# Advertising Related Articles

HTX News Center provides the latest articles and in-depth analysis on "Advertising", covering market trends, project updates, tech developments, and regulatory policies in the crypto industry.

OpenAI Caught, ChatGPT Actually Knows What You Bought on Other Websites

A recent investigation by independent researcher Buchodi revealed that OpenAI's ChatGPT can track and link user activity from other websites to a specific ChatGPT account via a specific cookie named `__obi`. The process involves three steps: first, when a user visits chatgpt.com, a temporary token is generated and linked to their account for 60 seconds. This token is then exchanged for a long-lasting cookie (`__obi`) with a one-year validity, configured to be sent across websites. Finally, when the user visits a site that has embedded OpenAI's tracking code, this cookie, along with details like viewed products, orders, and even location data (country, city, zip code), is sent back to OpenAI's servers. The data transmission occurs automatically upon page load, without any user interaction. The mechanism currently works on Android devices using Chrome but is blocked on Apple's Safari and iOS Chrome due to stricter privacy controls. While the tracking method itself is not new—similar to practices used by companies like Meta for years—the concern is that users' private conversations with ChatGPT and their external browsing/purchase history are now being consolidated under one account. Notably, the `__obi` cookie is categorized under "data analytics" rather than "advertising," meaning it remains active even if users disable ad tracking in settings. OpenAI acknowledged the inquiry but did not provide specific answers regarding the classification of the cookie or whether users who opt out of marketing are still affected. The researcher emphasizes that while data is confirmed to be sent to OpenAI, it's unclear how it is internally processed or linked to accounts.

marsbit09/21 10:51

OpenAI Caught, ChatGPT Actually Knows What You Bought on Other Websites

marsbit09/21 10:51

From Traffic Logic to Task Logic: AI Agent Is Ending the Free Lunch of the Internet

From Traffic to Task: How AI Agents Are Ending the Internet’s Free Lunch Andrej Karpathy recently highlighted the rise of an "agent-native" economy, requiring a rewrite of software, documents, and processes for AI agents. This shift raises a critical question: who pays for this transformation, and how is value redistributed? The answer lies in a fundamental change in cost structure, challenging the core internet business model built on near-zero marginal costs. The internet economy thrived on the premise that serving additional users incurred minimal extra costs, enabling profitable models like advertising, subscriptions, and commissions. Advertising, in particular, leveraged long user journeys—searching, comparing, browsing—to insert commercial touchpoints. AI agents disrupt this logic. They execute tasks—reasoning, reading files, calling tools—each consuming real GPU compute, power, and depreciation costs. Marginal cost per task is significant and varies greatly (MC > 0). While the cost per token may fall, agents tackle more complex tasks, increasing total expenditure—an AI version of Jevons's Paradox. Furthermore, fixed costs become perpetual due to the need for costly, frequent model retraining every 6-12 months. This cost structure prevents the internet's "profit explosion past breakeven" dynamic. The business logic shifts from monetizing attention to delivering value. The basic unit changes from a "click" to a "task." An agent doesn't just list links for "best renovation company"; it can compare quotes, check reputations, and schedule appointments. Its value lies in task completion and decision agency. This compresses the user journey, reducing the number of touchpoints where traditional ads can intervene. Ads must become more explicit within AI responses to maintain user trust in the agent's objectivity. This transition reshapes industry dynamics: 1. **Entry Points & Distribution:** Control shifts from search engines/app stores to agents that decide *how* a task is completed—which tools, data, and merchants are used. 2. **Platform & Middleman Economics:** Pure listing/intermediation platforms are vulnerable. Competitive moats shift from traffic control to owning indispensable supply, fulfillment, and execution capabilities that agents cannot bypass. 3. **Software & Service Boundaries Blur:** SaaS transitions from a tool (System of Record) used by employees to a capability (System of Action) that completes tasks autonomously. This expands the software market from IT budgets into larger labor and service budgets (e.g., customer service, research). In essence, AI agent economics ends the internet's simple "more users = more profit" equation. Scarcity is no longer just about traffic/attention but about high-value tasks and the data, permissions, and execution capabilities needed to complete them. As the fundamental economic unit changes from "click" to "task," the business models, platform power, and profit distribution built over the last two decades are set for a major recalculation.

marsbit09/07 19:16

From Traffic Logic to Task Logic: AI Agent Is Ending the Free Lunch of the Internet

marsbit09/07 19:16

Pro-Cryptocurrency PAC Wins Primaries but Loses $2M Race in Florida

A pro-cryptocurrency political action committee (PAC), Fairshake, achieved mixed results in recent U.S. primary elections, spending millions to influence races ahead of the 2026 midterms. Fairshake's affiliated PACs, Protect Progress and Defend American Jobs, supported candidates in Alaska, Florida, and Wyoming, spending approximately $3.6 million. Four out of five candidates backed by their ads won their primaries or advanced. These included Democrat Lois Frankel in Florida, Republican Nick Begich in Alaska, Republican Sydney Gruters in Florida, and Representative Harriet Hageman in Wyoming's Senate race. However, a high-profile $2 million negative ad campaign by Protect Progress failed to defeat Democrat Oliver Gilbert in Florida's 24th District. Gilbert won despite the ads, which he attributed to "Trump's techie-billionaire friends" and "crypto-scammers trying to buy a Democratic primary." A PAC representative defended the ads' factual accuracy. Fairshake, with a reported $193 million war chest, has spent over $130 million in the 2024 cycle supporting pro-crypto candidates and opposing critics. A spokesperson stated the PAC is "just getting started" to build the largest pro-crypto Congress in history. The outcomes could impact future crypto legislation, such as the pending CLARITY Act, as the 2026 elections may shift congressional control and determine which parties advance or block industry-related bills.

cryptonews.ru08/20 06:12

Pro-Cryptocurrency PAC Wins Primaries but Loses $2M Race in Florida

cryptonews.ru08/20 06:12

MSX US Stock Daily Observation: Baidu 2026 Q2 Earnings: AI Revenue Accounts for Half of Core Business for Two Consecutive Quarters, 283% GPU Cloud Growth Boosts Business Transformation

**MSX Daily US Stock Observation: Baidu 2026 Q2 Earnings Report** Baidu's Q2 2026 earnings present a clear picture of a company in transition. Total revenue declined slightly to RMB 313.25 billion, missing market expectations. This was driven by a continued double-digit contraction (-19%) in its core online marketing (advertising) business. The key highlight is the structural shift within the company's operations. AI-related revenue reached RMB 125 billion, accounting for approximately half of Baidu's core business income for the second consecutive quarter. Within AI, growth is heavily skewed towards infrastructure: AI cloud infrastructure revenue grew 50% year-over-year to RMB 73 billion, with GPU cloud revenue surging 283%. In contrast, AI application and AI-native marketing services revenue grew only 3% and remained flat, respectively, indicating slower commercialization on the application side. Profitability metrics were mixed. While adjusted operating profit and EBITDA exceeded expectations, demonstrating cost control, adjusted earnings per ADS of RMB 7.22 fell short by about 26%. This gap is largely attributed to depreciation and amortization costs linked to significant AI infrastructure investments, which continue to pressure bottom-line profit realization. The company holds a substantial cash position of RMB 2831 billion, providing runway for continued investment. The critical challenge ahead is translating the explosive growth in AI infrastructure (like GPU cloud) into stronger monetization from AI applications to bridge the gap between operational profit and per-share earnings.

Odaily星球日报08/19 16:10

MSX US Stock Daily Observation: Baidu 2026 Q2 Earnings: AI Revenue Accounts for Half of Core Business for Two Consecutive Quarters, 283% GPU Cloud Growth Boosts Business Transformation

Odaily星球日报08/19 16:10

活动图片