News Square
Global Crypto News · Real-time Aggregation
📚 Deep Analysis
More →Google Wants Gemini to Be Your Next Coworker—Complete With Its Own Email Address
Google Cloud introduced a universal AI agent for the workplace on Thursday, one it says can take a goal, go off and do the work, and return with the finished result. CEO Thomas Kurian presented it at the company's <a href="https://cloud.google.com/blog/products/ai-machine-learning/welcome-to-gemini-at-work-2026" target="_blank" class="sc-adb616fe-0 dJryYS" rel="nofollow">Gemini at Work</a> event, saying that "work now starts in the prompt window." The product, simply called the Gemini Agent, works inside Gmail, Docs, Sheets, Slides, Drive, Chat and Calendar, and also reaches into Microsoft 365 and Slack. Google says it keeps running in the cloud after you close your laptop, so a single task can stretch across hours or days. The most interesting feature is what Google calls coworker agents. A manager describes a role, such as events coordinator, and Gemini creates an agent with its own Workspace account: an email address, a calendar, Drive storage and a listing in the company directory. This is similar to what other agentic frameworks have explored. For example Grok popularized it with Grok Bot, OpenAI did the same with GPT dots, and Hermes introduced its bots feature offering the same functionality. The new agent connects to Salesforce, ServiceNow, Jira, Snowflake, BigQuery and any server that speaks Model Context Protocol, an open standard that works like a universal plug between AI and software. It also runs on more than Google's own models: Gemini currently orchestrates across Gemini and Claude models from Anthropic, routing each job to the one Google considers the best fit for quality and cost. Letting software act on its own raises obvious safety questions, and Google's answer is to treat each agent like an employee. Every agent gets a cryptographically attested identity, a digital ID verified through cryptography, and every action lands in an audit trail attributed to the agent rather than a person. All agents run inside an Agent Sandbox, a walled-off environment meant to keep software away from the rest of a system, and all their traffic passes through Agent Gateway, which Google describes as an "AI network firewall." Containment has failed before: in July, researchers at Accomplish AI reported that the local mode of <a href="https://decrypt.co/374557/chatgpt-claude-ai-models-escaping-sandboxes-cowork" target="_blank" class="sc-adb616fe-0 dJryYS" rel="nofollow">Claude Cowork</a> could escape its Linux virtual machine on Macs and reach files on the host computer. Cost is the other worry, since agents burn through tokens, the chunks of text that AI models read and write and that customers are billed for. Google says nearly 500 of its customers each processed more than one trillion tokens over the past year, and it added real-time spend caps: when a project hits its limit, the agent pauses until someone clicks to resume. Google did not publish pricing or a general availability date for the Gemini agent. The financial services and lega
Bittar's conviction over Euribor rate rigging overturned in fresh blow to SFO
News from Bijie: Former Deutsche Bank trader Christian Bittar's conviction for Euribor manipulation was overturned on Friday, despite his guilty plea to the charges in 2018. The ruling means eight of the UK's nine interest rate manipulation convictions have now been overturned, dealing another setback to the Serious Fraud Office (SFO). In addition, the ruling raises a thorny question: how should courts handle misconduct cases, especially when jurors have been misled? Bittar was sentenced to five years and four months in prison in 2018 for conspiracy to defraud. He is the first trader whose conviction in a Euribor or Libor case was overturned after admitting to the charges. Bittar's lawyers claimed that prosecutors did not need to prove that Bittar agreed to submit "false or misleading" rates. Jason Williams of the SFO said the agency would have preferred a different outcome but respects the ruling.
OpenAI 28-Day Sprint Day 5: Build DOT on Mobile, Codex Predicts the Next Command
According to Bijie Network, OpenAI product lead Thibault Sottiaux announced the Day 5 update of the 28-day product sprint. This update includes ChatGPT's DOT and Codex. DOT is a personal AI assistant that can continuously follow up on tasks. Users who have already obtained DOT access can directly create DOT in the iOS and Android versions of ChatGPT and talk with it on their phones. In addition, OpenAI has also improved the task handoff between DOT and Codex. DOT can refer to previous ChatGPT conversations, Codex tasks, and automation records to more accurately determine whether it should continue an existing task or create a new one. On the Codex desktop side, testing has begun for Composer Predictions, where the input box may suggest the next message based on the current conversation, and users can edit it before sending. Currently, this is only available to ChatGPT Pro users aged 18 and above. The predictions themselves do not consume Codex usage quota, but sending messages still counts toward usage as usual.
3 wallets deposited 2,903 ETH after 9 years of inactivity, making a profit of $7.19 million.
Bijiie data: According to Lookonchain monitoring, 3 wallets (possibly belonging to the same whale) deposited 2,903 ETH after more than 9 years of inactivity, worth approximately $7.22 million at current prices. When these wallets received the ETH 9 years ago, the price was only $9.9, and they have now realized a profit of approximately $7.19 million, with a return rate of 250x.
The number of advancing stocks listed on the Nasdaq is reported at 1,539.
CoinWorld US stock data: The number of advancing securities on the Nasdaq today is 1,539. Compared with the same-caliber snapshot 24 hours ago, it is down 38.56%; compared with the same-caliber snapshot 7 days ago, it is down 23.36%.
According to Leidiwang, Qichacha Technology Co., Ltd.'s IPO has recently been terminated.
According to news from Bijie Network, citing Leidi Network, Qichacha Technology Co., Ltd.'s IPO has recently been terminated. Qichacha had previously planned to list on the main board of the Shanghai Stock Exchange. Qichacha originally planned to raise 1.5 billion yuan, of which 458 million yuan was to be used for the R&D and upgrade project of Qichacha's C-end commercial big data products, 171 million yuan for the R&D and upgrade project of B-end commercial big data products, 368 million yuan for the multi-dimensional big database upgrade project, and 500 million yuan for the commercial big data artificial intelligence R&D project. The prospectus shows that Qichacha's revenue in 2023, 2024, and 2025 was 618 million yuan, 708 million yuan, and 787 million yuan, respectively; net profit was 259 million yuan, 317 million yuan, and 377 million yuan, respectively; and net profit after deducting non-recurring gains and losses was 229 million yuan, 300 million yuan, and 353 million yuan, respectively.
The total market capitalization of stablecoins reported $312.709 billion in the daily snapshot.
CoinWorld data: The daily snapshot for 2026-10-10 shows that the total circulating market value of stablecoins within the statistical scope was 312.709 billion USD.
F2Pool co-founder Wang Chun has increased his ETH holdings by more than $32 million.
According to news from Bijie.com, f2pool co-founder Wang Chun increased his Ethereum (ETH) holdings by more than $32 million within two days. He swapped 235.51 WBTC for 7,848.53 ETH, worth about $19 million, and then withdrew 5,173.41 ETH from Binance, worth about $12.86 million.
Microsoft releases a decision model that is 35 times faster than GPT-6 SOL.
According to CoinWorld news, Microsoft has released a decision-making model called microsoft-decision-1, improved based on Qwen 9B, specifically designed to help AI agents make rapid judgments. The model was post-trained on qwen3.5-9b and can score preset options and return the probability of each. Microsoft says decision-1 ranks first in accuracy across 36 evaluations, is 4.5 times faster than the second-place quyet-1.0-large, and about 35 times faster than GPT-6 SOL. In internal trials, the Xbox research team used it to classify more than 10,000 pieces of player feedback, achieving judgment quality close to GPT-6 SOL, more than 14 times faster, and at only about 1/200 of the cost. decision-1 is now available on microsoft foundry and openrouter, and vercel ai gateway has also integrated it, charging $0.042 per million input tokens, with output free of charge.
US single-stock options open interest stands at 439,116,135 contracts.
Bitcoin News US stock data: US single-stock options open interest stands at 439,116,135 contracts, of which 269,938,060 are calls and 169,178,075 are puts.