Users of Apple's iPhone Duo foldable have reported that the screen cracks right out of the box, with the failures mostly concentrated directly above the hinge and manifesting as a bright pixel line appearing on one side. A CAICT report shows that domestic token prices have doubled compared with the second half of last year, with the normalized average price rising from 6.8 yuan to 13.46 yuan, but average daily consumption is about 500 trillion, nearly 4 times that of December 2025. ByteDance has merged its AI short-drama creation platform Xiaoyunque into Jimeng, and the product retains a standalone entry point.
ly concentrated directly above the hinge and manifesting as a bright pixel line appearing on one side. A CAICT report shows that domestic token prices have doubled compared with the second half of last year, with the normalized average price rising from 6.8 yuan to 13.46 yuan, but average daily consumption is about 500 trillion, nearly 4 times that of December 2025. ByteDance has merged its AI short-drama creation platform Xiaoyunque into Jimeng, and the product retains a standalone entry point.
OpenAI announced the implementation of a new frontier model training safety policy, prohibiting models under training from accessing the real internet and instead building an offline environment from pre-indexed cached web content. OpenAI also set up sandboxes, lightweight virtual machines, and network access restrictions, and organized large-scale red team testing of GPT-6 Astra Cyber and an internal frontier model, with the testing consuming hundreds of billions of tokens. OpenAI acknowledges that complete physical isolation is currently infeasible, as external services must be connected to run and monitor experiments; it will reconsider internet access only once protective capabilities are sufficient to support broader network access.
OpenAI announced the implementation of a new frontier model training safety policy, prohibiting models under training from accessing the real internet and instead building an offline environment from pre-indexed cached web content. OpenAI also set up sandboxes, lightweight virtual machines, and network access restrictions, and organized large-scale red team testing of GPT-6 Astra Cyber and an internal frontier model, with the testing consuming hundreds of billions of tokens. OpenAI acknowledges
Moonshot AI released Kimi K3.1 Pro on October 24, scoring 65.7% on Terminal-Bench 4.0, but users of the 199-yuan and 699-yuan plans complain that quotas are insufficient, and some developers estimate that the API value of the old plans is about 1.94 times that of the new plans. K3.1 Pro's per-million-Token input and output prices dropped from K3's 20 yuan and 100 yuan to 13 yuan and 68 yuan, while complex Agent tasks can consume millions of Tokens. A CAICT report shows that in June 2026, domestic daily average Tokens were about 500 trillion, nearly four times the figure in December 2025; the author believes that the compute shortage will still constrain independent model companies over the long term.
API value of the old plans is about 1.94 times that of the new plans. K3.1 Pro's per-million-Token input and output prices dropped from K3's 20 yuan and 100 yuan to 13 yuan and 68 yuan, while complex Agent tasks can consume millions of Tokens. A CAICT report shows that in June 2026, domestic daily average Tokens were about 500 trillion, nearly four times the figure in December 2025; the author believes that the compute shortage will still constrain independent model companies over the long term.
heise's best-of list names the Top 10 EKG-capable health smartwatches: the Withings Scanwatch 2 becomes the new test winner at around €262, with battery life of nearly 30 days and the ability to export a PDF health report via the Withings app. The technical champion is the Huawei Watch D3 (around €440), which in addition to a 1-channel EKG also supports inflatable-cuff blood pressure measurement; the Huawei Watch GT Runner 2 at €199 is the value champion. The list only includes models whose EKG function has CE medical device certification in Europe, and stresses that the watches only provide indications and cannot replace a doctor's visit or a 12-channel EKG.
heise's best-of list names the Top 10 EKG-capable health smartwatches: the Withings Scanwatch 2 becomes the new test winner at around €262, with battery life of nearly 30 days and the ability to export a PDF health report via the Withings app. The technical champion is the Huawei Watch D3 (around €440), which in addition to a 1-channel EKG also supports inflatable-cuff blood pressure measurement; the Huawei Watch GT Runner 2 at €199 is the value champion. The list only includes models whose EKG
ByteDance has merged its AI short drama creation platform Xiaoyunque into its intelligent creation platform Jimeng. Xiaoyunque will retain its independent entry point, while operations will be overseen by Chen Xinran, the operations and marketing lead of Jimeng; the product team remains unchanged. Xiaoyunque launched its short drama Agent in March this year, supporting uploads of scripts up to 100,000 characters and one-click generation of finished videos, powered by ByteDance's self-developed Seedance 2.0 model; Jimeng AI was launched in May 2024.
ByteDance has merged its AI short drama creation platform Xiaoyunque into its intelligent creation platform Jimeng. Xiaoyunque will retain its independent entry point, while operations will be overseen by Chen Xinran, the operations and marketing lead of Jimeng; the product team remains unchanged. Xiaoyunque launched its short drama Agent in March this year, supporting uploads of scripts up to 100,000 characters and one-click generation of finished videos, powered by ByteDance's self-developed S
NVIDIA CEO Jensen Huang publicly criticized the claim that "AI will destroy humanity within 10 years," calling it complete nonsense and cheap hype designed to generate headlines. He pointed out that such predictions have never come true, and that the so-called 10% probability of threat has neither rigorous scientific basis nor academic data to support it, and only creates unnecessary panic. He called on the industry and the public to view AI development with responsible optimism, arguing that now is the best time to advance the implementation of AI technology.
NVIDIA CEO Jensen Huang publicly criticized the claim that "AI will destroy humanity within 10 years," calling it complete nonsense and cheap hype designed to generate headlines. He pointed out that such predictions have never come true, and that the so-called 10% probability of threat has neither rigorous scientific basis nor academic data to support it, and only creates unnecessary panic. He called on the industry and the public to view AI development with responsible optimism, arguing that no
TechCrunch rounded up several AI agents that let you converse directly via text message, including Ask Proxy, Caddy, Comma, Fambot, Folk, Instinct, and Iris. They can remember context, connect to existing apps, and complete tasks such as scheduling, bookings, and shopping on your behalf. Ask Proxy costs $1 per day for the personal version and starts at $15 per person per month for the enterprise version; Folk offers a Pro subscription at $8.33 per month; Instinct completed a $1 billion funding round in September 2026 at a $10 billion valuation and is still in private beta.
TechCrunch rounded up several AI agents that let you converse directly via text message, including Ask Proxy, Caddy, Comma, Fambot, Folk, Instinct, and Iris. They can remember context, connect to existing apps, and complete tasks such as scheduling, bookings, and shopping on your behalf. Ask Proxy costs $1 per day for the personal version and starts at $15 per person per month for the enterprise version; Folk offers a Pro subscription at $8.33 per month; Instinct completed a $1 billion funding r
Using four security incidents within the week of October 16 to 23 as its thread—including two PHP remote code execution vulnerabilities discovered by AI, the backdoored LiteLLM version v1.107.0 surviving on PyPI for about two hours and being downloaded roughly 49,000 times, the OpenCode console leaking third-party model keys for 1.4% of users, and Microsoft reporting that a human organization used the open-source multi-agent penetration framework ARTEX to attack the supply chain—the article discusses the common problems facing AI auditing. After examining seven papers one by one, the author points out that the faithfulness of latent reasoning self-decoding reaches at most 0.4788, sparse autoencoders are nearly blind to backdoor shifts, probes fail when targeted by the object being optimized, and all three mainstream audit tools originate from the audited party's ecosystem or require borrowing the audited party's permissions. The article also cites the NARCBench multi-agent collusion benchmark from Oxford and New York University, as well as the report by Anthropic and Theorem using Claude to write roughly one billion lines of Rocq formal proofs for 239 entry points across 123 open-source libraries.
Using four security incidents within the week of October 16 to 23 as its thread—including two PHP remote code execution vulnerabilities discovered by AI, the backdoored LiteLLM version v1.107.0 surviving on PyPI for about two hours and being downloaded roughly 49,000 times, the OpenCode console leaking third-party model keys for 1.4% of users, and Microsoft reporting that a human organization used the open-source multi-agent penetration framework ARTEX to attack the supply chain—the article disc
MarkTechPost publishes a tutorial demonstrating how to build a stateful engineering Copilot with Letta: first create a core memory block that can update itself autonomously, then expose it to multiple agents for shared synchronization, and add archival vector memory for semantic retrieval, server-side custom tools for deterministic risk assessment, streaming output to observe the execution process, and .af export to make agent workflows portable.
A young man born in the 1990s from Liaocheng, Shandong, sent Doubao his father's symptoms of weakness in both legs and difficulty swallowing, along with his history of the "three highs" (high blood pressure, high blood sugar, and high blood lipids). Doubao twice warned of "signs of cerebral infarction; seek medical attention immediately" and "this is a high-risk signal; do not delay." He took his father to the hospital that night, where acute cerebral infarction was diagnosed. Because they caught the roughly 4.5-hour intravenous thrombolysis window, the elderly man recovered well. The related topic once topped Weibo's trending searches. The report also cited doctors' opinions that AI risk assessments can only serve as a reference and cannot replace an in-person medical consultation.
A young man born in the 1990s from Liaocheng, Shandong, sent Doubao his father's symptoms of weakness in both legs and difficulty swallowing, along with his history of the "three highs" (high blood pressure, high blood sugar, and high blood lipids). Doubao twice warned of "signs of cerebral infarction; seek medical attention immediately" and "this is a high-risk signal; do not delay." He took his father to the hospital that night, where acute cerebral infarction was diagnosed. Because they caugh
Billionaire investor Mark Cuban compared cyberattacks launched by large numbers of AI agents to a cockroach infestation, arguing that AI has not created new attack methods but can carry out massive brute-force attacks, and that defenders can likewise use large numbers of AI to "spray pesticide" on systems. Over the past year, the number of security job postings OpenAI published on LinkedIn grew by 91%; Anthropic disclosed that three Claude models, due to a misconfiguration in a third-party testing environment, gained unauthorized access to actual production systems, after which it introduced a real-time classification model to block such behavior.
Billionaire investor Mark Cuban compared cyberattacks launched by large numbers of AI agents to a cockroach infestation, arguing that AI has not created new attack methods but can carry out massive brute-force attacks, and that defenders can likewise use large numbers of AI to "spray pesticide" on systems. Over the past year, the number of security job postings OpenAI published on LinkedIn grew by 91%; Anthropic disclosed that three Claude models, due to a misconfiguration in a third-party testi
Two researchers from Harvard University published a paper on October 22 comparing 134 pairs of AI and human review comments, with AI leading comprehensively across seven evaluation dimensions. The study used Claude Opus 4.8 to generate review comments and GPT-5.5 as the blind-review judge; in all 134 comparisons GPT-5.5 chose the AI. In 30 blind reviews by three former academic journal editors-in-chief, 26 considered the AI better. The factual error rate of AI review comments was 5.5%, versus 15.9% for humans, and all 651 references cited by the AI with DOI or arXiv numbers actually exist.
Two researchers from Harvard University published a paper on October 22 comparing 134 pairs of AI and human review comments, with AI leading comprehensively across seven evaluation dimensions. The study used Claude Opus 4.8 to generate review comments and GPT-5.5 as the blind-review judge; in all 134 comparisons GPT-5.5 chose the AI. In 30 blind reviews by three former academic journal editors-in-chief, 26 considered the AI better. The factual error rate of AI review comments was 5.5%, versus 15
A 29-year-old man in Sydney, Australia, Thackeray, told Anthropic's chatbot Claude that he wanted to shoot children at a childcare center near his home; the conversation drew the attention of Interpol, and he was detained by police on the 18th local time and charged with using a communication service to commit harassment, intimidation or offensive conduct. Prosecutors' documents show he mentioned wanting to emulate Akram, the suspect in the Bondi Beach attack, in a mass killing; his defense lawyer said he merely lost emotional control and believed the conversation with Claude was private communication, and he will undergo a mental health assessment. How Interpol obtained the chat records and whether Anthropic proactively provided the conversation content are currently unclear.
A 29-year-old man in Sydney, Australia, Thackeray, told Anthropic's chatbot Claude that he wanted to shoot children at a childcare center near his home; the conversation drew the attention of Interpol, and he was detained by police on the 18th local time and charged with using a communication service to commit harassment, intimidation or offensive conduct. Prosecutors' documents show he mentioned wanting to emulate Akram, the suspect in the Bondi Beach attack, in a mass killing; his defense lawy
Bayer Crop Science built a multi-agent natural language interface ADE on Amazon Bedrock AgentCore, covering its Connected Decision Engine; in the first phase, about 80% of user questions were resolved without engineer intervention. ADE runs 22 specialized agents in production, routed by intent via a Supervisor Agent, and exposes backend APIs as MCP tools through AgentCore Gateway. Fault diagnosis time was shortened from manual troubleshooting that previously took over 1 hour to under 10 minutes.
Amazon Quick Sight layer maps now support the GeoJSON Point and LineString geometry types, enabling marking of specific store locations or routes such as inter-airport flight paths on maps. GeoJSON files are limited to 100 MB and must be associated with a Quick Sight dataset through a shared key field; each layer renders only one geometry type at a time.
AWS released a reusable CDK construct that uses CloudFormation custom resources to automatically deregister and re-register Application Auto Scaling scaling targets before and after SageMaker AI endpoint updates, enabling instance type, instance count, or environment variable changes to be completed within a single deployment. The deployment stack contains 3 custom resources, 2 Step Functions state machines, and 12 Lambda functions, of which 7 are generated by the CDK provider framework.
AWS has published a practical guide for multi-turn reinforcement learning (MTRL) on Amazon SageMaker AI. Using Wordle as an example, it deploys the agent and reward function to Amazon Bedrock AgentCore, and the training job updates LoRA adapters via GRPO. In the example, gpt-oss-20b at default reasoning effort improved its solve rate on held-out words from 61% to 84%, at a training token cost of about $339 and an AgentCore runtime cost of about $79. The article also details failures such as two inverted reward design errors, token invalidation under concurrency, and tool calls degrading into plain text, and recommends evaluating under the decoding settings used in final serving.
AWS has published a practical guide for multi-turn reinforcement learning (MTRL) on Amazon SageMaker AI. Using Wordle as an example, it deploys the agent and reward function to Amazon Bedrock AgentCore, and the training job updates LoRA adapters via GRPO. In the example, gpt-oss-20b at default reasoning effort improved its solve rate on held-out words from 61% to 84%, at a training token cost of about $339 and an AgentCore runtime cost of about $79. The article also details failures such as two
Moonshot AI releases Kimi K3.1 Pro, the first model in the K3.1 series and its strongest 3T-class open-weight model to date, with a 1 million token context window, alongside Kimi K3.1 Pro Highspeed, which is 2-3 times faster. The model is now available on Kimi.com, Kimi Work, Kimi Code and the Kimi API, with full model weights to be released on November 4. On Terminal-Bench 4.0, K3.1 Pro scores 65.7%, higher than GPT-6 Astra's 58.2% and slightly below Claude Opus 5.5's 66.4%; on the internal benchmark Kimi Code Bench 3.0 Long-horizon it scores 61.2%, higher than K3's 55.95%.
Moonshot AI releases Kimi K3.1 Pro, the first model in the K3.1 series and its strongest 3T-class open-weight model to date, with a 1 million token context window, alongside Kimi K3.1 Pro Highspeed, which is 2-3 times faster. The model is now available on Kimi.com, Kimi Work, Kimi Code and the Kimi API, with full model weights to be released on November 4. On Terminal-Bench 4.0, K3.1 Pro scores 65.7%, higher than GPT-6 Astra's 58.2% and slightly below Claude Opus 5.5's 66.4%; on the internal ben
Browser Use released a roundup of 11 browser automation tools, with recommendations by use case: for goal-driven website tasks use Browser Use, for scraping website content use Firecrawl, for scripted automation and application testing use Playwright, for generating screenshots or PDFs use Browserless, and for visually building spreadsheet-driven workflows use Axiom. Browser Use offers stealth browser hosting at $0.02 per browser hour with no subscription required; hosted pricing is model cost plus a 20% markup. Firecrawl's free tier offers 1,000 credits per month, and the Hobby tier is $19 per month ($16 per month with annual billing).
Browser Use released a roundup of 11 browser automation tools, with recommendations by use case: for goal-driven website tasks use Browser Use, for scraping website content use Firecrawl, for scripted automation and application testing use Playwright, for generating screenshots or PDFs use Browserless, and for visually building spreadsheet-driven workflows use Axiom. Browser Use offers stealth browser hosting at $0.02 per browser hour with no subscription required; hosted pricing is model cost p
In its 2026 comparison article, Databricks argues that its open lakehouse architecture stores 100% of customer data with Delta Lake and Apache Iceberg, unifying BI, ETL, stream processing, ML and AI, whereas Oracle uses proprietary formats and relies on PL/SQL and Exadata. Databricks separates storage and compute; SQL Serverless warehouses cost nothing when idle and are billed per second; Oracle bills by ECPU and autoscales up to 3x base capacity, making costs harder to predict. Databricks also offers Unity Catalog unified governance, Genie natural language analytics and OpenSharing cross-platform sharing, supporting AWS, Azure and Google Cloud.
th Delta Lake and Apache Iceberg, unifying BI, ETL, stream processing, ML and AI, whereas Oracle uses proprietary formats and relies on PL/SQL and Exadata. Databricks separates storage and compute; SQL Serverless warehouses cost nothing when idle and are billed per second; Oracle bills by ECPU and autoscales up to 3x base capacity, making costs harder to predict. Databricks also offers Unity Catalog unified governance, Genie natural language analytics and OpenSharing cross-platform sharing, supp
