🚀 Zaprep: Your Socials on Steroids. 免费开始 — 每月自动发送 1,000 条私信，将互动转化为潜在客户。 ，每月 1,000 条自动私信。
最新 AI 资讯
Google Photos Could Soon Add a Map View to Ask Photos Search Results, Feature Spotted in APK Teardown
Google Photos might soon give Ask Photos a location-based way to display search results. A new map option has reportedly been discovered in a test version of the app, allowing photos found through location searches to appear according to where they were captured. The feature could let users explore groups of photos on a map, zoom into specific areas and open individual sets of images. However, Google has not announced the feature, and its presence in a test build does not guarantee a public rollout.
View
Binance now lets AI agents trade, but keeping them in check is largely up to users
Binance, the world’s largest crypto exchange with more than 300 million registered users, on Thursday launched a platform that lets AI agents analyze markets and execute trades on users’ behalf, bringing autonomous AI directly into the business of managing real money. CalledAgent OS, the platform lets developers connect AI applications and agents to Binance’s financial infrastructure. It brings the exchange’s existing tools and services such as Binance APIs, Binance Wallet Agentic Hub, Binance x402 transaction verification and payment facilitator API, and Binance Skill Hub, along with newly introduced support for itsModel Context Protocol (MCP). The platform also works with tools including OpenAI’s ChatGPT and Codex, Anthropic’s Claude Code, and Cursor, allowing users to authorize agents to access market data, view account information, and execute trades. However, as the AI race moves away from chatbots that answer questions to agents capable of taking action, Binance is putting much of the responsibility for keeping them in check on users, who ultimately have to decide what agents can access and trade and set limits on what they can do. “Instead of total freedom, we put the power in users’ hands to give them the granular access control of what they can do through the agent,” said Jeff Li, vice president of product at Binance, in an interview. “We put [the control] at the account level to protect the users’ funds.” Binance does that primarily through dedicated “sub-accounts”, which users can assign to agents and configure for specific activities, such as spot or futures trading. Withdrawals from those sub-accounts are blocked by default, Li told TechCrunch, creating a sandbox around an agent’s activity. Users can also choose whether an AI agent must seek approval for every order or can execute trades autonomously once its permissions are configured, a Binance representative said. Binance does not impose a separate cap on how much an AI agent can trade or lose, so the amount a user transfers into the sub-account effectively serves as the limit. Asked whether Binance can see what leads an agent to make a particular trade, Li said the reasoning happens outside its systems, either on the user’s computer or within their chosen AI application. “We really cannot see the reasoning of what the user’s action is,” he said. That means Binance can monitor an agent’s resulting trading activity, but has limited visibility into whether a decision was influenced by faulty information or manipulation. Li again pointed to the sub-account as the main line of defense when asked what would happen if an agent were manipulated through a prompt-injection attack or otherwise compromised. Binance also said its existing security, risk-control, and anti-money-laundering policies for subaccount APIs apply to Agent OS at launch. Trading is one of the first use cases Binance is targeting. Nonetheless, Li said agents could monitor markets, conduct research and risk analysis, react to signals, and autonomously place orders or execute strategies such as arbitrage. Agent OS is also designed to connect agents to payments and on-chain activity. Through Binance’s x402 integration, agents can send and settle payments, while its Agentic Wallet allows them to interact with tokens and decentralized-finance protocols. Unlike exchange trading, where Binance does not impose a separate cap on how much an agent can trade or lose within its subaccount, Agentic Wallet transactions carry Binance-set daily limits. Regular swaps are capped at $50,000 a day, DeFi transactions have a default $100,000 daily limit, and x402 payments are limited to $20 a day, according to the company. Li said Agent OS was Binance’s “first step” toward giving developers a platform to build AI-powered applications that can act across crypto and traditional markets. Binance is not alone in opening its infrastructure to AI agents. Rival crypto exchanges have been moving in the same direction, using MCP and other developer tools to give AI applications direct access to market data and trading systems. In March, Krakenlaunchedan open-source command-line tool with a built-in MCP server that allows AI agents to execute actions including spot and futures trades. Coinbase followed in June withCoinbase for Agents, which connects AI agents directly to users’ accounts and allows them to trade, make payments and execute other financial workflows within user-set limits. Similarly, OKXenabled agentic tradingon its platform by bringing an open-source MCP toolkit earlier this year.
View
OpenAI Brings GPT-5.6 Terra and Luna to Amazon Bedrock in India
The move gives Indian businesses access to in-country inference, with Luna targeting high-volume workloads at lower per-interaction costs.
View
Google Offers Indian Students One Year of Gemini AI Plus for Free
The offer also includes higher usage limits, study notebooks, Deep Research through Gemini Live and new AI-powered learning tools in Google Search.
View
OpenAI’s Codex Helps Asana Cut 5-Year Migration to 2 Weeks, Slashes Costs 500x
The project cost about $12,000 in model and infrastructure usage, compared to roughly $6 million that Asana had previously estimated.
View
AI is Killing the 8-Hour Workday. What Enterprises Will Measure Instead
AI is also challenging the assumption that spending more time on a task necessarily means creating more value.
View
KissanAI Is Obsessed With Getting Its Vision Model for Farmers Right
KissanAI wants to turn crop images into actionable intelligence that agricultural businesses can plug into their existing platforms.
View
Claude Can Now Take Greater Control Over Gmail, Google Drive With Updated Google Workspace Connectors
Claude users have had the option to connect their Google Workspace apps, including Gmail, Google Calendar, and Google Drive, for a while now. However, the connectors missed a few basic abilities. Now, Anthropic has updated the Google Workspace connectors for its AI chatbot, Claude. The update brings the ability to allow Claude to reply to email threads, send emails, and do more in Gmail. On top of this, with the updated Google Workspace connectors, Claude can now manage files and folders in Google Drive for users. This adds to the existing features that are offered with Google Drive, Gmail, and Google Calendar connectors in Claude.
View
From Kadapa to New York, and Now Raising Millions to Build Voice AI
Maya 2 supports 11 languages, each with a voice native to that language. The first-party API is 8x cheaper than Sarvam’s latest Bulbul v3 model.
View
Cognition CEO denies report that SpaceX tried to acquire the startup
Elon Musk’s SpaceX attempted to acquire AI coding startup Cognition as it works to catch up to OpenAI, Anthropic, and Google in the AI race,Bloomberg reported Wednesday, citing sources familiar with the matter. Cognition CEO Scott Wu disputed the report soon after it published, writing on X that the story was inaccurate and that Cognition “is not for sale,” adding that the two companies haven’t been in talks. The report comes a few days after SpaceX’s$60 billion acquisitionof Cursor, another AI coding startup, whose dealclosedlast week. SpaceX acquired Musk’s AI company, xAI, earlier this year. It then went public in ablockbuster IPO inJune, with its market capitalization rising to nearly $2.3 trillion at its peak. SpaceX has sold investors on its AI ambitions, which include eventually buildingdata centers in space. But the xAI business remains relatively early-stage and has fallen behind competitors. It’s also had to contend with its chatbot Grok’s penchant for controversy, including last year’s“MechaHitler”incident and this year’snonconsensual sexual imagery scandals, as it tries to win over enterprise customers. Last week,Musk told SpaceX’s employeesthat in about “four or five years, AI will be 99% of the value” of the company, but achieving that feat will require SpaceX to pull in much more revenue from AI. AI-assisted coding has emerged as one of the clearest ways to monetize the technology. Anthropic’s meteoric growth, fueled in large part by Claude Code, is proof of that. Bringing Cursor into the fold was part of that equation, and the companies were already working together before the acquisition closed. This month, Cursor and SpaceX jointly released Grok 4.6, a new model that scores higher on benchmarks for coding and complex multi-step agentic tasks. Adding Cognition and its coding agent Devin — along with an enterprise customer base that includes Mercedes-Benz, Citi, and Goldman Sachs — would have given SpaceX another way to deepen its push into AI coding and compete for enterprise customers. Bloomberg reports that the deal talks are no longer active, but that the companies are still discussing working together — potentially with Cognition using SpaceX’s computing capacity, which the company isselling to other AI playerslike Anthropic until it needs that capacity for itself. Wu didn’t address this specific claim in his denial. Cognition remains one of the largest independent AI software coding startups that hasn’t yet been gobbled up by a major AI model maker. The company in late May raised a$1 billion roundat a $25 billion post-money valuation, and Bloomberg reports it’s now in early talks for a new round of funding at a$40 billion valuation. Cognition made headlines last year when it acquired the remaining assets of competitor Windsurf afterGoogle DeepMind acqui-hiredthe startup’s CEO and top research in a $2.4 billion deal for talent and licensing rights. After the merger,Cognition laid off 30 employeesand offered buyouts to the remaining 200 Windsurf employees. Those who decided to stay faced strict operational expectations, like a more than 80-hour workweek and six days in the office. That sort of wartime work ethic wouldn’t be out of place in a Musk company. Musk has said he works up to 120 hours a week and often sleeps on office or factory floors. SpaceX and Cognition did not respond to requests for comment.
View
OpenAI seeks to one-up Anthropic with new customer privacy protections
As AI models have become more powerful, the potential for those models to be misused has grown — as has a clamor for safety guardrails that can stop such abuse from happening. AI companies must now walk a delicate tight rope between respecting their enterprise customers’ privacy while also watching usage for possible issues. Sensing an opportunity to one-up its rival Anthropic, OpenAI just announced a privacy-centric safety approach to monitoring for misuse. The companyis previewinga new service to select customers that it calls Private Safety Processing. This is an automated system that watches for potential abuse while simultaneously retaining none of the customer’s data. This system clearly runs counter to Anthropic’srecently announceddata-retention policy. The policy, which has aggravated some customers, enables the AI lab to keep user data (all of their sessions — and the conversations therein) for a period of 30 days, when it comes to “covered models.” Those models include all Mythos-class models and “future models with similar capabilities,” the company says. This policy, which was announced in July, was designed for the purposes of safety, allowing the lab to sift and analyze potential impropriety. However, ithas deeply concernedsome enterprises that handle large amounts of sensitive data and don’t want it harbored (or inspected) by the AI lab. OpenAI — like most other AI companies — already afford customers a relative level of privacy by adhering to a policy known as Zero Data Retention. ZDR uses agents within the OpenAI API to monitor for abuse on a per session basis. In this way, customer data isn’t retained by the company but companies are still able to scan for bad activity without the need for human intervention. It’s worth noting that Anthropic also largely abides by ZDR — except when it comes to “covered models,” like Fable. OpenAI says that Private Safety Processing is a new technology that widens ZDR’s scope. It describes it as a form of long-horizon safety monitoring that assesses the inputs and outputs of multiple conversations — not just one. Again, the monitoring is conducted by an agent, which, if triggered, catches interactions and analyzes them across sessions for signs of potential misuse. The new tech helps OpenAI detect malicious use of AI that takes place over multiple sessions, a spokesperson told TechCrunch. A bad actor — hypothetically someone trying to engineer malware for a cyberattack — may spread out their requests to avoid detection. Private Safety Processing can analyze those multiple conversations for signs of abuse without human review of a user’s conversations. In the case where the system is triggered, it may send a “narrowly defined signal” to OpenAI that warns of a specific type of activity, the company says. Based on that signal, OpenAI can then decide whether “enforcement is necessary,” it says. If so, OpenAI will reach out to the customer for more context or to work with them on the issue and a customer may choose to share data with OpenAI at their discretion, the spokesperson said. By contrast, Anthropic notes that human review of customer data can occur, but only “through a controlled access path” that involves “a small set of approved reviewers.” Every one of those review sessions is “recorded in a tamper-proof log that reviewers cannot suppress or modify,” the company says. The corporate competition between OpenAI and Anthropic is tense at the moment, with both companies looking for any opportunity to gain an advantage on the other. Arecent reportshowed that OpenAI’s Q2 grew more slowly than Anthropic. Anthropic’s annualized revenue run rateis now reportedly $65 billion. Anthropic investors have said itcould IPO at $2 trillion, while OpenAIis also working on its IPO.
View
Stripe didn’t really buy OpenRouter because of the ‘singularity’
Stripeconfirmedon Wednesday that it was buying OpenRouter. While the company didn’t disclose the deal price, sources told the New York Times that it paid $7.5 billion. That’s a huge step up from OpenRouter’s$1.3 billion valuation in May. To put that price in context, the founders alone will reportedly receive $1.5 billion from the sale — more than the startup’s entire valuation just three months ago. Investors will get the remaining $6 billion, according to the NYT. Stripe reportedly had to outbid others interested in the fast-growing startup,including Databricks. But the question is: what does a payments giant want with a startup that routes prompts between different AI models? The short and funny answer, according to a leaked letter from Stripe’s founders to its investors about the deal, is: the singularity. “It’s a fuzzy and perhaps already overworked term but we decided that January 1 marked the beginning of the singularity and we’ve been operating on that basis,” they wrote in the letter,published by Eric Newcomer, and verified by TechCrunch. The singularity is supposed to mean the point at which humans and the tech we’ve created merge to become a new species. This is obviously a tongue-and-cheek reference (as Patrick Collison admitted whenusing the term it at his company’sconference in April). We’re fairly certain Stripe’s founders, the brothers Patrick and John Collison, don’t think humanity started turning into The Borg eight months ago. But they have referred to the economic uptick that AI is bringing to Stripe. With AI, more companies are being launched and more of them are using Stripe’s offerings. Stripe says that 88% of the Forbes AI 50 are using its products, including OpenAI and Anthropic, as do100% of Brex’s fastest-growing startups.No one knows how AI and agents will change the economy of the future, but everyone is certain it will change it dramatically. That still doesn’t explain why Stripe wants a company mostly known for helping developers manage their model usage. Stripe’s founders acknowledged that their customer bases overlap. “OpenRouter is exceptionally useful for any developer and Stripe is one of the world’s largest developer platforms,” the founders write in their letter. No doubt that just using OpenRouter internally will probably offer significant benefits to Stripe and make it easier to roll out future model-agnostic agentic offerings, too. It seems as if OpenRouter will continue to operate independently after the deal closes in a few weeks, or so the startup promised inits own blog post, saying that its “product, mission, and current commitments remain unchanged.” Still, until now, most of Stripe’s large acquisitions have been related to helping people collect and manage incoming cash. Buying OpenRouter looks like a move to other side of the ledger, too: expense management, beginning with AI expenses. This acquisition “is Stripe’s deliberate attempt to embed itself into the middle of capital flows in the AI era,” said PitchBook’s research analyst Franco Granda. It’s joining an unusual assortment of companies also entering token expense management. Databricks developed its ownAI gateway.Rippling just launched one focused onemployee AI spend and ROI. Ramp just launched one,also for AI expense management. And the list goes on. For Stripe, buying the granddaddy of popular AI gateways for developers gives it insight into how coders are using AI. But it also gains a lever on AI demand itself. OpenRouter will grant it “some degree of power over suppliers such as the frontier labs themselves, as well as hyperscalers and neoclouds,” Granda said. It may not be the Borg, but payments plus token expense management and a model router? That’s a lot of power.
View
