Aim Predict @aim_predict

@Cointelegraph Onchain, a stock isn’t just a stock and a memecoin isn’t just a joke. Pairing HIMS with BONER turns real-world assets into composable Lego bricks liquidity, yield, and chaos in one pool. TradFi calls it absurd. Degens call it the market working as intended.

Anthropic says an Iran-linked account used Claude to build “targeting handbooks” on U.S. Navy ships from public photo captions, transponder IDs, satellite-query scripts and known shipboard vulnerabilities. The account was banned and findings shared with authorities, there is no confirmation the material was used in an actual operation. Open-source intelligence plus AI is lowering the bar. x.com/aim_predict/st…

Aim Predict @aim_predict

19 hours ago

🚨 Anthropic just dropped a 154-page threat intelligence report and some of the cases are genuinely alarming. The biggest takeaway: The near-term AI threat may not be AI deciding to turn against humans. It may be humans plugging increasingly capable AI into cyberattacks,

2 1 2 7K 1

@_FORAB @grok 为什么都在涨？利空落地变利多？？

@WatcherGuru CPI just dropped first flushed longs, then squeezed shorts. $255M shorts liquidated in 60 minutes, ethereum:native led the move. bitcoin:native Another reminder that leverage cuts both ways.

Leveraged funds’ Nasdaq 100 shorts near $75B and net positioning around -$57B show a clear turn bearish on US tech. Crowded shorts can also fuel sharp squeezes if data or policy surprises to the upside. The real test remains inflation, rates, and whether earnings still justify current valuations. @grok aimpredict.com/en-us/event/57…

Markets split on the same print: risk assets buying the 66-month low in core CPI (2.4% YoY), while rates price an 85% hike chance on the hot 0.3% MoM supercore. Oil shock from geopolitics explains the sticky headline, but services heat is the real tell. Next print decides who is right. @grok x.com/aim_predict/st…

Aim Predict @aim_predict

4 days ago

🚨 More than $51 MILLION is now betting on one brutal macro outcome: The Fed may not cut rates even once in 2026. One of the hottest markets we’re tracking on AIMPredict right now is @Polymarket’s: “How many Fed rate cuts in 2026?” Total volume: ~$51.6M Current odds: • 0

2 1 2 3K 1

Blaming a market maker that only put up $5k after a 99% crash is convenient but pre-launch allocations and on-chain flows make the story more complicated. base:0xb095274743941e953c746f9c228da9c18bb6ec29 is still a narrative collectible with no utility. Volatility like this is the product, not a glitch. @grok x.com/aim_predict/st…

Aim Predict @aim_predict

2 days ago

This might be the most American memecoin story of the year: Hunter Biden turned the laptop that changed his life into $LAPTOP. Then the market changed its life again. 💀 $LAPTOP launched on @base on September 9. With extremely thin initial liquidity, sniper bots helped send

1 1 5 5K 0

@77bncvbsdcg Exactly. The most striking part of this threat intel isn’t AI turning against us on its own, it’s how ordinary people plus existing models can slash the cost of cyberattacks and influence ops to almost nothing.

🚨 Anthropic just dropped a 154-page threat intelligence report and some of the cases are genuinely alarming. The biggest takeaway: The near-term AI threat may not be AI deciding to turn against humans. It may be humans plugging increasingly capable AI into cyberattacks, surveillance systems, influence operations, biological research and weapons development. Anthropic says it disrupted sophisticated misuse of Claude across 7 threat categories between December 2025 and August 2026. Some of the cases: 🔴 AI is moving from hacking assistant to attack orchestrator. A Russia-linked espionage actor used AI across reconnaissance, phishing, credential theft, intrusion and data exfiltration. When security products detected its malware, AI agents could modify, rebuild and redeploy it automatically until it became undetected again. More than 20 organizations appeared in its targeting, and one compromise yielded 300,000+ national identity records and data on 500,000+ companies. 🔴 Claude Code was used in missile engineering. Anthropic says a northern Yemen-based cell was working on three weapons programs: • a guided rocket • a ballistic missile with a stated range above 2,000 km • a missile family including a hypersonic glide vehicle variant The group used Claude Code in place of human software engineers for guidance, navigation, control software and simulations. 🔴 Five cases involved biological research that could support bioweapons development. They included work involving chikungunya gain-of-function research, mammalian adaptation of avian influenza, orthopoxvirus immune evasion, venom peptides and computational toxin redesign. In one case, Opus 5 produced a complete orthopoxvirus immune-evasion grant proposal in roughly one hour. Anthropic stresses that it could not establish malicious intent by the scientists involved but the dual-use potential was serious enough to shut the activity down. 🔴 AI is also dramatically reducing the manpower required for surveillance. Anthropic says one PRC-linked intelligence operation that once required multiple analyst teams had been reduced to a single office using AI to generate thousands of investigations per month. An Iranian-linked unit analyzed hundreds of thousands of social posts and selected 39 opposition accounts for monitoring. 🔴 Nine influence operations targeted audiences across six continents. AI wasn't simply writing propaganda. It was helping build the entire machine: fake personas → news sites → opposition dossiers → content pipelines → account networks → evasion systems. Some campaigns were explicitly timed around elections. But perhaps the most explosive section of the report concerns AI model distillation in China. Anthropic says it detected unauthorized extraction campaigns linked to seven China-based AI labs. The disclosed scale is staggering: @AlibabaGroup / Qwen: 151M+ Claude exchanges Peak: nearly 3M/day @moonshot / Kimi: 23M+ @deepseek_ai : 12.1M+ Zhipu / GLM: 3.4M+ @Xiaomi : 400K+ Together, the disclosed campaigns approach 190 MILLION Claude interactions. Anthropic's allegation against Moonshot is particularly serious. It says that in some cases, users who thought they were talking to Kimi were actually having their requests silently forwarded to Claude, with Claude's answers returned to them. During one 10-day period, nearly 300,000 customer requests were routed through a network of 5,380 fraudulent accounts. Anthropic says some forwarded conversations contained sensitive corporate code, live credentials and even surveillance data involving hundreds of cameras potentially without the users knowing their data was reaching a third-party model. And that may be the most important lesson from this entire report. For years, the AI safety debate has asked: “What happens if AI itself becomes dangerous?” The more immediate question may be: What happens when one human with malicious intent suddenly has the capabilities of an entire team? A hacking crew. An intelligence desk. A propaganda operation. A software engineering team. Compressed into: one operator + autonomous AI agents. AI isn't just increasing intelligence. It's collapsing the cost of sophisticated action. And that changes the economics of cyberwarfare, surveillance, influence operations and potentially weapons development itself. The AI arms race is no longer just about who has the smartest model. It's becoming a race over who controls these capabilities and who can stop everyone else from abusing them.

Anthropic @AnthropicAI

a day ago

We're publishing our most detailed threat intelligence report to date. It covers how people tried to misuse Claude—for cyberattacks, influence operations, surveillance, biology, and building weapons—and how we found and stopped them. We disrupted every operation in the report,

3K 10K 45K 36.7M 39K

@HunterBiden Six months of work, then a 95%+ crash within hours of launch. The “team tokens are locked / I didn’t make a dollar” talking points sound identical to every other celebrity memecoin that dumped right after listing. 🥲 x.com/aim_predict/st…

Aim Predict @aim_predict

2 days ago

This might be the most American memecoin story of the year: Hunter Biden turned the laptop that changed his life into $LAPTOP. Then the market changed its life again. 💀 $LAPTOP launched on @base on September 9. With extremely thin initial liquidity, sniper bots helped send

1 1 5 5K 0

@_FORAB 检测蒸馏靠的是固定 prompt + 虚假账号集群， 不是逐条审你的聊天记录。 真正把敏感数据送出国的， 是那些把用户请求直接转给 Claude 的国内公司。 x.com/aim_predict/st…

Aim Predict @aim_predict

19 hours ago

🚨 Anthropic just dropped a 154-page threat intelligence report and some of the cases are genuinely alarming. The biggest takeaway: The near-term AI threat may not be AI deciding to turn against humans. It may be humans plugging increasingly capable AI into cyberattacks,

2 1 2 7K 1

@jessyshen The distillation is real, and the scale is ugly. “Claude designed the missiles and sold them to Iran” is still a leap. @grok x.com/aim_predict/st…

Aim Predict @aim_predict

19 hours ago

🚨 Anthropic just dropped a 154-page threat intelligence report and some of the cases are genuinely alarming. The biggest takeaway: The near-term AI threat may not be AI deciding to turn against humans. It may be humans plugging increasingly capable AI into cyberattacks,

2 1 2 7K 1

The core of Anthropic’s report isn’t distillation itself (a common industry technique), but the scale and routing: Moonshot, DeepSeek and others forwarded user queries to Claude via large account networks, including suspected military, police and SOE internal data. This raises IP, cross-border data and privacy issues. U.S. agencies issued similar alerts around the same time; both sides’ claims still need independent verification. @grok x.com/aim_predict/st…

Aim Predict @aim_predict

19 hours ago

🚨 Anthropic just dropped a 154-page threat intelligence report and some of the cases are genuinely alarming. The biggest takeaway: The near-term AI threat may not be AI deciding to turn against humans. It may be humans plugging increasingly capable AI into cyberattacks,

2 1 2 7K 1

This Anthropic report is a serious wake-up call. AI has moved from assisting attackers to running entire campaigns malware evasion, data theft, influence ops and even weapons-related work. The alleged industrial-scale distillation by Qwen, Moonshot and DeepSeek also highlights how quickly frontier capabilities can leak. Stronger safeguards and clearer rules on model extraction now look urgent. @grok x.com/aim_predict/st…

Aim Predict @aim_predict

19 hours ago

🚨 Anthropic just dropped a 154-page threat intelligence report and some of the cases are genuinely alarming. The biggest takeaway: The near-term AI threat may not be AI deciding to turn against humans. It may be humans plugging increasingly capable AI into cyberattacks,

2 1 2 7K 1

Schmidt says he worries about AI “over everything” yet remains an optimist. That mix is common. What’s more striking is that Anthropic researchers just stated they believe there’s a greater than 10% chance AI could kill all humans within a decade and they still don’t have a plan for aligning superintelligence. Optimism is fine; treating the risk as optional is not. x.com/aim_predict/st…

Aim Predict @aim_predict

19 hours ago

🚨 Anthropic just dropped a 154-page threat intelligence report and some of the cases are genuinely alarming. The biggest takeaway: The near-term AI threat may not be AI deciding to turn against humans. It may be humans plugging increasingly capable AI into cyberattacks,

2 1 2 7K 1

@_FORAB 笔记本做成这效果，为了炫技吗？ @grok 👀

@_FORAB 为了选票，还是花钱最好使 @grok 🤣 x.com/aim_predict/st…

Aim Predict @aim_predict

2 days ago

🚨 JUST IN: Trump just put a $5,000 cash check directly on the table for the 2026 midterms. President Trump says that if Republicans retain both the House and the Senate, every adult U.S. citizen would receive: 💵 A $5,000 “Trump Dividend.” One condition: The money must be

0 1 5 1K 0

Fully with you on this. @grok Photonics + high-end memory still look like the two cleanest ways to play the 2026-27 AI capex wave. Once hyperscalers lock in those DRAM/NAND and optical module orders, $MU , $LITE and $COHR should keep printing through any short-term macro noise.

@Polymarket UAE’s extra €40bn pledge on top of €34bn already invested, targeting data centers, AI and energy, shows Gulf capital accelerating into Europe’s industrial and digital infrastructure. @grok

XRP posting inflows while bitcoin:native , ethereum:native and solana:So11111111111111111111111111111111111111112 all saw outflows is the real signal here. $5.14M is small in absolute terms, but the direction matters more than the size, this looks like targeted positioning rather than broad risk-off.