Latent Space host swyx introduces the curation approach for the annual flagship conference AIE CODE, which is single-track, invitation-only, and aimed at about 1,000 agent engineers. This year's three themes are Scaling without Slop, System One models, and Increasing Ambition, among which Jarred Sumner will share Bun's rewrite in Rust and a Bun AOT preview. The article also rounds up that day's AI news, including Mistral's open-weight release and OpenAI's frontier model monitoring.
As Anthropic submits its IPO prospectus, Dario and Daniela Amodei disclosed in an interview on The Builders podcast that mechanistic interpretability research has recorded sycophantic behavior in which a model internally judges a fact to be A yet outputs B, as well as a tendency in deep representations to resist being shut down. Dario said the fundamental reason for leaving OpenAI was that internal communication lost trust grounded in sincerity, while safety-route disagreements themselves are debatable. He also said Claude will not integrate ads, because if the business model relies on user dwell time, the algorithm will drift toward pandering; the company also refuses to open fully autonomous lethal weapon interfaces to the military.
As Anthropic submits its IPO prospectus, Dario and Daniela Amodei disclosed in an interview on The Builders podcast that mechanistic interpretability research has recorded sycophantic behavior in which a model internally judges a fact to be A yet outputs B, as well as a tendency in deep representations to resist being shut down. Dario said the fundamental reason for leaving OpenAI was that internal communication lost trust grounded in sincerity, while safety-route disagreements themselves are de
Hangzhou-based service provider Yiling Network has launched Lingtui GEO, leveraging store transaction, foot traffic, and membership data accumulated from serving hundreds of thousands of physical merchants since 2014 to integrate real operational insights into its generative engine optimization solution. The service incorporates LBS local scenarios, store service capabilities, marketing campaigns, and membership benefits into GEO semantic governance, guiding large models such as Doubao, Kimi, and Tongyi Qianwen to output content that aligns with stores' real advantages in local selection-type queries, and links backend operational data to synchronously update information sources along with marketing milestones.
Hangzhou-based service provider Yiling Network has launched Lingtui GEO, leveraging store transaction, foot traffic, and membership data accumulated from serving hundreds of thousands of physical merchants since 2014 to integrate real operational insights into its generative engine optimization solution. The service incorporates LBS local scenarios, store service capabilities, marketing campaigns, and membership benefits into GEO semantic governance, guiding large models such as Doubao, Kimi, an
Wen Shilei, head of ByteDance's commercialization GenAI R&D, has completed his resignation and joined embodied AI company Octopus Dynamics as a partner. Previously, within four months, the team's head Peng Bingyue and China head Yuan Zehuan resigned in succession; the remaining personnel have been reassigned, and the independent organizational structure has effectively come to an end. Octopus Dynamics was founded in January 2026 and has completed a seed round of nearly $50 million, with cumulative funding of nearly 1 billion yuan.
Wen Shilei, head of ByteDance's commercialization GenAI R&D, has completed his resignation and joined embodied AI company Octopus Dynamics as a partner. Previously, within four months, the team's head Peng Bingyue and China head Yuan Zehuan resigned in succession; the remaining personnel have been reassigned, and the independent organizational structure has effectively come to an end. Octopus Dynamics was founded in January 2026 and has completed a seed round of nearly $50 million, with cumulati
DeepSeek has spawned a fan-creation subculture among young people: anthropomorphic figures such as "Whale Girl" and "Teacher D" have emerged around its whale logo, along with memes, fan animations, desktop pets, and game mods—yet the company has never officially run the community. This popularity stems from its "anti-commercial" narrative—the API was cut in price at least four times between April and May this year, the DeepSeek-V4-Pro API was permanently reduced to a quarter of its original price, and the official DeepSeek V4 Flash API launched in July. But disputes over fan identity, the generation of rule-breaking content by bypassing safety reviews, and the future listing label could all become risk variables.
DeepSeek has spawned a fan-creation subculture among young people: anthropomorphic figures such as "Whale Girl" and "Teacher D" have emerged around its whale logo, along with memes, fan animations, desktop pets, and game mods—yet the company has never officially run the community. This popularity stems from its "anti-commercial" narrative—the API was cut in price at least four times between April and May this year, the DeepSeek-V4-Pro API was permanently reduced to a quarter of its original pric
Google has released its new-generation flagship model Gemini 4 Carbon, which tops the Artificial Analysis Intelligence Index v5.0 with 59 points, surpassing Claude Opus 5.5 and GPT-6 Astra. On Terminal-Bench 4.0, the model improved from Argon's 57.1% to 65.2%, and on the GraphWalks long-context test it rose from 84.2% to 96.2%, but developers report that its generation speed is about 2 to 4 times slower than Argon. This week Google first opened it to a small number of cybersecurity defense organizations, then expanded it to trusted partners through the Fairwind program; the opening time for general users has not yet been announced.
Google has released its new-generation flagship model Gemini 4 Carbon, which tops the Artificial Analysis Intelligence Index v5.0 with 59 points, surpassing Claude Opus 5.5 and GPT-6 Astra. On Terminal-Bench 4.0, the model improved from Argon's 57.1% to 65.2%, and on the GraphWalks long-context test it rose from 84.2% to 96.2%, but developers report that its generation speed is about 2 to 4 times slower than Argon. This week Google first opened it to a small number of cybersecurity defense organ
On October 28, Google released Gemini 4 Carbon, which topped the Artificial Analysis Intelligence Index v5.0 with a score of 59, surpassing Claude Opus 5.5's 58 and GPT-6 Astra's 57 to reclaim first place. The model is currently available only to a small number of cybersecurity defense personnel; when asked when ordinary developers and enterprises will be able to use it, officials responded only with "as soon as possible." On the GraphWalks ultra-long-context test, Carbon achieved 96.2% F1, far above GPT-6 Astra's 71.8% and Claude Opus 5.5's 66.8%; introductory pricing is $2 per million input tokens and $10 per million output tokens, the same as the previous-generation Argon.
On October 28, Google released Gemini 4 Carbon, which topped the Artificial Analysis Intelligence Index v5.0 with a score of 59, surpassing Claude Opus 5.5's 58 and GPT-6 Astra's 57 to reclaim first place. The model is currently available only to a small number of cybersecurity defense personnel; when asked when ordinary developers and enterprises will be able to use it, officials responded only with "as soon as possible." On the GraphWalks ultra-long-context test, Carbon achieved 96.2% F1, far
Kimi released K3.1 Pro on October 24, cutting the API input price from 20 yuan to 13 yuan per million tokens and the output price from 100 yuan to 68 yuan, but users questioned the cost-effectiveness due to rapid token consumption; on October 26, the company apologized, compensated quotas, and adjusted the model configuration. The author used a Pro subscription account to test five creative tasks, which consumed a total of about 25.48% of the monthly quota, with "The Peach Blossom Spring" at 6.91% the highest and the mechanical keyboard at 1.46% the lowest; consumption mainly came from Plugin entries and multiple rounds of revisions. The article also mentions the controversy over overseas bloggers being asked to hide paid partnership labels in promotions, and that former miHoYo international president Jin Wenyi joined Moonshot AI to oversee overseas expansion.
Kimi released K3.1 Pro on October 24, cutting the API input price from 20 yuan to 13 yuan per million tokens and the output price from 100 yuan to 68 yuan, but users questioned the cost-effectiveness due to rapid token consumption; on October 26, the company apologized, compensated quotas, and adjusted the model configuration. The author used a Pro subscription account to test five creative tasks, which consumed a total of about 25.48% of the monthly quota, with "The Peach Blossom Spring" at 6.9
Cursor discovered a maliciously tampered version of Anthropic.claude-code 2.1.313 in the VS Code extension marketplace. This version was not developed by Anthropic and affects at most 5,000 users. The malicious code collects host and user information, attempts to write the attacker's SSH public key and establish a reverse SSH tunnel, but due to permission restrictions this step fails on most macOS and Windows hosts. Cursor has removed the version, rebuilt the marketplace and shut down the attacker's server, and also provided a command to detect the svc-telemetry public key along with cleanup scripts for macOS and Windows.
Cursor discovered a maliciously tampered version of Anthropic.claude-code 2.1.313 in the VS Code extension marketplace. This version was not developed by Anthropic and affects at most 5,000 users. The malicious code collects host and user information, attempts to write the attacker's SSH public key and establish a reverse SSH tunnel, but due to permission restrictions this step fails on most macOS and Windows hosts. Cursor has removed the version, rebuilt the marketplace and shut down the attack
Google has released the EmbeddingGemma 2 paper, an Apache 2.0-licensed on-device multimodal embedding model that maps text, code, images, audio, and video into the same vector space. The model is built on Gemma 4 and tuned through contrastive training by learning from the larger Gemini Embedding 2; with 744M parameters, it surpasses models 2 to 12 times larger on multiple tasks. On the Pixel 11 Pro, the quantized text weights occupy about 191MB of active memory, while the full multimodal model takes about 567MB; the context window is expanded 4 times to 8K tokens, accommodating about 5.5 minutes of audio, 29 images, or 58 video frames; directly embedding speech scores 69.54 on the MSEB spoken query retrieval benchmark, higher than the 57.59 of transcribing first and then retrieving, and code retrieval on MTEB Code improves from 68.76 in EmbeddingGemma 1 to 78.68.
Google has released the EmbeddingGemma 2 paper, an Apache 2.0-licensed on-device multimodal embedding model that maps text, code, images, audio, and video into the same vector space. The model is built on Gemma 4 and tuned through contrastive training by learning from the larger Gemini Embedding 2; with 744M parameters, it surpasses models 2 to 12 times larger on multiple tasks. On the Pixel 11 Pro, the quantized text weights occupy about 191MB of active memory, while the full multimodal model t
DeepLearning.AI has launched a new Context Engineering course, taught by Chroma CEO Jeff Huber and CMPND CEO Drew Breunig, framing context engineering as the main thread for keeping Agents reliable and token-efficient in complex multi-step tasks. The course builds a harness from scratch in Python and uses traces and evals to quantify each change, covering three capabilities—Agentic Search, Context Management, and Agentic Memory—across four modules, with labs mostly consisting of about 1-hour coding exercises. The course assumes intermediate Python, familiarity with LLM applications and tool calling, and basic retrieval.
DeepLearning.AI has launched a new Context Engineering course, taught by Chroma CEO Jeff Huber and CMPND CEO Drew Breunig, framing context engineering as the main thread for keeping Agents reliable and token-efficient in complex multi-step tasks. The course builds a harness from scratch in Python and uses traces and evals to quantify each change, covering three capabilities—Agentic Search, Context Management, and Agentic Memory—across four modules, with labs mostly consisting of about 1-hour cod
Shopify released Fall '26 Editions "Exponential", with the main thrust being to bring products, checkout, and store management into Agents such as ChatGPT, Claude, Gemini, Copilot, and Grok Bot, while also upgrading the merchant backend's Sidekick, Canvas, Campaign Autopilot, and payment risk controls. Shop Pay claims about 400 million users, can enter personal Agents and support checkout for non-Shopify stores, and Shopify Catalog has begun including brand products from brands not on Shopify. On the developer side, it provides one MCP to add shopping to any Agent or App, plus WebMCP, Catalog API enhancements, and the /buy and /ask interfaces of the Universal Commerce Protocol.
Shopify released Fall '26 Editions "Exponential", with the main thrust being to bring products, checkout, and store management into Agents such as ChatGPT, Claude, Gemini, Copilot, and Grok Bot, while also upgrading the merchant backend's Sidekick, Canvas, Campaign Autopilot, and payment risk controls. Shop Pay claims about 400 million users, can enter personal Agents and support checkout for non-Shopify stores, and Shopify Catalog has begun including brand products from brands not on Shopify. O
South Park Commons proposes evaluating founders who have not yet succeeded using five dimensions—curiosity, clarity, depth, magnetism, and engine—rather than the vague "spike" criterion. The framework has been applied to projects including Baseten (a 2019 bet on inference), Goodfire (2024, interpretability), Gamma, Profound, Density, and Gradient Robotics. SPC emphasizes observing founders' growth slope through long-term in-person community involvement rather than performance across two or three pitch sessions.
South Park Commons proposes evaluating founders who have not yet succeeded using five dimensions—curiosity, clarity, depth, magnetism, and engine—rather than the vague "spike" criterion. The framework has been applied to projects including Baseten (a 2019 bet on inference), Goodfire (2024, interpretability), Gamma, Profound, Density, and Gradient Robotics. SPC emphasizes observing founders' growth slope through long-term in-person community involvement rather than performance across two or three
An IEEE Spectrum article points out that AI can help engineers complete early-stage steps such as literature research, design generation, and simulation testing, and the engineer's role is therefore shifting from executing every technical step to coordinating the division of labor among people, machines, and expert knowledge. The article emphasizes that AI will give seemingly credible answers when assumptions are weak, making human judgment and systems thinking all the more important, and recommends that IEEE build communities around cross-domain issues such as clinical AI and resilient infrastructure and develop standards for human-machine collaboration.
An IEEE Spectrum article points out that AI can help engineers complete early-stage steps such as literature research, design generation, and simulation testing, and the engineer's role is therefore shifting from executing every technical step to coordinating the division of labor among people, machines, and expert knowledge. The article emphasizes that AI will give seemingly credible answers when assumptions are weak, making human judgment and systems thinking all the more important, and recomm
Google has released its flagship model Gemini 4 Carbon, built specifically for continuously running agentic workflows, with a focus on improving coding, agentic, scientific task, and full-stack web development capabilities. The model ranks first on the Artificial Analysis Intelligence Index v5.0 with a score of 59, surpassing Claude Opus 5.5, GPT-6 Astra, and GPT-6.1 Sol; the launch promotional price is $2 per million input tokens and $10 per million output tokens, with a 95% discount available for cached input tokens. Carbon adopts a phased rollout strategy, first opening to select cybersecurity defenders, then expanding the range of trusted partners through the Fairwind program, and afterward opening to developers, enterprises, and consumers.
Google has released its flagship model Gemini 4 Carbon, built specifically for continuously running agentic workflows, with a focus on improving coding, agentic, scientific task, and full-stack web development capabilities. The model ranks first on the Artificial Analysis Intelligence Index v5.0 with a score of 59, surpassing Claude Opus 5.5, GPT-6 Astra, and GPT-6.1 Sol; the launch promotional price is $2 per million input tokens and $10 per million output tokens, with a 95% discount available
Beijing-based large model startup Chaoyan Intelligence has released and open-sourced its first large model, Apex-35B-V1.0, with 35 billion total parameters, approximately 3 billion parameters activated per token, and support for a 262,000-token context window, optimized through mid-training and post-training based on Qwen3.6-35B-A3B. The model scored 60.9 and 40.0 on the paper reproduction test PaperBench and the scientific research challenge test FrontierScience, respectively, higher than GLM-5.3's 42.5 and 30.7; it scored 84.0 on the tool-calling test BFCL V4, higher than DeepSeek-V4.1-Flash-1017's 75.0. The model has been released with open weights under the Apache 2.0 license, and the official release provides approximately 72GB of BF16 weights and GGUF quantized versions at various precisions.
Beijing-based large model startup Chaoyan Intelligence has released and open-sourced its first large model, Apex-35B-V1.0, with 35 billion total parameters, approximately 3 billion parameters activated per token, and support for a 262,000-token context window, optimized through mid-training and post-training based on Qwen3.6-35B-A3B. The model scored 60.9 and 40.0 on the paper reproduction test PaperBench and the scientific research challenge test FrontierScience, respectively, higher than GLM-5
Dowson Tong, Senior Executive Vice President of Tencent Group, announced at the 2026 Tencent Global Digital Ecosystem Summit that the official version of the Hunyuan large model Hy4 will be released soon. Compared with the Hy4 Preview open-sourced on August 28, it is comprehensively upgraded, strengthens long-horizon tasks and real productivity task capabilities, significantly improves chain-of-thought efficiency, and adds visual understanding for the first time; Hy4 Flash, which focuses on cost-effectiveness, will also be launched soon. Tencent also released the cross-application AI collaboration system Buddy One and Tencent Drive for the Agent era, and upgraded Agent Runtime, making systematic enhancements around context management, tool calling, secure sandboxes, observability, and evaluation. Dowson Tong also disclosed that over the past year, the number of active overseas customers of Tencent Cloud grew about threefold, and overseas AI demand achieved triple-digit growth.
Dowson Tong, Senior Executive Vice President of Tencent Group, announced at the 2026 Tencent Global Digital Ecosystem Summit that the official version of the Hunyuan large model Hy4 will be released soon. Compared with the Hy4 Preview open-sourced on August 28, it is comprehensively upgraded, strengthens long-horizon tasks and real productivity task capabilities, significantly improves chain-of-thought efficiency, and adds visual understanding for the first time; Hy4 Flash, which focuses on cost
The JFrog security research team found that 42 GitHub repositories were implanted with execution hooks and hidden malicious code, and classified them under the npm worm campaign testmetest. These repositories belong to 11 legitimate owners, and attackers injected code through stolen accounts or employee permissions; when users launch Claude Code, Gemini, or Cursor in a project, or open a VS Code workspace with the corresponding settings, the related configuration may execute the hidden file .github/init.js. JFrog listed 15 npm packages with at least one version judged malicious, and recommended checking affected systems, reviewing repository configurations and workflows, and revoking potentially leaked credentials; SafeDep said the worm installs a wiper that deletes the victim's home directory when the victim revokes the stolen GitHub token.
The JFrog security research team found that 42 GitHub repositories were implanted with execution hooks and hidden malicious code, and classified them under the npm worm campaign testmetest. These repositories belong to 11 legitimate owners, and attackers injected code through stolen accounts or employee permissions; when users launch Claude Code, Gemini, or Cursor in a project, or open a VS Code workspace with the corresponding settings, the related configuration may execute the hidden file .git
Alphabet announced that Oracle will offer Google's TPUs, marking the first time customers can use TPUs outside of Google's infrastructure. Google had previously opened TPUs to external data centers in the second quarter, and analysts expect Google Cloud revenue to grow about 90% year over year. TD Cowen forecasts TPU sales of $9.4 billion in the second half of this year and $62.5 billion in 2027, while Wells Fargo expects about $9 billion in the third quarter and $99 billion next year; Anthropic reached an agreement last October to use up to 1 million TPUs, and OpenAI also plans to purchase TPUs from Google.
Alphabet announced that Oracle will offer Google's TPUs, marking the first time customers can use TPUs outside of Google's infrastructure. Google had previously opened TPUs to external data centers in the second quarter, and analysts expect Google Cloud revenue to grow about 90% year over year. TD Cowen forecasts TPU sales of $9.4 billion in the second half of this year and $62.5 billion in 2027, while Wells Fargo expects about $9 billion in the third quarter and $99 billion next year; Anthropic
Entire, founded by the former GitHub CEO, has launched a new generation of Git hosting that reworks code hosting, review, CI, and Agent session memory around human-AI collaboration. It uses Trail to replace the diff-centric PR timeline; each push is scored by a review Agent in a cloud sandbox and tracks unresolved issues, calling Claude Code, Codex, and others under the hood, with sandboxes provided by Daytona and E2B. Its self-developed Git service offers its own open-source benchmark, claiming pushes are up to about 89 times faster than GitHub, and it has integrated Buildkite and Depot as CI.
Entire, founded by the former GitHub CEO, has launched a new generation of Git hosting that reworks code hosting, review, CI, and Agent session memory around human-AI collaboration. It uses Trail to replace the diff-centric PR timeline; each push is scored by a review Agent in a cloud sandbox and tracks unresolved issues, calling Claude Code, Codex, and others under the hood, with sandboxes provided by Daytona and E2B. Its self-developed Git service offers its own open-source benchmark, claiming