←Search results

By relevance · Total 1168 results Clear search

Friday

  1. TMTPost · AGISelected

    Google Releases Gemini 4 Carbon, Topping the Artificial Analysis Intelligence Index but Not Yet Publicly Available

    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

    Matched fieldsBody

  2. TMTPost · AGI

    Kimi K3.1 Pro hands-on test: Why does quota consumption still spark controversy after the API price cut?

    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

    Matched fieldsBody

  3. TMTPost · AGI

    Gemini 4 Carbon's Top Ranking Fails to Mask Disappointing Earnings: The Bill for Google's AI Gamble

    Alphabet's third-quarter revenue rose 25% year-over-year to $127.58 billion, while net profit fell 25% year-over-year to $26.17 billion, free cash flow was negative more than $9 billion, and the stock fell nearly 10% in after-hours trading. Capital expenditures of $51.4 billion were mostly directed toward AI infrastructure, and CFO Anat Ashkenazi maintained the full-year 2026 capital expenditure range of $195 billion to $205 billion unchanged. Hours before the earnings release, Google launched Gemini 4 Carbon, which topped the Artificial Analysis Intelligence Index with a score of 59, but it has not yet opened a public API and is initially available only to some cybersecurity professionals.

    Alphabet's third-quarter revenue rose 25% year-over-year to $127.58 billion, while net profit fell 25% year-over-year to $26.17 billion, free cash flow was negative more than $9 billion, and the stock fell nearly 10% in after-hours trading. Capital expenditures of $51.4 billion were mostly directed toward AI infrastructure, and CFO Anat Ashkenazi maintained the full-year 2026 capital expenditure range of $195 billion to $205 billion unchanged. Hours before the earnings release, Google launched G

    Matched fieldsBody

  4. TMTPost · AGI

    Edge AI Daily Morning Brief (October 30): Intel Plunges 21%, Anthropic…

    Intel reported third-quarter 2026 revenue of $16.5 billion and earnings per share of 52 cents, beating expectations, but its fourth-quarter revenue guidance of $15 billion to $16 billion fell short of the market's expectation of about $17.1 billion, and its gross margin guidance dropped from 43% to 38%; its stock price plunged about 21% in after-hours trading on October 29. Anthropic filed for a Nasdaq IPO on October 26 (ticker ANT), with revenue of $16.286 billion in the first half of 2026, up about 1,300% year over year, and second-quarter operating profit of $974 million. The U.S. Department of Justice has opened inquiries into whether Nvidia's acquisition of Hugging Face for about $12.93 billion affects AI competition.

    Intel reported third-quarter 2026 revenue of $16.5 billion and earnings per share of 52 cents, beating expectations, but its fourth-quarter revenue guidance of $15 billion to $16 billion fell short of the market's expectation of about $17.1 billion, and its gross margin guidance dropped from 43% to 38%; its stock price plunged about 21% in after-hours trading on October 29. Anthropic filed for a Nasdaq IPO on October 26 (ticker ANT), with revenue of $16.286 billion in the first half of 2026, up

    Matched fieldsBody

  5. TMTPost · AGI

    Microsoft, Google, and Amazon Earnings Side by Side: The AI Infrastructure Race Reaches a Watershed

    The latest earnings reports from the three cloud providers Microsoft, Google, and Amazon show cloud businesses accelerating across the board: AWS third-quarter revenue was $45.98 billion, up 39.3% year over year; Azure posted $33.165 billion for the quarter, up 48%; Google Cloud revenue was $30.71 billion, more than doubling year over year; and their backlogs reached $578 billion, $677 billion, and $561 billion, respectively. Over the same period, all three posted negative free cash flow: Amazon had a net outflow of $42.36 billion over the past 12 months, Alphabet was negative $9.07 billion, and Meta had a net outflow of $2.12 billion, while only Microsoft remained positive at $18.6 billion. Market reactions diverged; Alphabet fell 12.64% in a single day on slowing advertising growth and its swing to negative cash flow, and the article argues that the competitive bottleneck is shifting from chips to power, land, and grid-connection approvals.

    The latest earnings reports from the three cloud providers Microsoft, Google, and Amazon show cloud businesses accelerating across the board: AWS third-quarter revenue was $45.98 billion, up 39.3% year over year; Azure posted $33.165 billion for the quarter, up 48%; Google Cloud revenue was $30.71 billion, more than doubling year over year; and their backlogs reached $578 billion, $677 billion, and $561 billion, respectively. Over the same period, all three posted negative free cash flow: Amazon

    Matched fieldsBody

  6. Unite.AI

    Cohere Completes First Close of Series E, Targeting Up to C$7 Billion in Funding at a C$35.7 Billion Valuation

    On October 29, 2026, Cohere announced the first close of its Series E financing, targeting up to C$7 billion in funding at a corresponding valuation of C$35.7 billion, led by a company under Schwarz Group, with Bell Ventures, Bullhound Capital, and the Government of Canada joining as new investors. Cohere said this is the largest private financing for a Canadian tech company, and the funds will be used to advance its model and product roadmap, verticalized AI solutions for regulated industries, and long-term compute infrastructure. The company also disclosed that as of Q3 2026, its annual recurring revenue grew 165% year-over-year, with global customers increasing to 300.

    On October 29, 2026, Cohere announced the first close of its Series E financing, targeting up to C$7 billion in funding at a corresponding valuation of C$35.7 billion, led by a company under Schwarz Group, with Bell Ventures, Bullhound Capital, and the Government of Canada joining as new investors. Cohere said this is the largest private financing for a Canadian tech company, and the funds will be used to advance its model and product roadmap, verticalized AI solutions for regulated industries,

    Matched fieldsBody

  7. Unite.AI

    OpenAI expands its misalignment monitoring system to cover training, evaluation, and external deployment

    On October 29, 2026, OpenAI announced an expanded multi-layer misalignment behavior monitoring system covering training, evaluation, and internal use, and, after modifications, extended to external deployment. The system monitors every reinforcement learning run and evaluation for models at least as capable as GPT-6 Astra, with a four-layer architecture that escalates step by step from activation probes to safety and alignment investigation agents. In independent red team testing conducted by METR and Redwood Research from July 21 to August 11, two client-controllable request headers that could disable the new agent monitoring were found, along with a compression event gap from May to July 2026 in which approximately 10% of eligible internal Codex traffic was not monitored.

    On October 29, 2026, OpenAI announced an expanded multi-layer misalignment behavior monitoring system covering training, evaluation, and internal use, and, after modifications, extended to external deployment. The system monitors every reinforcement learning run and evaluation for models at least as capable as GPT-6 Astra, with a four-layer architecture that escalates step by step from activation probes to safety and alignment investigation agents. In independent red team testing conducted by ME

    Matched fieldsBody

  8. Unite.AI

    Research reveals why AI agents fail silently, and how to measure such failures

    Unite.AI, reviewing multiple studies, points out that when LLM agents make errors they often still return fluent answers and success status codes, making them hard for conventional monitoring to detect. A prompt-format experiment at ICLR 2024 showed that merely changing delimiters, spaces, and capitalization caused LLaMA-2-13B's accuracy on the same task to fluctuate by as much as 76 percentage points; in τ-bench, a GPT-4o agent solved about 61% of retail tasks in a single attempt, but the proportion of 8 consecutive fully correct runs dropped to about 25%. The article recommends putting prompts and tool descriptions under version control, repeating tests multiple times, recording the tools' actual returns in traces, and validating LLM-as-a-judge scores with human annotations.

    Unite.AI, reviewing multiple studies, points out that when LLM agents make errors they often still return fluent answers and success status codes, making them hard for conventional monitoring to detect. A prompt-format experiment at ICLR 2024 showed that merely changing delimiters, spaces, and capitalization caused LLaMA-2-13B's accuracy on the same task to fluctuate by as much as 76 percentage points; in τ-bench, a GPT-4o agent solved about 61% of retail tasks in a single attempt, but the propo

    Matched fieldsBody

  9. Unite.AI

    Mintive Raises $7 Million to Provide Independent Identity and Permission Management for AI Agents

    Mintive has completed a new $7 million funding round led by Cowboy Ventures, bringing total funding to $11 million. The company says it currently handles 2.5 million agent calls per day, serves more than 50 paying enterprise customers, and has processed over 200 million cumulative calls in seven months. Its platform assigns each agent an independent identity, scoped permissions, and traceable operation records; underlying service credentials remain within the gateway, agents access tools through short-lived tokens, and administrators can revoke an individual agent's token separately.

    Mintive has completed a new $7 million funding round led by Cowboy Ventures, bringing total funding to $11 million. The company says it currently handles 2.5 million agent calls per day, serves more than 50 paying enterprise customers, and has processed over 200 million cumulative calls in seven months. Its platform assigns each agent an independent identity, scoped permissions, and traceable operation records; underlying service credentials remain within the gateway, agents access tools through

    Matched fieldsBody

  10. Unite.AI

    Coveo Acquires Y Combinator-Backed Product Data Startup Anglera

    On October 29, 2026, Coveo announced the acquisition of Anglera, a San Francisco-based startup backed by Y Combinator, folding the latter's AI agent-powered product data enrichment platform into the Coveo Commerce product line. The transaction amount was not disclosed. Anglera extracts product attributes from sources such as manufacturer specifications, PDFs, and images; each enriched value comes with citations and a confidence score, uncertain values are routed to human review, and it syncs bidirectionally with PIM, ERP, and e-commerce platforms. Citing a NielsenIQ estimate, Coveo said missing product attributes can reduce product discoverability by up to 30%; in a 30-day A/B test, roughly 300 SKU variants saw a 120% increase in organic landing page traffic and a 270% increase in AI-referred visits.

    On October 29, 2026, Coveo announced the acquisition of Anglera, a San Francisco-based startup backed by Y Combinator, folding the latter's AI agent-powered product data enrichment platform into the Coveo Commerce product line. The transaction amount was not disclosed. Anglera extracts product attributes from sources such as manufacturer specifications, PDFs, and images; each enriched value comes with citations and a confidence score, uncertain values are routed to human review, and it syncs bid

    Matched fieldsBody

  11. English text is being prepared.

    Dapeng P1 PRO MAX Eye-Tracking Edition Released: 120–200Hz Sampling Rate, Tracking Accuracy Within 1°

    Dapeng officially launched the enterprise-grade VR headset P1 PRO MAX Eye-Tracking Edition, adding eye tracking to its existing VR capabilities. It supports recording data such as the user's gaze position, with an eye-tracking sampling rate of 120–200Hz and tracking accuracy controlled within 1°. The device targets visual training, scientific research experiments, and gaze interaction development, and offers supporting software customization that can organize interaction tasks, data recording, and result display according to project requirements.

    Dapeng officially launched the enterprise-grade VR headset P1 PRO MAX Eye-Tracking Edition, adding eye tracking to its existing VR capabilities. It supports recording data such as the user's gaze position, with an eye-tracking sampling rate of 120–200Hz and tracking accuracy controlled within 1°. The device targets visual training, scientific research experiments, and gaze interaction development, and offers supporting software customization that can organize interaction tasks, data recording, a

    Matched fieldsBody

  12. The Elec

    Akamai Launches AI Inference and Agent Sandbox Services, Expanding Its Cloud Services Footprint

    On October 29, Akamai released beta versions of two products, Akamai Inference and Akamai Agent Sandbox, allowing users to run AI models on Akamai Cloud and securely execute code generated by AI agents. Akamai Inference supports 12 open-weight models, including OpenAI's gpt-oss-120b, Google's Gemma 4, and Nvidia's Nemotron 3 Ultra, and provides an OpenAI-compatible API. Agent Sandbox is based on WebAssembly isolation technology and restricts code access to computing resources, network connections, and system functions.

    On October 29, Akamai released beta versions of two products, Akamai Inference and Akamai Agent Sandbox, allowing users to run AI models on Akamai Cloud and securely execute code generated by AI agents. Akamai Inference supports 12 open-weight models, including OpenAI's gpt-oss-120b, Google's Gemma 4, and Nvidia's Nemotron 3 Ultra, and provides an OpenAI-compatible API. Agent Sandbox is based on WebAssembly isolation technology and restricts code access to computing resources, network connection

    Matched fieldsBody

  13. The Elec

    Alphabet's Q3 2026 cloud revenue grew 103% year-over-year, with free cash flow at negative $9.07 billion

    Alphabet announced its Q3 2026 earnings: revenue of $127.6 billion came in below analysts' expectations of $128.2 billion, while Google Cloud revenue grew 103% year-over-year to $30.71 billion. Net profit fell about 25% year-over-year to $26.17 billion, and free cash flow was negative $9.07 billion, further deteriorating from negative $5.9 billion in the previous quarter, with shares falling nearly 10% after hours. CFO Anat Ashkenazi said capital expenditures will increase significantly in 2027 due to AI technology development, and on the same day Alphabet released its new-generation model Gemini 4 Carbon.

    Alphabet announced its Q3 2026 earnings: revenue of $127.6 billion came in below analysts' expectations of $128.2 billion, while Google Cloud revenue grew 103% year-over-year to $30.71 billion. Net profit fell about 25% year-over-year to $26.17 billion, and free cash flow was negative $9.07 billion, further deteriorating from negative $5.9 billion in the previous quarter, with shares falling nearly 10% after hours. CFO Anat Ashkenazi said capital expenditures will increase significantly in 2027

    Matched fieldsBody

Thursday

  1. Analytics India Magazine

    Sarvam AI deploys Indus Cowork to 2,500 government officials in Maharashtra

    Sarvam AI's Indus Cowork, launched on October 28 by Chief Minister Devendra Fadnavis, is being deployed to 2,500 government officials in Maharashtra for drafting documents, analyzing reports, summarizing files and retrieving government records. The state approved Sarvam's sovereign AI proposal in August with a two-year budget of 1.126 billion rupees; besides Indus Cowork, it also includes Samvaad, a voice and WhatsApp system for citizen communication, and connects the Saaras speech recognition and Bulbul text-to-speech models to the state's MahaAI platform. The platform supports 22 languages including Marathi, and offers tools such as GR Expert for retrieving government resolutions, email summarization and RFP bid evaluation; Sarvam says the system runs on securely audited infrastructure located within India.

    Sarvam AI's Indus Cowork, launched on October 28 by Chief Minister Devendra Fadnavis, is being deployed to 2,500 government officials in Maharashtra for drafting documents, analyzing reports, summarizing files and retrieving government records. The state approved Sarvam's sovereign AI proposal in August with a two-year budget of 1.126 billion rupees; besides Indus Cowork, it also includes Samvaad, a voice and WhatsApp system for citizen communication, and connects the Saaras speech recognition a

    Matched fieldsBody

  2. Analytics India Magazine

    The economics of continuous inference: from cost per token to cost per task

    AI inference is shifting from a one-off computational event to a continuous production workload, whose cost depends on traffic patterns, latency targets, model size, context length, and concurrency. The Stanford 2025 AI Index shows that the query price for GPT-3.5-level performance on MMLU fell from $20 per million tokens in November 2022 to $0.07 per million tokens for Gemini 1.5 Flash-8B in October 2024, a decline of more than 280x in about 18 months. But reasoning models and agents generate more tokens, call models multiple times, and occupy accelerator memory for extended periods, increasingly separating cost per token from cost per completed task; production environments should measure cost per successful task at an agreed service level.

    AI inference is shifting from a one-off computational event to a continuous production workload, whose cost depends on traffic patterns, latency targets, model size, context length, and concurrency. The Stanford 2025 AI Index shows that the query price for GPT-3.5-level performance on MMLU fell from $20 per million tokens in November 2022 to $0.07 per million tokens for Gemini 1.5 Flash-8B in October 2024, a decline of more than 280x in about 18 months. But reasoning models and agents generate m

    Matched fieldsBody

  3. gihyo.jp

    OpenAI Offers Up to $500 Monthly API Credit for ChatGPT Pro and Business

    Starting October 28, 2026, OpenAI will provide eligible ChatGPT Pro and Business subscribers with up to $500 per month in API credits, usable for app and agent development. Pro's three monthly price tiers of $100, $200, and $500 correspond to credits of the same amount; Business accumulates at $20 per month per Standard seat and $100 per month per Premium seat, with a workspace-wide cap of $500. Annual Business plans receive up to $6,000 in credits for the full year at once, initially prorated based on the remaining days in the billing cycle, with unused portions expiring at the end of the period. These credits cannot be used to call Codex via API key, nor do they increase the usage limits of ChatGPT or the logged-in version of Codex; users must link an API organization in billing settings and enable it manually.

    Starting October 28, 2026, OpenAI will provide eligible ChatGPT Pro and Business subscribers with up to $500 per month in API credits, usable for app and agent development. Pro's three monthly price tiers of $100, $200, and $500 correspond to credits of the same amount; Business accumulates at $20 per month per Standard seat and $100 per month per Premium seat, with a workspace-wide cap of $500. Annual Business plans receive up to $6,000 in credits for the full year at once, initially prorated b

    Matched fieldsBody

  4. Unite.AI

    Solink Secures Over $100 Million in Funding Led by Vertu Capital to Accelerate AI Vision…

    Ottawa-based Solink has announced the completion of over $100 million in new funding, led by Vertu Capital with participation from Goldman Sachs Alternatives' Growth Equity, among others. The funds will be used to accelerate its AI Vision Intelligence deployment. Its platform transforms existing security cameras into real-time business intelligence, covering more than 35,000 customer sites across 65 countries and managing 500,000 camera feeds. Solink previously completed a $60 million Series C funding round in 2023.

    Ottawa-based Solink has announced the completion of over $100 million in new funding, led by Vertu Capital with participation from Goldman Sachs Alternatives' Growth Equity, among others. The funds will be used to accelerate its AI Vision Intelligence deployment. Its platform transforms existing security cameras into real-time business intelligence, covering more than 35,000 customer sites across 65 countries and managing 500,000 camera feeds. Solink previously completed a $60 million Series C f

    Matched fieldsBody

  5. Unite.AI

    BBERT uses AI to identify hidden bacterial DNA in soil samples

    Researchers from the Hebrew University and other institutions have developed BBERT, which can identify bacterial sequences in DNA fragments as short as 100 bases, achieving about 90% accuracy in distinguishing bacterial from non-bacterial DNA in benchmark tests. The model is based on BERT and learns the context of nucleotide sequences, achieving about 88% accuracy in identifying protein-coding fragments and about 96% accuracy in determining reading frames. In simulated mixed samples, filtering with BBERT first reduced the average assembly time from about 66 hours to 25 hours while retaining about 96% average bacterial genome coverage.

    Researchers from the Hebrew University and other institutions have developed BBERT, which can identify bacterial sequences in DNA fragments as short as 100 bases, achieving about 90% accuracy in distinguishing bacterial from non-bacterial DNA in benchmark tests. The model is based on BERT and learns the context of nucleotide sequences, achieving about 88% accuracy in identifying protein-coding fragments and about 96% accuracy in determining reading frames. In simulated mixed samples, filtering w

    Matched fieldsBody

Wednesday

  1. gihyo.jp

    OpenAI DevDay Exchange Tokyo Recap: Practical Uses of dots and Codex

    OpenAI held DevDay Exchange Tokyo in Tokyo, with more than 500 developers attending, focusing on how to use dots, Codex, and the ChatGPT plugin ecosystem. Japan is the world's No. 2 market by weekly active users of dots and No. 1 in Asia-Pacific, and ChatGPT Enterprise seat count grew more than 2x year over year. GPT-6.1 Sol's input and output token unit price is 1/5 that of GPT-6 Astra, and it comes with an Ultrafast high-speed mode.

    Matched fieldsBody

  2. The New Stack

    AI agents need to be replaceable without losing approved business operations

    The New Stack published an article discussing how to safely take over work after an AI agent crashes: before dispatching a refund, the application first writes the amount, authorization, idempotency key, and status to transactional storage, and the replacement agent inherits the original operation rather than re-reasoning. The article points out that durable execution frameworks such as LangGraph and Temporal can restore progress, but cannot determine whether the payer has already submitted the request; therefore, persistent records must be validated at the tool boundary, and the MCP services of Langfuse and ClickStack should be used to read-only investigate calls with unclear results.

    The New Stack published an article discussing how to safely take over work after an AI agent crashes: before dispatching a refund, the application first writes the amount, authorization, idempotency key, and status to transactional storage, and the replacement agent inherits the original operation rather than re-reasoning. The article points out that durable execution frameworks such as LangGraph and Temporal can restore progress, but cannot determine whether the payer has already submitted the

    Matched fieldsBody