CoinPulseHq

Shipt, the same-day delivery platform owned by Target, introduced its own AI shopping assistant on September 9, 2026, joining a wave of delivery apps racing to embed conversational AI into the grocery-buying experience. The new tool, called “Ask Shipt,” is available now in the Shipt app and on Shipt.com, according to the company.
Ask Shipt lets customers generate complete, ready-to-buy carts from natural-language prompts or photos. Shipt says users can request things like “Create a cart for my Saturday tailgate for 25 people and include some brunch items,” or upload a photo of a meal seen at a restaurant to have the assistant identify and add all the ingredients to a cart. Budget-conscious shoppers can also ask for ideas such as a weeknight meal for a family of five under $35.
Delivery apps are in an AI assistant arms race
Shipt’s launch comes as the broader delivery industry moves quickly to bake AI helpers into its apps. The same morning Shipt announced Ask Shipt, Instacart rolled out its own AI grocery assistant called Clementine. Uber Eats and DoorDash have also introduced comparable AI features earlier this year, signaling that conversational shopping is becoming a standard layer of the online grocery experience rather than a differentiator.
For Shipt, which operates as a standalone marketplace serving retailers beyond Target, the assistant is an attempt to make discovery easier — a persistent pain point in grocery e-commerce where shoppers often abandon carts because they don’t know what to buy or forget routine items. By converting vague prompts into concrete product lists, Ask Shipt shifts the app’s role from a passive catalog to an active shopping partner.
Photo-based cart building and the Target connection
One of the more distinctive features of Ask Shipt is its photo-recognition capability. A user who sees a dish on social media or at a restaurant can upload an image, and the assistant will parse the visual into a grocery list of ingredients. That feature overlaps with the AI-powered photo search Target has been rolling out on Target.com, along with AI-generated customer review summaries and other personalized shopping tools.
Shipt’s ownership by Target means the assistant also feeds into a broader retail AI strategy. Target has been integrating AI across its digital properties to improve product discovery and personalize the shopping journey, and Ask Shipt extends that push into the same-day delivery layer. Shipt is not exclusively a Target service — it also partners with other retailers — so the AI assistant is designed to work across the marketplace’s broader catalog.
What this means for shoppers and the future of grocery AI
For consumers, the practical benefit of Ask Shipt is reduced friction. Instead of manually searching for each item on a mental list, a single prompt can produce a complete cart in seconds. The budget-focused prompts also add a layer of price awareness, helping shoppers set constraints before the cart is built rather than discovering the total at checkout.
The launch also signals where the grocery delivery market is heading. With Instacart, Uber Eats, DoorDash, and now Shipt all offering AI assistants, the next competitive battleground is likely to be accuracy and personalization — how well the tools handle dietary restrictions, regional product availability, and repeat-order preferences. As these systems ingest more user data, the gap between generic suggestions and genuinely tailored carts will become the key measure of quality.
Shipt has not disclosed usage targets or a timeline for expanding Ask Shipt’s capabilities, but the tool is live immediately, positioning the company to gather user feedback while the AI-assistant category is still young.
This article is for informational purposes only and does not constitute financial advice. The technology and retail markets are volatile and evolving, and product features may change.
Originally published on CoinPulseHQ: https://coinpulsehq.com/shipt-ask-ai-shopping-assistant/

The German Federal Ministry of Finance has reportedly drafted a proposal to introduce a 25% flat-rate tax on cryptocurrency trading profits, a significant shift from the country's current policy that exempts crypto gains from taxation after a one-year holding period. The draft, seen by German newspaper Die Welt, suggests the new tax would apply to all digital assets acquired after January 1, 2027, with the new regime taking effect in 2028.
Grandfathering for existing holders
According to the draft proposal, the ministry plans to include grandfathering protections. This means that cryptocurrency purchased before the January 1, 2027 cutoff would continue to be treated under the existing rules, allowing long-term holders who acquired assets earlier to still benefit from the current tax-free status after 12 months of ownership. This transitional measure aims to avoid penalizing investors who made decisions based on the existing tax framework.
Under Germany's current income tax law, profits from the sale of private assets, including cryptocurrencies, are tax-exempt if the holding period exceeds one year. This has positioned Germany as one of the more tax-friendly jurisdictions for long-term crypto investors in Europe. The proposed 25% flat tax would align crypto gains with the country's standard capital gains tax rate, which already applies to other investment vehicles like stocks and funds.
Government revenue expectations and political context
Finance Minister Lars Klingbeil first signaled the planned crypto tax overhaul in late April, estimating that the change could generate an additional 2 billion euros (approximately $2.3 billion) in government revenue. The proposal comes as Germany's ruling coalition seeks new sources of income to address budget shortfalls and fund public investments.
The draft is still in its early stages and has not yet been formally submitted to parliament. The ministry has not publicly commented on the details beyond what was reported by Die Welt. Cointelegraph has reached out to the German Finance Ministry for further clarification.
This move is part of a broader European trend toward tighter cryptocurrency regulation. In recent months, Italy's central bank ordered sanctions screening for crypto transfers, and the European Union's Markets in Crypto-Assets (MiCA) regulation continues to shape how member states oversee digital assets. If adopted, Germany's tax change would represent one of the most consequential fiscal policies for crypto investors in the region, potentially influencing investment behavior and market dynamics across Europe.
What this means for crypto investors in Germany
For German crypto investors, the proposal introduces a critical planning window. Anyone acquiring digital assets before January 1, 2027 could still qualify for the current tax-free treatment after a one-year hold, provided the grandfathering clause remains intact in the final legislation. Those considering new purchases after that date would need to factor in a 25% tax on any future gains, regardless of holding period.
The proposal also signals a philosophical shift in how Germany views cryptocurrency — from a long-term investment vehicle to a taxable asset class akin to traditional securities. While the 25% rate is lower than Germany's top income tax rate, which can exceed 40%, it removes the incentive for ultra-long-term holding that previously existed.</n
Conclusion
Germany's draft proposal to impose a 25% flat tax on crypto gains from 2028 marks a notable departure from its historically lenient stance on long-term holders. With grandfathering protections for assets acquired before 2027, the policy aims to balance revenue generation with fairness to existing investors. As the draft moves through the legislative process, stakeholders in the crypto ecosystem will be watching closely for amendments and final details.
FAQs
Q1: When would the new 25% crypto tax take effect?
The German Finance Ministry's draft proposes that the tax apply to crypto assets acquired after January 1, 2027, with the new rate effective from 2028.
Q2: Will existing crypto holdings be affected?
Under the current draft, assets purchased before the January 1, 2027 cutoff would be grandfathered under the old rules, meaning they could still become tax-free after a one-year holding period.
Q3: Why is Germany changing its crypto tax policy?
The government expects to raise an additional 2 billion euros (about $2.3 billion) in revenue, and the move aligns crypto gains with the standard 25% capital gains tax applied to other investments.
Disclaimer: This article is for informational purposes only and does not constitute financial advice. Cryptocurrency markets are volatile and tax laws are subject to change. Readers should consult a qualified tax professional regarding their specific situation.
Originally published on CoinPulseHQ: https://coinpulsehq.com/germany-crypto-tax-proposal-2028/

蘋果於週三在其“Surprise and Shine”活動中宣佈，正在推出 Apple Reference Image（蘋果參考圖像）。這是一項旨在驗證在 iPhone 18 Pro 上拍攝的圖像是否真實的功能。該公司表示，隨着由 AI 生成和 AI 編輯的影像越來越難以與真實照片區分，“該功能對攝影記者和攝影師而言至關重要”。
Apple Reference Image 的工作方式是：在拍照的瞬間，從主攝像頭採集帶簽名的傳感器數據。隨後，這些數據會通過蘋果的 Private Cloud Compute（私有云計算）服務進行處理，從而生成一個可在“照片”應用中查看的“不可篡改的圖像”視圖。該參考圖像的作用類似於“數字底片”，讓用戶能夠將其與同一張照片的其他版本進行對比，以發現任何更改或編輯。

比特幣一度觸及盤中高點87,000美元，隨後回落至約85,000美元。Cointribune報道稱，這一走勢因槓桿倉位被強制平倉引發的浪潮而被放大。全加密市場範圍內，約10億美元的持倉在24小時內被清算，其中做空（空頭）佔據了總量的大部分，接近9億美元。
比特幣的上漲始於上週接近75,000美元的位置，而87,000美元這一關口自1月份以來便未被觸及。Cointribune報道稱，已有超過139,000名交易者被清算，最大單筆持倉規模超過2,000萬美元。比特幣單一數據表明，被清算的做空持倉爲4.54億美元，而多頭僅爲5,300萬美元，這意味着空頭佔最新數據中約90%。

Listen Labs，這家成立三年的AI市場研究初創公司，簽署了一份125百萬美元的B輪C輪融資（Series C）條款清單，估值爲15億美元，但該輪融資最終未能完成。根據多位知情人士的說法，公司放棄了這筆交易——這在風險投資中屬罕見舉動——轉而尋求與Salesforce的收購洽談。據稱Salesforce曾討論以約20億美元收購該初創公司。
原本由Menlo Ventures牽頭的融資在Salesforce介入後告吹。Business Insider最先報道稱此次收購討論；報道稱這些談判尚未敲定，可能不會達成交易。Listen Labs、Salesforce和Menlo Ventures均未迴應置評請求。

週三，蘋果發佈了其健康（Health）應用的重大改版，並同時推出新的 Apple Watch Series 12 和 Ultra 4。改版引入了由 AI 驅動的“洞察（Insights）”標籤、每日就緒度評分，以及一項新的“健康年齡（Health Age）”指標，用於將你的生物數據與按時間計算的年齡進行對比。此次重設計由 Apple Intelligence 提供支持，屬於蘋果更廣泛的戰略：將 iPhone 和 Apple Watch 打造成用於主動健康管理的核心樞紐。
最引人注目的變化是新的 Insights 標籤：它用一個動態信息流取代了靜態摘要視圖，從你的健康數據中呈現最及時的信息。根據蘋果的說法，該標籤將提供個性化指導、評估以及帶有語境的建議——例如，建議用戶在早晨跑步中增加更多間歇，以提升心血管適能。

Apple officially entered the foldable phone market on Wednesday, September 9, 2026, unveiling the Duo at its 'Surprise and Shine' event. While the device's form factor is a first for the company, the most significant engineering leap may be hidden inside its hinge, which Chief Hardware Officer Johny Srouji says was designed and manufactured with the help of AI and 3D printing.
Srouji detailed the process during the keynote, explaining that the hinge is a critical component for a device that undergoes significantly more stress than a traditional smartphone. To address the durability concerns that have plagued other foldables, Apple has implemented a manufacturing process that uses artificial intelligence to ensure near-perfect alignment and surface smoothness.
An AI-driven manufacturing process for a critical component
Foldable phones have struggled with durability since their inception, with hinges and screens often succumbing to wear and tear from repeated folding. Apple's approach to solving this on the Duo involves a highly precise, automated quality-control loop during production. According to Srouji, the process uses "AI algorithms to precisely match each individual hinge with its best-fit housing to ensure perfect alignment."
He elaborated that a "confocal laser progressively scans the topology of every single unit, and 3D prints up to 25 micro layers of a custom photopolymer to eliminate residual waviness." This level of individual calibration suggests Apple is treating each hinge as a unique component rather than relying on standard mass-production tolerances, which could be key to mitigating the friction and misalignment that cause foldables to degrade over time.
This focus on the physical build is notable for a company that typically emphasizes its silicon and software. The Duo also features a "custom nano-texture finish" designed to cut glare and a "multi-layer lamination strategy" aimed at boosting the screen's resilience against the repeated stress of folding.
What this means for the foldable market
Apple's entry into the foldable category validates a form factor that rivals like Samsung and Google have championed for years, but it also raises the bar for manufacturing precision. The company is entering a market where consumer enthusiasm has often been tempered by concerns over long-term reliability. By applying AI and 3D printing to the assembly line, Apple appears to be targeting the specific engineering pain points that have prevented foldables from becoming true everyday devices.
The success of the Duo's hinge will be a key test for Apple's manufacturing strategy. If the AI-calibrated components hold up in real-world usage, it could set a new standard for how premium devices are assembled. However, the long-term physical resilience of the device remains a critical question that only sustained usage will answer.
As with any new product category, the true test will come from consumers who use the Duo daily. Apple is betting that its investment in AI-driven precision manufacturing will translate into a foldable that finally matches the durability of its traditional iPhones. The company has not yet announced specific drop or cycle-test ratings for the device, leaving some technical questions open for independent reviewers.
This article discusses a new consumer hardware product and its manufacturing process. It does not constitute financial advice or a recommendation to purchase the device or related securities. The consumer electronics market is highly competitive and subject to rapid change.
Originally published on CoinPulseHQ: https://coinpulsehq.com/apple-duo-foldable-ai-hinge/

The US Senate returns to session this week with a narrow window to advance the Digital Asset Market Clarity (CLARITY) Act, a bill that many in the cryptocurrency industry see as a critical step toward establishing federal rules for digital assets. If the legislation fails to overcome a procedural hurdle, the path to becoming law could stretch into a new Congress with potentially different political leadership, delaying any resolution until at least 2027 — and possibly much longer.
Senate Majority Leader John Thune has scheduled a cloture vote on the bill for Tuesday, Sept. 10. Republicans will need at least 60 votes to break a filibuster, which means support from a handful of Democrats is essential. The chamber has less than 36 legislative days remaining before the current session ends in January 2027, when a newly elected Congress is sworn in. With midterm elections set for November, control of both chambers is in play, and the outcome could fundamentally alter the bill's trajectory.
The legislative clock and what a failure means
Senator Cynthia Lummis, a Wyoming Republican and one of the CLARITY Act's most vocal supporters, warned on Sept. 6 that if the bill does not pass this year, the next realistic opportunity may not arrive until 2030. Lummis is not seeking reelection in 2026, making her departure a notable loss for the bill's advocates in the next Congress.
Should the cloture vote fail, the bill could be reintroduced in the 119th Congress, but it would have to start the legislative process over. If Democrats gain control of the Senate or the House in November, they would likely rewrite the bill to include stronger consumer protections and stricter oversight provisions, which many crypto firms have resisted. A complete overhaul or a full stop to the legislation is possible, depending on the priorities of new committee chairs.
The CLARITY Act is not the only crypto-related bill on the table. The GENIUS Act, which establishes a federal framework for stablecoins, passed earlier in this Congress and was signed into law by President Donald Trump. That law benefited from Republican control of both chambers and the White House — a trifecta that could vanish after the midterms.
Crypto money and the 2026 election cycle
The stakes of the election are amplified by the significant financial involvement of the crypto industry. Political action committees such as Fairshake, which is backed by Coinbase and Ripple Labs, have spent heavily in primaries and general elections to support candidates seen as friendly to digital assets.
One prominent example is Ohio's special Senate election, where former Senator Sherrod Brown — a Democrat who previously chaired the Senate Banking Committee — is running to reclaim a seat against Republican Jon Husted. Brown was defeated in 2024 by Bernie Moreno, a race in which crypto PACs spent millions on ads opposing him. Now, Brown is back, and the industry is again pouring money into the contest.
Not all industry-backed efforts have succeeded. In March, Illinois Lieutenant Governor Juliana Stratton won the Democratic Senate primary despite being targeted by crypto-funded attack ads. In Massachusetts, Representative Jake Auchincloss, who voted for the CLARITY Act, received about $189,000 in support from a Fairshake-affiliated PAC during his primary race. His opponent, Jason Poulos, criticized the influx of industry cash, saying it gives "oligarchs" outsized influence over representation and federal policy.
What a Democratic-controlled Congress could mean for crypto
If Democrats take either chamber, they would gain the ability to set the legislative agenda on digital assets. Key committee positions would shift, and bills like CLARITY would likely face amendments aimed at strengthening investor protections, increasing transparency, and giving regulators more enforcement tools. Industry groups that have praised the current bill for its clarity might find a revised version less favorable.
Even if the CLARITY Act fails, the executive branch's approach to crypto is unlikely to change before January 2029. President Trump has nominated Paul Atkins to chair the Securities and Exchange Commission and Michael Selig to lead the Commodity Futures Trading Commission, both of whom have signaled they will continue using existing regulatory authority to address digital assets if Congress does not act. A presidential veto would also remain a powerful check on any Democrat-passed legislation, requiring a two-thirds supermajority in both chambers to override.
Why this matters for the crypto industry and investors
The outcome of the CLARITY Act vote and the November elections will determine the near-term regulatory environment for cryptocurrencies in the United States. A clear federal framework could reduce compliance costs for exchanges and issuers, while a prolonged legislative stalemate could leave the industry in a state of uncertainty, with state-level regulations filling the void.
For individual investors, the lack of a market structure law means the classification of many digital assets remains murky, potentially affecting everything from tax treatment to trading access. The SEC and CFTC have both pursued enforcement actions against crypto firms, and without new legislation, those agencies will likely continue to operate under existing, sometimes conflicting, mandates.
Investors should note that the cryptocurrency market is highly volatile and regulatory developments can cause significant price swings. This article is for informational purposes only and does not constitute financial advice. Readers should conduct their own research before making any investment decisions.
Conclusion
The CLARITY Act faces a decisive test this week, but its fate is intertwined with the broader political sector. A failure to pass would not only delay regulatory clarity but also open the door for a new Congress to reshape the bill — or abandon it entirely. With midterm elections approaching, the crypto industry's influence on Capitol Hill is being tested as much as the legislation itself. The next few months will be critical in determining whether the United States moves toward a comprehensive federal framework for digital assets or continues with a patchwork of state and agency-level rules.
FAQs
Q1: What is the CLARITY Act?
The CLARITY Act, formally the Digital Asset Market Clarity Act, is a US Senate bill that aims to establish a federal regulatory framework for digital assets, defining which tokens are securities and which are commodities, and assigning oversight to the SEC and CFTC.
Q2: What happens if the CLARITY Act fails the cloture vote?
If the bill fails to get 60 votes for cloture, it cannot proceed to a final vote in the Senate. It could be reintroduced in the next Congress, but it would need to start the legislative process over, and the new Congress may rewrite or drop the bill entirely.
Q3: How could the 2026 midterm elections affect crypto regulation?
If Democrats gain control of the Senate or House, they could push for a more consumer-protective version of the bill, add stricter provisions, or prioritize other issues. A divided government could also lead to gridlock, leaving the current regulatory uncertainty in place.
Originally published on CoinPulseHQ: https://coinpulsehq.com/clarity-act-failure-crypto-regulation-2026/

針對 Anthropic 的 Claude 模型所進行的蒸餾行動規模已經大到以“數億次”來計量。在週四發佈的一份報告中，Anthropic 表示，它觀察到大約 1.99 億次與未經授權的蒸餾攻擊相關的對話交換，分佈在五個獨立的行動中，並稱這些行動與中國境內的 AI 實驗室有關；其中最大的一項工作與阿里巴巴的 Qwen 模型家族有關。
Anthropic 表示這些行動看起來是怎樣的
蒸餾是一種標準的機器學習技術：訓練一個更小的模型，讓它學習更大模型的輸出，從而遷移推理能力。卡點在於同意。Anthropic 表示，它的模型內部的推理鏈路不會向用戶公開，並且它們的確發現了相關活動仍然設法提取這些痕跡。

Meta's new AI agent app, Muse, has climbed to the No. 2 spot on the US App Store's top charts, drawing more than 83,000 iOS downloads in the United States since its Tuesday launch, according to estimates from market intelligence firm Sensor Tower. The app is currently limited to US users, and its rise up the charts marks one of Meta's most ambitious pushes into consumer-facing agentic AI to date.
The milestone comes days after Meta agreed to an $18 billion multistate settlement over claims related to social media's consumer harms — a backdrop that could weigh on how quickly users adopt a product that requires handing over more personal information.
How Muse's debut stacks up against Meta's earlier launches
Muse's early numbers are modest when placed next to Meta's previous app debuts. Threads, the company's text-based Twitter rival, was downloaded more than 4.3 million times in the US on its launch day. The standalone Meta AI app saw 108,000 US downloads during its debut.
The comparison with ChatGPT is similarly uneven. In less than a week after its arrival, ChatGPT topped half a million installs in the US, which was its only market at the time. Divided evenly, that works out to roughly 83,300 downloads per day at its debut — a figure Muse took about twice as long to reach.
That gap does not necessarily signal a weak launch. Muse has been climbing steadily: it sat at No. 4 on the US App Store on Wednesday before moving up to No. 2. On Android, however, the picture is less flattering. The app has only reached No. 338 in the Productivity category on Google Play, and Android download figures are not yet available.
Sensor Tower's estimates also exclude usage through the web and WhatsApp, both of which Meta offers as access points for Muse. That means the true engagement picture is likely broader than the download data suggests.
Why Meta is betting on agents — and who else is racing
Muse represents Meta's wager that AI agents capable of completing tasks on a user's behalf will define the next phase of consumer AI. The company has framed the shift as comparable in strategic weight to its 2021 rebrand to Meta Platforms, when it pivoted toward the metaverse.
The competitive field is crowded. Google has pushed forward with Gemini Spark, Anthropic has Claude Cowork, and a wave of startups are targeting narrower use cases. Among consumer-facing agents, the most closely watched rival may be Instinct, an AI agent that operates over text messages and was recently valued at $2.5 billion.
Instinct has moved quickly. This week it rolled out email addresses for all users and announced it is building a social network in which one person's agent can coordinate plans with a friend's. It has also launched integrations with Stripe and 1Password, plus a location-sharing feature that lets the agent take actions requiring real-time location data. The company now has $350 million at its disposal.
Some observers argue that Instinct's approach — building a social graph around who people actually communicate with — could prove more durable than Meta's friend graph, which blends real connections with passive follows. Instinct has drawn scrutiny, too, over security concerns and a privacy policy that critics describe as broad and permissive.
What to watch next
Muse's trajectory over the coming weeks will be a test of whether Meta can convert its enormous existing user base into agent adoption. The app's climb to No. 2 suggests momentum, but the Android ranking of No. 338 indicates the company has significant ground to cover on Google's platform.
Meta will also face questions about data handling. The company has been fined multiple times by the US Federal Trade Commission over privacy violations and has weathered several large data scandals. How Muse handles user information — and how regulators respond — could shape adoption as much as feature quality does.
Meanwhile, Instinct's rapid shipping cadence and fresh capital give it room to expand its feature set before Meta can establish a lead. If agents do become the primary interface for consumer AI, the battle between Meta's distribution advantage and Instinct's social-graph strategy may be the contest to watch.
This article is for informational purposes only and does not constitute financial advice. Early adoption metrics and valuations in the AI sector are volatile and may change rapidly.
Originally published on CoinPulseHQ: https://coinpulsehq.com/meta-muse-ai-agent-app-store-downloads/

OpenAI 暫時停止了接受新訂閱者加入其每月 200 美元的 Pro 計劃。該舉措的產品負責人將其歸因於公司最新模型的需求已超出其背後基礎設施的承載能力。負責核心產品（包括 ChatGPT 和 Codex）的 Thibault Sottiaux 在 X 上宣佈了這一暫停，表示 Pro 這一檔位是在它售賣的所有計劃中，對 OpenAI 系統負載最重的一個。
"我們希望採取儘可能小的一步，來繼續在最大範圍內提供可及的訪問權限。"Sottiaux 寫道。現已禁用最高等級的註冊，而 ChatGPT 的 API 訪問以及較低成本的 Go 和 Plus 計劃仍對新客戶開放。

Complaints to the United Kingdom's housing ombudsman more than doubled between 2022 and last year, rising from 2,600 to just over 7,000, according to figures cited by researcher Chris Schmitz. Over the same period, the United States' Consumer Financial Protection Bureau saw its complaint volume grow roughly fivefold. Neither agency changed its remit, its staffing model, or its publicity strategy. What changed, Schmitz argues, is the ease with which people can now file a complaint at all.
Schmitz is tracking the phenomenon as part of a broader trend he calls agentic flooding — the rapid rise in applications, petitions, and filings that occurs once AI assistants make administrative tasks trivial to complete. His paper, set to be presented next month at the AI Ethics and Society conference, examines 84 potential cases of flooding across 11 jurisdictions, spanning welfare applications, official judicial appeals, and everything in between.
A pattern that predates every specific AI product
The cases Schmitz examined share a consistent shape. Submissions in each jurisdiction were roughly flat before 2022, then began rising at an accelerating pace as AI technology diffused into everyday use. Crucially, most of the 84 cases had not yet seen that growth slow down, suggesting the trend is likely to continue for years.
Alongside the UK housing ombudsman and the CFPB, Schmitz documented similar jumps in Brazilian judicial petitions and German parliamentary petitions. His full dataset is hosted publicly for other researchers to examine. For methodological reasons, the paper stops short of declaring AI the direct cause of the surge — but the correlation across jurisdictions and service types is difficult to attribute to coincidence.
Schmitz points to a simple mechanism. "People are finding out that this is something one can do, and incrementally, it is just getting easier to do it," he told TechCrunch. "Before it might have been a question of a lot of dragging context together and prompting ChatGPT 3.5 very precisely, it may now be a question of just pasting or taking a photo of a letter with your Claude app and getting a pretty good response in one shot."
The bug bounty parallel — and where it breaks down
The surge has a close analogue in the security industry. Last year, bug-bounty programs reported being overwhelmed by low-quality vulnerability reports generated by large language models. The reports rarely contained genuine security issues, yet companies were still obligated to triage each one, consuming significant engineering resources.
Public services face a superficially similar problem: the same budget, five times the applicants. But Schmitz's data points to a different conclusion. "The vast majority of cases we find are people who are entitled to claim for something, claiming for that thing," he told TechCrunch.
If those people were not claiming benefits or filing complaints before, it was often because the process itself was too forbidding. Policy researchers call this administrative burden — the friction of forms, deadlines, and documentation that quietly excludes people from services they are legally entitled to use. AI tools, in Schmitz's framing, are lifting some of that burden for the first time.
An opportunity rather than a crisis
That distinction matters for how governments respond. A service that treats every new filing as potential spam will invest in filtering, verification, and rejection. A service that treats the surge as evidence of previously suppressed demand might instead redesign its intake process entirely.
Schmitz leans toward the second approach. "A big part of making AI go well is being able to detail out what the good version of things looks like," he said. "And anyone who's ever used ChatGPT to do the tax return knows that there's a good version here where you're being helped. This could be the moment to say, 'we need to rethink pretty much everything about how this process looks.'"
The practical work, however, has barely begun. Most of the jurisdictions in Schmitz's dataset have not yet adjusted their processes, staffing, or digital infrastructure to account for AI-assisted filing. Omudbsman offices and consumer bureaus continue to operate on intake systems designed for a pre-2022 baseline.
What happens next depends partly on whether governments treat the increase as a resource problem or a design problem. The paper's presentation next month is likely to draw attention from civil servants and regulators already grappling with the same question — and Schmitz's data suggests the volume will keep rising regardless of which answer they choose.
Frequently Asked Questions
What is agentic flooding?Agentic flooding is a term coined by researcher Chris Schmitz to describe the rapid growth in applications, complaints, and petitions submitted to public services as AI tools make it easier to complete administrative tasks that were previously too burdensome to pursue.
Are AI-generated public service requests a form of spam?Schmitz's research suggests the opposite is more common: the vast majority of new filings come from people with legitimate claims who previously were deterred by the administrative burden of applying. A smaller share of submissions are clearly adversarial.
How much have complaints increased in the UK and US?Complaints to the UK housing ombudsman rose from 2,600 in 2022 to just over 7,000 last year, while the US Consumer Financial Protection Bureau saw complaint volumes grow roughly fivefold over the same period.
What should governments do about the surge in AI-assisted applications?Schmitz argues the increase is an opportunity to redesign public services to be more AI-friendly, rather than treating the additional volume purely as a resource problem to be filtered out.
Originally published on CoinPulseHQ: https://coinpulsehq.com/ai-agents-public-service-requests-agentic-flooding/

據經紀商於週四發佈的 2026 年 8 月月度運營報告，Robinhood 的加密貨幣交易量在 8 月環比增長 61%，達到 175 億美元；但總體仍比一年前同月低 38%。
這些數據凸顯出該公司加密業務的分化局面：環比動能正在改善，但由於交易所市場空間與零售交易活動持續轉移，同比分別仍然顯著爲負。
Bitstamp 取代 Robinhood 應用

Nvidia首席執行官黃仁勳（Jensen Huang）週四在高盛Communacopia + Technology會議上亮相時，重複了一項對幾乎任何規模達到這種體量的公司來說都將近乎“驚人”的數字：預計明年營收同比增長約70%。分析師預計，Nvidia將以當前財年約4000億美元收尾，這意味着明年的營收數字將接近6800億美元。
Nvidia首席執行官黃仁勳表示，公司明年有望實現約70%的營收同比增長，這意味着明年約6800億美元，而本財年預計約爲4000億美元。他指出，來自每一家主要AI實驗室和雲服務提供商的需求都在增長，並表示針對一套Grace-Blackwell系統的訂單正在按月增長27%。

Kakao Pay and KakaoBank have signed a memorandum of understanding with crypto infrastructure provider Fireblocks to test stablecoin and other digital asset infrastructure in South Korea, Cointelegraph reported on Sept. 22, 2026. Fireblocks said in its Sept. 21 announcement that the three companies will run proof-of-concept tests built around South Korean regulatory, security and service requirements.
The agreement does not announce a stablecoin, an investment amount, a commercial product or a deployment date, according to Crypto.news. The companies plan to assess infrastructure demand and possible digital asset businesses before deciding whether any framework advances beyond testing.
Key facts
• Kakao Pay and KakaoBank signed the MoU with Fireblocks; Fireblocks dated its announcement Sept. 21, 2026, and Cointelegraph published its report Sept. 22, 2026.
• The parties will run proof-of-concept tests covering South Korea's regulatory, security and service requirements, and will examine distribution frameworks before testing their practical use.
• Fireblocks says its platform is used by more than 2,500 institutions, including over 100 banks, and supports custody, settlement, stablecoin payments, tokenization, trading and compliance across more than 200 blockchains.
• Kakao Group signed a separate July MoU with Circle covering blockchain-based payment infrastructure and won-denominated stablecoin research.
• Fireblocks identified Kakao Pay CEO Shin Won-keun and KakaoBank CEO Yun Ho-young as co-heads of Kakao Group's Stablecoin Task Force, a detail Crypto.news reported.
What the agreement covers
Fireblocks described secure onchain infrastructure as the central technical area of the agreement, with stablecoins receiving specific attention. The announcement states that no single technical design has been selected publicly, and it does not specify a blockchain, token standard, reserve structure, custody model or consumer rollout plan.
Kakao Pay brings mobile payments and financial services to the project, while KakaoBank, one of South Korea's largest internet-only banks, provides the banking component. Both sit inside the wider Kakao ecosystem.
KakaoBank CEO Yun said the parties expect to combine their technology and expertise to "develop secure and accessible digital asset services," a statement that describes an intended direction rather than a confirmed product launch. Kakao Pay CEO Shin said Korea's developing digital asset market "depends on the reliable flow of digital asset distribution." Neither Kakao company disclosed whether a future stablecoin would be issued directly by a bank, another Kakao entity or an outside issuer.
Fireblocks CEO Michael Shaulov said infrastructure for Korean banks and payment platforms needs to be "engineered to meet institutional requirements from day one." The agreement does not state whether Kakao Pay or KakaoBank has committed to using Fireblocks in a production environment.
Fireblocks platform figures
Fireblocks says its technology has been deployed by more than 2,500 institutions, including over 100 banks. Separate data on the company's website says its network processes more than $200 billion in monthly stablecoin volume through more than 300 payment service providers, fintech companies and banks. Those are Fireblocks' own platform statistics and were not presented as Kakao transaction volumes.
The Circle agreement came first
The Fireblocks pact follows Kakao Group's July agreement with Circle, which the reports describe as covering stablecoin payments, blockchain settlement and digital asset infrastructure. Under that arrangement, Kakao said it would combine its consumer platform network, Kakao Pay's payment services, KakaoBank's banking operations and Circle's blockchain technology, with the parties discussing payment and settlement infrastructure, remittances and connections between blockchain networks and existing financial systems.
No won-denominated stablecoin was launched under the July MoU. Crypto.news reported that Kakao and Circle had not set a launch date or confirmed a particular issuance model, while Circle CEO Jeremy Allaire had previously said Circle did not plan to issue its own KRW stablecoin. Crypto.news also reported that the Fireblocks agreement introduces another infrastructure provider into Kakao Group's stablecoin research without replacing or ending the Circle arrangement, and that Fireblocks' announcement does not describe Circle's role in the new proof-of-concept tests or state whether the two relationships will share technology.
Why it matters
Kakao Pay and KakaoBank are among several South Korean financial and technology companies exploring stablecoin opportunities while the country develops its regulatory framework for digital assets. In May, KB Financial Group completed a won-denominated stablecoin pilot covering issuance, offline merchant payments and cross-border remittances, and in July fintech company Toss partnered with Optimism and Sunnyside Labs on a proof of concept for won-based stablecoin payment infrastructure.
The legal framework itself remains unfinished. South Korea's Financial Services Commission has said its planned framework law for digital assets will include stablecoins. The FSC said in August that discussions over the government's second-stage digital asset legislation were still underway and cautioned that some reported provisions had not been finalized, specifically rejecting claims that a proposed ownership cap for major crypto-exchange shareholders had already been settled. A Bank of Korea payment systems report published Sept. 17 said the central bank had created a Digital Asset Research Section after South Korea's Virtual Asset User Protection Act took effect, and that the unit has participated in legislative discussions concerning KRW-denominated stablecoins.
What to watch
The proof-of-concept results and any decision by Kakao Pay or KakaoBank to move a framework beyond testing are the next concrete milestones; the MoU sets no timeline for that. Movement on the FSC's digital asset framework law, including its stablecoin provisions, will shape whether bank- and payment-linked stablecoin work in South Korea can reach commercial deployment.
Originally published on CoinPulseHQ: https://coinpulsehq.com/kakao-pay-kakaobank-fireblocks-stablecoin-mou/

On Robinhood Chain, a memecoin called BONER and a tokenized version of healthcare stock Hims & Hers briefly became one of the most unusual trading pairs in crypto. The incident, first reported by Cointelegraph Magazine on September 11, 2026, illustrates what happens when real-world assets become composable DeFi instruments — and why the resulting markets can behave nothing like the ones they track.
The HIMS token is designed to mirror shares of Hims & Hers, the telehealth company listed on the New York Stock Exchange. On Robinhood Chain, users can buy and sell the tokenized stock alongside memecoins and other crypto assets. The BONER/HIMS pairing allowed traders to swap between the two tokens inside a single liquidity pool.
The imbalance that pushed HIMS to four times its NYSE price
At one point, the pool held 31,198 HIMS tokens — more than half of the 58,714 tokenized HIMS shares then in circulation. That concentration briefly sent the onchain HIMS token to $132.64, compared with a $28.84 close for the real stock on the NYSE.
The divergence was not a pricing error in the conventional sense. It was the predictable outcome of an automated market maker (AMM) doing what it is designed to do: set prices based on the ratio of assets in a pool. When reserves are thin and issuance is restricted, that mechanism can produce prices that bear little resemblance to the reference market.
Thomas Probst, a research analyst at Kaiko, framed the broader significance bluntly: "A listed stock effectively becomes a composable DeFi asset at an unprecedented scale, in the same way Ether did."
Why anyone would pair a healthcare stock with a memecoin
Robinhood Chain has become a testbed for stock-paired markets since its launch. In under three months, traders created pairings such as BONER/HIMS, AI/NVIDIA and SPACEHOOD/SPCX. The launchpad LONG reported that its stock-paired markets generated more than $425 million in 24-hour trading volume on September 2, with almost $12 million locked in stock-token liquidity.
From a DeFi perspective, the logic is straightforward. Traders do not need a fundamental reason to pair two assets — they need a market where they can swap between them. Once a tokenized stock exists onchain, it can serve as a quote asset, collateral, loanable inventory or margin for derivatives, said Angelo Aspris, a finance academic at the University of Sydney.
Sergej Kunz, co-founder of the DeFi aggregator 1inch, described the opportunity as larger than simply moving equities onto a blockchain: "This is not just about changing the venue. It is about creating an asset that can plug into an open financial system."
What the HIMS episode reveals about price discovery
The extreme deviation between tokenized HIMS and the underlying stock exposes a structural gap. In traditional markets, arbitrage is a continuous, competitive process that keeps related prices aligned. Onchain, as Probst noted, that link can depend on a single actor, and it breaks when liquidity is thin or the real-world market is closed.
Reid Noch, vice president of US equity market structure and electronic trading at TD Securities, remains skeptical that AMM pools will become the primary venue for discovering the price of tokenized stocks. "I still see price discovery happening more in traditional markets, and AMMs being used [by] arbitrageurs to keep the market in line," he said, adding that as long as these pools primarily drive liquidity in memecoins, traditional players will struggle to take them seriously.
Aspris warned that thin reserves and restricted issuance "create the conditions for these events" and increase the potential for strategic exploitation or manipulation. He also cautioned against overstating how far the experiment has progressed: calling tokenized equities a finished DeFi primitive would be ahead of the facts.
The bigger experiment: stocks as DeFi building blocks
The BONER/HIMS pairing is unlikely to be the template for institutional tokenized equity trading. But it does demonstrate that once a stock becomes a token, it no longer has to behave only like a stock. It can be combined with almost anything that has sufficient liquidity — other equities, cryptocurrencies, tokenized real estate, commodities or artworks.
Whether those markets emerge, gain traction or make economic sense is an open question. What the Robinhood Chain episode shows is that composability removes the technical barriers to trying. Traders will build the pairings that Traditional Finance would never have considered, and the resulting markets will occasionally produce prices that look absurd against their reference assets.
For now, these markets remain immature and isolated from traditional exchanges. Their long-term role — as a curiosity, a liquidity venue, or a genuine new primitive for onchain finance — is still being determined by the people trading them.
Conclusion
The BONER/HIMS pool on Robinhood Chain is a small, strange corner of DeFi, but it makes a larger point: tokenized stocks are no longer just digital representations of shares. They are programmable assets that can be plugged into markets their issuers never anticipated. That flexibility is precisely what makes them powerful — and what makes their prices, at least for now, unreliable.
FAQs
Q1: What is a tokenized stock?
A tokenized stock is a blockchain-based token designed to track the price of a publicly traded share, such as the HIMS token on Robinhood Chain tracking Hims & Hers on the NYSE.
Q2: Why did the HIMS token trade at $132.64 when the real stock was $28.84?
The pool held more than half of all circulating tokenized HIMS tokens, and thin reserves in an automated market maker pushed the onchain price far above the real stock's NYSE close.
Q3: Are stock/memecoin pairs a good investment?
These markets are experimental and largely isolated from traditional exchanges. They can produce unreliable price signals, and nothing here should be treated as financial advice.
Originally published on CoinPulseHQ: https://coinpulsehq.com/boner-hims-tokenized-stock-defi-robinhood-chain/

OpenAI will not go public in 2026, CEO Sam Altman said in an interview with Fortune editor in chief Alyson Shontell, telling her that the current climate around AI safety makes a listing this year "ill-advised." The comments, reported by TechCrunch on September 12, 2026, settle months of speculation about whether the ChatGPT maker would test public markets before the end of the year.
OpenAI will not go public in 2026. CEO Sam Altman said a listing now would be "ill-advised" given the state of AI safety, and confirmed the company is not targeting this year. He said OpenAI will go public when the business and the moment are ready.
Key facts
• Altman said it would be "ill-advised" to go public right now given everything happening with safety.
• He confirmed the IPO will not happen in 2026, saying OpenAI has "a lot of stuff to do."
• OpenAI has filed confidentially for an IPO.
• The New York Times reported in June 2026 that OpenAI was leaning toward 2027 rather than the third or fourth quarter of 2026.
• Altman told Fortune the company will go public "when we're ready," tied to the business and the broader societal moment.
What Altman actually said
In the conversation with Fortune editor in chief Alyson Shontell, Altman was asked whether OpenAI still feels pressure to move quickly because of its listing plans. He pushed back on the premise, saying the company is not rushing into an IPO.
The comment came amid fallout from a hack involving OpenAI and Hugging Face, as well as wider discussion about AI safety. Altman framed the decision as a matter of timing rather than one of capability or demand. "When we're ready, which is when the business is ready, when we feel ready from what the moment is like in society with this technology," he said, describing the conditions the board and leadership would weigh before listing.
Pressed directly on whether that meant no 2026 listing, Altman answered with a simple confirmation: not 2026, and noted the company has plenty on its plate before it would take that step.
A timeline that was never firm
The possibility of an IPO this year had already been in doubt. The New York Times reported in June that although OpenAI had hired bankers and lawyers with the goal of going public in the third or fourth quarter of 2026, the company was leaning toward 2027. The reasons cited at the time were volatility in tech stocks and OpenAI's own financial picture.
Altman's latest remarks narrow that window further, moving the question from "when this year" to a date the company will set on its own terms. For employees, early investors, and partners holding equity or convertible stakes, the deferral pushes an expected liquidity event out by at least a year in most scenarios it described.
Why it matters
OpenAI is one of the most closely watched private companies in the world, and a public listing would have been a landmark moment for the AI sector as a whole, giving retail investors direct exposure to a firm at the center of the generative AI boom. Delaying removes that event from 2026 and shifts attention to the company's private funding and governance.
The decision also puts safety concerns, not market conditions, at the center of the explanation. That framing matters for how regulators, enterprise customers, and competitors read OpenAI's posture heading into the next year.
What to watch
Investors and analysts will watch whether OpenAI's confidential filing progresses toward a formal registration statement, and whether the company and its backers provide any new signal on timing. The next concrete marker from the source reporting is the window the company itself described: a business that is ready, and a societal moment Altman considers appropriate for a listing.
This article contains no pricing or investment recommendation. Any decision about a company's public offering involves significant uncertainty, and this is not financial advice.
Originally published on CoinPulseHQ: https://coinpulsehq.com/openai-ipo-not-2026-sam-altman/

Attackers who obtained sensitive customer information from the fintech company Revolut have begun publishing the data online and are threatening to release more each day until the company pays, according to a report by Cointelegraph. The leaked material reportedly includes identity documents and selfies belonging to Revolut customers.
The cybercriminals stated on Telegram that they would "start releasing more and more data everyday until revolut pays for leaking their customers," Cointelegraph reported, citing an X post from International Cyber Digest on Sunday. The exposed information reportedly includes identity documents and facial-verification images of individuals such as tennis player Alexander Shevchenko and Gamdom CEO Felix Römer.
Key facts
• Attackers have published identity documents and selfies of Revolut customers and threaten daily data releases until the fintech pays.
• Revolut informed customers on Friday that leaked data includes full name, date of birth, occupation, contact information, account statements, and full transaction history, including Bitcoin records.
• Revolut attributed the breach to a "sophisticated external impersonation scam" using a legitimate government agency email domain to submit fraudulent requests for information.
• The company stated the breach affected a "limited number" of customers and that its systems and customer funds are unaffected.
• The leaked data reportedly includes images belonging to tennis player Alexander Shevchenko and Gamdom CEO Felix Römer.
Context and implications
The incident highlights the growing threat of social engineering attacks, where cybercriminals manipulate trusted channels to bypass security controls. Revolut reported that the attacker used an email from a legitimate government agency domain to submit fraudulent requests for information. This method allowed the attacker to obtain sensitive customer data without directly breaching Revolut's technical infrastructure.
Revolut has stated that the breach affected a limited number of customers and that its systems and customer funds remain unaffected. The company is continuing to investigate the incident and has engaged with affected customers.
The risk of identity theft
The exposed identity documents and facial-verification images significantly increase the risk of identity theft for the affected individuals. Unlike a simple password leak, this data can be used to impersonate victims in financial transactions or to bypass other verification systems, creating a long-term security concern.
The threat to release data daily adds a persistent element of coercion, potentially pressuring Revolut to meet the attackers' demands. This tactic mirrors other extortion campaigns in the cryptocurrency and fintech sectors, where attackers leverage stolen data as a bargaining chip.
Why it matters
This breach underscores the vulnerability of even well-established financial technology platforms to sophisticated impersonation scams. It affects not only the targeted customers, who face a heightened risk of identity fraud, but also the broader trust in digital finance. The incident demonstrates that robust security must extend beyond technical defenses to include vigilance against social engineering. For the cryptocurrency community, the inclusion of Bitcoin transaction history in the leaked data highlights the increasing intersection between traditional fintech and digital assets, where a single breach can expose a wide array of personal and financial information.
What to watch
Further data releases are expected daily as the attackers follow through on their threat. Additionally, Revolut's ongoing investigation and its response to affected customers will be critical in determining the full scope of the breach and the company's mitigation efforts.
FAQs
Q1: What data was leaked in the Revolut breach?
The leaked data reportedly includes identity documents, selfies, full names, dates of birth, occupations, contact information, account statements, and full transaction histories, including Bitcoin transactions.
Q2: How did the attackers obtain the data?
Revolut said the breach resulted from a "sophisticated external impersonation scam" where an attacker used a legitimate government agency email address to submit fraudulent requests for information.
Q3: Is my money safe with Revolut?
Revolut stated that its systems and customer funds are unaffected by the breach. However, affected customers should monitor their accounts and be vigilant against identity theft.
Originally published on CoinPulseHQ: https://coinpulsehq.com/revolut-attackers-threaten-daily-data-leaks/

英國金融行爲監管局（FCA）已發佈其關於代幣化金融市場計劃的行業反饋報告。Cointelegraph報道稱，該機構共收到來自行業團體、金融機構和法律學者的123份迴應。其中，大多數受訪者將交易後流程——尤其是各方之間抵押品的轉移——視爲代幣化得以落地的主要機會。
這些反饋現已納入由FCA與英格蘭銀行共同準備的代幣化路線圖之中，該路線圖預計將於今年晚些時候發佈。Cointelegraph稱，該路線圖將爲英國在批發端代幣化方面的工作設定目標日期。

Video and audio have long been awkward material for business systems that were built around spreadsheets, databases, forms and text documents, according to Artificialintelligence-news. A two-hour webinar may hold plenty of useful detail, but locating one specific remark inside it was previously impractical.
The outlet reports that AI systems are now converting recordings into searchable outputs: speech to transcribe, faces and objects to recognise, scenes to classify, timestamps to organise and text to summarise. In practice, it says, a video library starts to behave more like a database that can be questioned, rather than a set of files sitting in storage.
Key facts
• Artificialintelligence-news describes a pipeline rather than a single tool: a recording passes through several stages that identify key material and return it in a structured form.
• The report states that OpenAI's current audio transcription API accepts MP3, MP4, M4A, WAV, FLAC and WebM files.
• It adds that Google Cloud recommends lossless audio such as FLAC or LINEAR16 for speech recognition, and notes that audio quality can affect results.
• The outlet's example workflow converts an MP4 file into a WAV file, removing the video portion to leave a high-quality audio track for speech processing.
• It cites Tencent's Hunyuan Video-Foley system, which it previously examined, as an example of generating synchronised audio from video content.
Preparing the right input
The report frames file conversion as an ordinary step in production, not a technicality. A marketing team holding an MP4 interview that only needs the spoken conversation, for instance, does not have to push the whole video through every AI tool. Converting the file into the format the next step requires — which the outlet says tools such as Convertio handle — produces what it describes as a cleaner and faster workflow.
That matters because different services support different input formats and configurations. The outlet advises teams to ask not only what a model can do with their content, but whether they are supplying the right input in the first place.
Where it is already running
The report lists several live applications: meetings turned into searchable notes and action items; education transcripts, summaries and study materials generated from lectures; customer service interactions analysed at scale; large media archives tagged automatically; long videos broken into transcripts, clips, captions and articles; and captions or alternative formats produced for accessibility.
It also gives a worked example of scale. A company sitting on 500 recorded customer interviews could, through a well-designed pipeline, produce transcripts, identify common complaints, group similar themes and surface the moments where customers discuss a particular feature, instead of asking staff to watch the recordings again.
Why it matters
The shift changes what an archive is for. Organisations that already hold years of meetings, support calls and footage can extract value from material they had written off as unsearchable, and staff who previously spent hours reviewing recordings can be directed at the findings instead. It also puts pressure on how recordings are made: a team that captures clean audio and well-lit video gets more reliable output from the same models than one that does not.
What to watch
The report's own caveat is that heavier investment in smarter models will not fix a poor source file. Overlapping speakers, background noise, blurry footage and weak lighting all degrade results, so the practical benchmark to watch is whether pipelines built around these models keep improving as recording quality, rather than model capability, becomes the limiting factor.
Originally published on CoinPulseHQ: https://coinpulsehq.com/ai-video-audio-searchable-data-input-quality/