discussing whether AI coding agents can produce high-quality code. Key takeaways include the idea that "verification is overrated" — having a massive number of tests doesn't equal system reliability — and that agents excel at tactical tasks while system design still depends on human judgment.
Kent C. Dodds attends the infra.nyc event at the WorkOS office in NYC, where engineers discuss systems and infrastructure challenges. infra.nyc is a curated monthly meetup community for software infrastructure builders with members from companies like OpenAI, Anthropic, Stripe, and Cloudflare.
A live broadcast announcement featuring @dabit3 and @xujennyc, tagged with Devin developers, inviting people to join the conversation.
The post discusses the announcement that ICML 2027 will be held in Rio de Janeiro, noting community reactions including safety concerns and reports of racism similar to feedback around ICLR 2026, while asking how conference locations are chosen and how these issues could be addressed.
LangChain is hosting a live virtual workshop on Agent Runtime on November 19th, covering how to build, deploy, and trace agents using LangSmith Managed Deep Agents.
At a roundtable forum at the 2026 Yunqiao Conference, Tang Feihu, Kimi’s head of developer relations, appeared dressed as rival DeepSeek’s “Whale Girl” character, sparking heated discussion in AI circles. Known online as “Kojima Minako,” Tang is a legendary programmer who dropped out of Harbin Institute of Technology after gaining admission without an entrance exam, swept programming competitions, worked at Google’s Shanghai office, and founded Andromeda Technology.
Clement Delangue shares that he is at the DealBook event with Andrew Ng to advocate for open science and open-source AI, calling it more important than ever.
A transcript of Manus co-founder Zhang Tao's talk at Peking University, covering why they scrapped their AI browser project, how Cursor and Claude 3.5 Sonnet's long-horizon planning and tool-calling capabilities pointed to an opportunity in general-purpose agents, and how they build user trust through transparent execution and session replay.
NASA's annual Space Apps Challenge hackathon drew a record 181,835 participants across 593 events worldwide, a 59% increase over last year, using NASA open data to tackle space and Earth science challenges.
At Interrupt NYC, LangChain founder Harrison Chase sat down with OpenRouter CEO Alex Atallah for a fireside chat covering the drivers behind growing model diversity, the Open Responses API standard, traffic patterns across open and closed-weight models, and a framework for choosing models by "entropy" of tasks.
A nostalgic blog post sharing scanned pages of the zine/program from the Defcon 5 California Caravan (1997), recounting the author's early DEF CON road-trip experiences and hacker-scene memories.
Shane Parrish shares his conversation with Jon McNeill, former President of Tesla and co-founder/CEO of DVx Ventures, covering topics like impressing Elon Musk and building Tesla's culture of speed.
A promotional post recommending a free one-hour masterclass by SpaceX AI engineer Lauren Tan on how to build a GrokBot agent team and autonomous loops from scratch, covering real-world role assignments, workflows handling thousands of PRs per month, a minimal self-check loop, and token-saving codebase guardrails.
Cerebras is hosting a 24-hour online hackathon on Nov 21–22 featuring Kimi K3.1 Pro and OpenClaw, with $5,000 in prizes across X, Reddit, and a TPS speed bounty, plus increased private-preview access for participants.
A researcher asks for advice on whether to contact the AC before the ICLR 2027 rebuttal window closes, since two of three reviewers haven't responded and the average score sits at 2.67 (3/3/2).
MIT OpenCourseWare offers a free IAP 2024 workshop teaching CAD through reverse-engineering examples, with materials and playlists covering Solidworks, Creo, Fusion 360, Inventor, Onshape, and NX.
In the SPLASH'26 keynote, Chris Lattner outlined the motivation and design of Mojo — a unified systems language with Python syntax, open-sourced under Apache 2.0, targeting CPUs and six AI accelerator architectures. Through zero-cost abstractions, linear types, and low-level access via MLIR, it aims to solve hardware fragmentation and unlock full performance.
A talk by Dave Abrahams on using mutable value semantics as a foundation for safe systems programming, allowing in-place modification while preserving predictable value behavior; only metadata is available as the transcript could not be retrieved.
LangChain is hosting a webinar this Friday with Harrison Chase discussing the future of AI agents, how they are built, and what's needed to make them reliable and scalable in production.