日期	订阅者增长	提及	频道
30 十月	0
29 十月	0
28 十月	0
27 十月	0
26 十月	0
25 十月	0
24 十月	0
23 十月	0
22 十月	0
21 十月	0
20 十月	0
19 十月	0
18 十月	0
17 十月	0
16 十月	0
15 十月	0
14 十月	0
13 十月	0
12 十月	0
11 十月	0
10 十月	0
09 十月	0
08 十月	0
07 十月	0
06 十月	0
05 十月	0
04 十月	0
03 十月	0
02 十月	0
01 十月	0

频道帖子

WhatsApp Accidentally Exposed Its AI Chatbot's Hidden Instructions Around 80 pages on WhatsApp's website leaked internal instructions for Meta's chatbot, including a section titled "WHAT NOT TO SAY." Meta confirmed a bug caused the exposure. The leak offered a rare look at how Meta shapes its AI's behavior behind the scenes. @Vitrine

2	Meta Quietly Removes AI Training From Its Web Crawling Terms Meta has updated its upcoming Commercial Terms, dropping the AI training clause from its crawling policy. However, the permission to crawl websites linked to business users remains in place. Site owners tied to Meta's business services should note: Meta can still crawl their content — just no longer explicitly for AI training purposes. @Vitrine	41
3	YouTube Shorts Gets New Creator Bonuses and Shopping Upgrades 📹 YouTube is testing bonus incentives for creators who invite others to Shorts. The platform is also expanding its shopping affiliates program and updating shopping analytics tools. This could make Shorts more attractive for creators looking to monetize their content. @Vitrine	32
4	Bluesky hit 46M users — here's what brands need to know 🐦 Bluesky grew nearly 60% in 2025 and now has over 46 million users. There are no ads, so reach depends entirely on useful posts and genuine engagement. Its Following feed is chronological, meaning content gets buried fast. Consistency and real conversation matter more on Bluesky than on algorithm-driven platforms. @Vitrine	31
5	⚠️ Meta Now Scrapes Your Website If You Use Business Accounts Starting Oct. 30, Meta's updated terms of service mean that using business features on their platforms grants consent for Meta's web scrapers to collect data from your website — including news content — to train AI models. There is no opt-out option available. If you run a business or news site linked to Meta, this change affects you directly. @Vitrine	18
6	Instagram Is Cracking Down on AI-Generated Slop 🤖 Instagram head Adam Mosseri says the platform is actively working to reduce AI-generated replica content. He raised the issue in a weekly Q&A and a separate Threads post. This signals Instagram may soon take stronger action against low-effort AI content flooding users' feeds. @Vitrine	15
7	X Rolls Out Unified Subscription Plans With Five Pricing Tiers X now offers bundled subscription packages across five price points, including a free tier. The plans consolidate X's services into integrated options for different types of users. This gives users more flexibility in choosing a plan that fits their needs and budget. @Vitrine	13
8	🍎 Apple Explains Why It's Blocking Ad-Tracking in Safari Apple has revealed why Safari started blocking certain identity companies last month. When users enter their email on a site, some firms convert it into a cross-site tracking ID — bypassing cookies. Apple wants this stopped, directly targeting LiveRamp, ID5, and The Trade Desk's UID 2.0. These companies were among the first cut off in Safari's September update. @Vitrine	12
9	Meta Joins Global Crackdown on West African Scam Networks 🌍 Meta, alongside Google, Microsoft, and others, partnered with law enforcement to disrupt criminal scam networks operating in West Africa. The coordinated effort targets organized groups responsible for large-scale fraud schemes affecting victims worldwide. This marks a rare cross-industry push to tackle scam operations at their source. @Vitrine	12
10	Billion Dollar Boy Group acquires LinkedIn creator agency to push into B2B Influencer agency Billion Dollar Boy has acquired Creator Authority, a LinkedIn-focused creator marketing firm. The move marks a strategic push into B2B influencer marketing, a growing but underserved space. The acquisition puts the company on track to exceed $100M in global revenue next year. 📈 B2B brands are increasingly turning to creators to reach professional audiences. @Vitrine	17
11	🤖 Robot toy "Eiliko" sends your conversations to servers in China The Eiliko, a cute AI companion robot marketed to kids, transmits voice data to Chinese servers via Wi-Fi. Security researchers warn this data could be vulnerable to hacking. Anyone using Eiliko risks having personal conversations stored outside their control. @Vitrine	15
12	⚠️ Hackers Breached AI Coding Tool OpenCode, Stealing User API Keys Attackers accessed part of OpenCode's platform and stole API keys from users, running up charges on their AI accounts. OpenCode is a popular open-source alternative to Claude Code and Cursor. Notably, AI tools flagged during the investigation refused to analyze the malicious code involved. If you use OpenCode, check your API usage and rotate your keys immediately. @Vitrine	12
13	Meta Quietly Updates Its Terms of Service 📄 Meta has revised its Terms of Service to close potential legal loopholes related to data scraping. The previous wording left room for misuse, so the company updated the language to tighten its data protections. Users and developers relying on Meta's platforms should review the updated agreement. @Vitrine	11
14	TikTok Payments in 2026: You Need Less Than You Think 💡 You don't need a massive following to earn on TikTok. The Creator Rewards Program requires 10,000 followers and 100,000 views in 30 days — but brand deals and affiliate marketing have no follower minimum. Brands prioritize audience fit over follower count, so smaller creators can still land paid opportunities. @Vitrine	9
15	📡 Spyware Now Spreading Through Targeted Ads — No Click Required Citizen Lab researchers have published the first technical evidence of mercenary spyware being delivered through online advertising. The attack requires no interaction from the target — simply receiving an ad can infect a device. This marks a significant escalation in surveillance technology, potentially affecting anyone exposed to targeted ads on their phone. @Vitrine	14
16	Apple Bans AI "Nudify" Apps Hiding Behind Fake Descriptions Apple has updated its App Store guidelines to block AI apps that disguise themselves as regular tools to sneak past review — specifically targeting so-called "nudify" apps. The new rules require AI apps to moderate offensive content or face removal. Developers who try to obscure an app's true purpose during review will now be in direct violation of Apple's policies. @Vitrine	9
17	🔍 Social Media Bans Under Political Scrutiny Before U.S. Midterms Two liberal accounts were disciplined on separate platforms just weeks before the U.S. midterm elections. The "Musk" documentary was permanently suspended from TikTok ads over political content — confirmed via emails, not automation. Every platform action now carries political weight, putting social media companies under intense public pressure. @Vitrine	7
18	💼 LinkedIn Is Getting Digital Business Cards LinkedIn is rolling out scannable digital business cards for its users. They can be saved to your phone's digital wallet and shared in person to make offline networking easier. This could replace the need for physical business cards at events and meetings. @Vitrine	7
19	AI Tool Blamed for Wrongly Suspending Anti-Trump Instagram Account Meta says an Anthropic Claude-linked tool caused the suspension of the No Kings Coalition's Instagram account on October 17. The No Kings Coalition is an organization opposed to Trump. Meta VP of Communications Andy Stone shared the explanation on Sunday. AI moderation errors can have real consequences for political groups trying to organize online. @Vitrine	7
20	AI Is Making Scams Harder to Spot ⚠️ Trading standards officers warn that AI is intensifying existing scams by powering fake websites and voice or face clones of real people. These tools make fraudulent content more convincing and harder for everyday users to detect. If something feels off — an unexpected offer, an urgent request, or a familiar face asking for money — verify through official channels before acting. @Vitrine	6

查看所有帖子