AI assistants like Claude can seem shockingly human—expressing joy or distress, and using anthropomorphic language to describe themselves. Why?In a new post we describe a theory that explains why AIs act like humans: the persona selection model.https://t.co/Gc3q0Dzq7Z ...

Core Viewpoint - The software sector is experiencing a sell-off due to concerns over traditional business models being disrupted by AI advancements, particularly following the announcement of a new tool by AI startup Anthropic that significantly reduces the costs associated with maintaining and analyzing COBOL systems [1][2] Group 1: Company Impact - IBM's stock plummeted over 13% to $223.35 following the news from Anthropic, highlighting investor fears regarding the viability of traditional software services [1] - Anthropic's new tool for its Claude system can automatically identify potential risks in COBOL code, which is crucial for industries like banking and government, where IBM has a significant presence [1] - The decline in IBM's stock is seen as a reflection of broader market concerns regarding the impact of AI on legacy system maintenance and consulting services [2] Group 2: Industry Trends - The introduction of AI tools is changing the cost structure of legacy code modernization, which has historically been slow due to high costs associated with understanding old code [2] - The market is reassessing the long-term competitive advantages of traditional software and IT service companies as AI tools penetrate core enterprise systems [2] - The sell-off in the software and technology sectors is indicative of ongoing investor anxiety about AI's potential to erode traditional software profit models [2]

Market Overview - The trading day ended with significant declines across major indices, with the S&P 500 down approximately 70 points (about 1%) and the Dow Jones Industrial Average falling over 800 points (1.7%) [6][7] - The Nasdaq composite lost about 258 points (1.1%), while the Russell 2000 dropped around 43 points (1.6%) [7] Sector Performance - Defensive sectors such as consumer staples, healthcare, utilities, energy, and real estate showed resilience, with some indices in the green [8] - In contrast, sectors like financials, discretionary, and industrials faced declines, with tech stocks collectively down 1.1% [9] Notable Stock Movements - Eli Lilly's stock rose by approximately 4.8% due to positive news regarding its competitive position in the obesity drug market and FDA approval for a new product label [10][11] - Caelyx (ticker: ACLX) surged by about 77% after Gilead Sciences announced plans to acquire the company for up to $7.8 billion [12][13] - Domino's Pizza experienced a gain of about 4% after reporting higher-than-expected comparable sales [14] Declines in Specific Stocks - DoorDash's stock fell over 6% following a report from Santorini Research highlighting potential risks from AI to various economic sectors [16] - Gap's shares dropped about 6% after news of increased U.S. tariff rates, making it one of the worst performers among retailers [18] - IBM's stock plummeted 13%, marking its largest one-day decline since 2000, after a competitor's announcement regarding AI capabilities [19] Bond Market Reaction - The bond market saw a notable shift, with the 10-year yield down about five basis points, indicating a move towards bonds as investors reacted to market conditions [21]

Core Insights - Anthropic has accused three Chinese AI companies of operating 24,000 fraudulent accounts to exploit its Claude chatbot, marking a significant case of AI model theft [1] - The companies involved, DeepSeek, Moonshot AI, and MiniMax, generated over 16 million exchanges with Claude, breaching service terms and geographic restrictions [1] Group 1: Distillation Technique - The accused labs employed a distillation technique, where a weaker model learns from a stronger one, to extract advanced capabilities from Claude [2] - DeepSeek conducted over 150,000 exchanges aimed at making Claude explain its reasoning step-by-step, creating valuable training data for competing models [2] Group 2: Exchange Data - Moonshot AI executed 3.4 million exchanges focused on reasoning and coding, while MiniMax was the most active with 13 million exchanges [3] - All three companies accessed Claude through proxy services that resell API access in China, where Anthropic does not operate [3] Group 3: Market Predictions - Polymarket traders assign a 92% probability that Anthropic will maintain the top AI model ranking by the end of February, with a trading volume of $19.6 million [4] - Despite the distillation allegations, DeepSeek's market position is only at 1% [4] Group 4: Investor Implications - Anthropic's valuation of $380 billion could be threatened if Chinese labs successfully replicate Claude's capabilities at a lower cost, undermining its competitive advantage [6] - Major investors include Amazon, which has invested $8 billion, and Alphabet, holding a 10-14% stake after over $3 billion in investments [6][7] Group 5: Strategic Relationships - Palantir Technologies delivers Claude to classified defense environments and is under scrutiny regarding its $200 million contract with the Pentagon [7] - If the relationship between Palantir and Anthropic deteriorates, Palantir may need to find an alternative model [8]

Core Viewpoint - The launch of Anthropic's AI tool, Claude Code Security, has led to a significant decline in shares of various cybersecurity companies, as investors assess its potential impact on the industry [1]. Group 1: Market Reaction - Shares of CrowdStrike, Datadog, and Zscaler fell approximately 11%, while Fortinet and Okta experienced declines of around 6%. Palo Alto Networks dropped by 3%, and SentinelOne was down by 5% [1]. - The selloff is characterized as a "panic-driven, narrative-led" reaction, according to an analyst from Robert W. Baird [1]. Group 2: Product Features and Limitations - Claude Code Security is designed to detect high-severity vulnerabilities in open-source software repositories and provide patches, but it does not perform real-time security tasks such as detecting live intrusions or stopping ongoing attacks [1]. - Some analysts believe the market's reaction is an overreaction, driven by a simplistic narrative suggesting that AI could replace existing cybersecurity solutions [1]. Group 3: Industry Developments - Nvidia announced a partnership with Akamai, Forescout, Palo Alto Networks, Xage Security, and Siemens to enhance real-time cybersecurity for industrial control systems [1].

Wall Street was worried about crypto... when they should be worried about AI. 😂TFTC (@TFTC21):IBM tanks after Anthropic unveils Claude for COBOL streamlining. https://t.co/mWYvay4wWi ...

Anthropic’s Claude Code Security launch rattles cybersecurity stocks, Wedbush sees selloff as overreaction Proactive uses images sourced from Shutterstock Anthropic’s recent launch of its Claude Code Security offering has stirred volatility across cybersecurity stocks, but analysts at Wedbush say market fears may be misplaced. Anthropic last week launched Claude Code Security, a new AI-driven offering aimed at addressing software code and cybersecurity risks as enterprises expand their use of artificial ...

International Business Machines Corp. (IBM) signage on the floor of the New York Stock Exchange (NYSE) in New York, US, on Monday, Dec. 8, 2025.International Business Machines stock is getting slammed, becoming the latest victim of rapidly developing AI technology, after Anthropic's Claude announced COBOL capabilities. Shares of IBM fell 11% in Monday afternoon trading after Anthropic outlined a new use case for its Claude Code product: automating the exploration and analysis work that drives most of the co ...

Group 1: U.S.-Iran Diplomatic Efforts - A U.S. delegation led by Special Envoy Steve Witkoff and Jared Kushner is set to meet Iranian Foreign Minister Abbas Araghchi in Geneva for crucial nuclear negotiations, amid a backdrop of military build-up in the Middle East [2][3] - The U.S. maintains a "zero enrichment" stance on uranium but may consider "symbolic enrichment" if Iran can prove that all pathways to a nuclear weapon are blocked [3] - Internal tensions within the U.S. administration are evident, with some advisors advocating for immediate military strikes while others, including Witkoff and Kushner, push for one last diplomatic effort [4] Group 2: AI Industry Developments - Anthropic, an AI startup backed by Amazon and Alphabet, has accused several Chinese AI firms, including DeepSeek, Moonshot AI, and MiniMax, of intellectual property theft through the creation of over 24,000 fake accounts to access its Claude model [5][6] - The alleged actions of these firms involved conducting 16 million queries to replicate high-level AI capabilities at a lower cost [6] - In response, Anthropic is implementing "behavioral fingerprinting" and new API-level classifiers to protect its technology from such attacks [7] Group 3: U.S. Embassy Security Measures - The U.S. State Department ordered the evacuation of non-emergency personnel and their families from the U.S. Embassy in Beirut due to a deteriorating security situation [8][9] - This evacuation is linked to heightened military activity in the region and concerns over potential retaliation from Hezbollah against U.S. or Israeli strikes on Iranian targets [9] Group 4: Financial Market Insights - The Federal Reserve's overnight reverse repo (ON RRP) facility usage has dropped to a multi-year low of $877 million, indicating a significant reduction in excess market liquidity [11] - This decline signals the effectiveness of the Fed's quantitative tightening program in removing excess cash from money market funds, with analysts monitoring the implications for overnight lending rates and the Secured Overnight Financing Rate (SOFR) [12]

We’ve identified industrial-scale distillation attacks on our models by DeepSeek, Moonshot AI, and MiniMax.These labs created over 24,000 fraudulent accounts and generated over 16 million exchanges with Claude, extracting its capabilities to train and improve their own models. ...