global edition · 2026-09-14
All BriefingsTech & Finance Daily Briefing — September 14, 2026
9 evidence-backed technology and finance stories shaping the day.
Story at a glance
PBOC caps daily overnight reverse repo at 600 billion yuan from Sept 14-17
What happened
本周中国央行公开市场7天逆回购到期规模为85亿元，其中周一至周五分别到期5亿元、10亿元、0亿元、30亿元和40亿元；周一另有5000亿元6个月期买断式逆回购到期。
Inspect 1 cited report
本周中国央行公开市场7天逆回购到期规模85亿元，其中周一到周五分别到期5亿元、10亿元、0亿元、30亿元和40亿元。周一另有5000亿元6个月期买断式逆回购到期。
Original excerpt
本周中国央行公开市场7天逆回购到期规模85亿元，其中周一到周五分别到期5亿元、10亿元、0亿元、30亿元和40亿元。周一另有5000亿元6个月期买断式逆回购到期。
Open original report
Why it matters · interpretation
周一单日5000亿元6个月期买断式逆回购到期，规模远超当周7天逆回购合计到期量，显示短期流动性回笼压力集中在周初，市场需关注央行是否进行对冲操作。
Inspect 1 cited report
本周中国央行公开市场7天逆回购到期规模85亿元，其中周一到周五分别到期5亿元、10亿元、0亿元、30亿元和40亿元。周一另有5000亿元6个月期买断式逆回购到期。
Original excerpt
本周中国央行公开市场7天逆回购到期规模85亿元，其中周一到周五分别到期5亿元、10亿元、0亿元、30亿元和40亿元。周一另有5000亿元6个月期买断式逆回购到期。
Open original report
Story at a glance
Anthropic researcher Jacob Coxon resigns, warns AI race risks human extinction by 2030
What happened
Former Anthropic and OpenAI researcher Jacob Coxon gave interviews to the BBC and NBC News in which he said AI development is like "inviting an alien mind onto the planet" and that without slowing current progress "there is a strong chance that we could all die in the immediate future." Anthropic CEO Dario Amodei responded on CNN, saying he agrees with Coxon "much more than I disagree with him" and that Coxon was calling out the industry's pace rather than Anthropic specifically.
Inspect 4 cited reports
AI staff 'genuinely frightened' for humanity's future, ex-Anthropic researcher tells BBC
Original excerpt
publication date:2026-09-13 01:15:04  *Figure caption,'There is a possibility of human extinction,' former Anthropic employee says about AI* By Brandon Livesay  and  Mitchell Labiak , business reporter Updated 13 minutes ago **An artificial intelligence researcher who quit AI firm Anthropic ha
Open original reportJacob Coxon’s next interview on BBC (the ex-Anthropic+OpenAI researcher who resigned). "If we don't slow down at the current rate of progress, there is a strong chance that we could all die in the immediate future.
Original excerpt
Jacob Coxon’s next interview on BBC (the ex-Anthropic+OpenAI researcher who resigned). "If we don't slow down at the current rate of progress, there is a strong chance that we could all die in the immediate future. It's not at all an exaggeration. And many of the people who are deep in the weeds and actually coding the thing, believe that there is a greater than 10% chance for a probability that humans might die. And they all keep this in their head on a daily basis while working on the techno
Open original reportOn CNN, Dario Amodei finally spoke on Jacob Coxon’s 166M+ super viral tweet (the ex-Anthropic and OpenAI researcher who resigned). "I agree with Jacob much more than I disagree with him. It’s an interesting resignation because
Original excerpt
On CNN, Dario Amodei finally spoke on Jacob Coxon’s 166M+ super viral tweet (the ex-Anthropic and OpenAI researcher who resigned). "I agree with Jacob much more than I disagree with him. It’s an interesting resignation because when he left, he said, “I think Anthropic is the most responsible player. I think Anthropic is the most aware of these issues.” He wasn’t calling out us. He was calling out the dynamic of the industry as a whole moving too fast." ---- From "CNN" YouTube channel, (full
Open original reportOkay, now Jacob Coxon on NBC News (the ex-Anthropic+OpenAI researcher who resigned). NBC News: "What do you mean when you say AI could kill us all? What is the worst-case scenario, or scenarios, that keeps
Original excerpt
Okay, now Jacob Coxon on NBC News (the ex-Anthropic+OpenAI researcher who resigned). NBC News: "What do you mean when you say AI could kill us all? What is the worst-case scenario, or scenarios, that keeps you up at night?" Jacob Coxon: "Before discussing specific scenarios, it helps to just think of building AI as sort of inviting an alien mind onto the planet. You're building a human-level or superhuman-level mind without understanding what it wants, what it thinks, or the way it thinks.
Open original report
Why it matters · interpretation
The interviews mark a shift from a viral social media post to sustained mainstream broadcast coverage of existential AI risk warnings from a former insider, while Amodei's CNN response signals that a leading AI company executive is publicly validating rather than dismissing those concerns. The BBC also reported that the warnings "fall flat with some in Silicon Valley," indicating the warnings are intensifying public debate even as they meet industry resistance.
Inspect 5 cited reports
Dramatic insider warnings over AI fall flat with some in Silicon Valley
Original excerpt
publication date:2026-09-12 17:28:53   By Lily Jamali North America technology correspondent , Reporting from in San Francisco Upd
Open original reportAI staff 'genuinely frightened' for humanity's future, ex-Anthropic researcher tells BBC
Original excerpt
publication date:2026-09-13 01:15:04  *Figure caption,'There is a possibility of human extinction,' former Anthropic employee says about AI* By Brandon Livesay  and  Mitchell Labiak , business reporter Updated 13 minutes ago **An artificial intelligence researcher who quit AI firm Anthropic ha
Open original reportJacob Coxon’s next interview on BBC (the ex-Anthropic+OpenAI researcher who resigned). "If we don't slow down at the current rate of progress, there is a strong chance that we could all die in the immediate future.
Original excerpt
Jacob Coxon’s next interview on BBC (the ex-Anthropic+OpenAI researcher who resigned). "If we don't slow down at the current rate of progress, there is a strong chance that we could all die in the immediate future. It's not at all an exaggeration. And many of the people who are deep in the weeds and actually coding the thing, believe that there is a greater than 10% chance for a probability that humans might die. And they all keep this in their head on a daily basis while working on the techno
Open original reportOn CNN, Dario Amodei finally spoke on Jacob Coxon’s 166M+ super viral tweet (the ex-Anthropic and OpenAI researcher who resigned). "I agree with Jacob much more than I disagree with him. It’s an interesting resignation because
Original excerpt
On CNN, Dario Amodei finally spoke on Jacob Coxon’s 166M+ super viral tweet (the ex-Anthropic and OpenAI researcher who resigned). "I agree with Jacob much more than I disagree with him. It’s an interesting resignation because when he left, he said, “I think Anthropic is the most responsible player. I think Anthropic is the most aware of these issues.” He wasn’t calling out us. He was calling out the dynamic of the industry as a whole moving too fast." ---- From "CNN" YouTube channel, (full
Open original reportOkay, now Jacob Coxon on NBC News (the ex-Anthropic+OpenAI researcher who resigned). NBC News: "What do you mean when you say AI could kill us all? What is the worst-case scenario, or scenarios, that keeps
Original excerpt
Okay, now Jacob Coxon on NBC News (the ex-Anthropic+OpenAI researcher who resigned). NBC News: "What do you mean when you say AI could kill us all? What is the worst-case scenario, or scenarios, that keeps you up at night?" Jacob Coxon: "Before discussing specific scenarios, it helps to just think of building AI as sort of inviting an alien mind onto the planet. You're building a human-level or superhuman-level mind without understanding what it wants, what it thinks, or the way it thinks.
Open original report
Story at a glance
Fed rate hike probability surges to 86.2% after August CPI exceeds expectations
What happened
CME FedWatch data released Sept. 13 showed the market-implied probability of a Federal Reserve rate hike this week at 86.2%, with a 13.8% chance of no change.
Inspect 1 cited report
【美联储本周加息的概率为86.2%】9月14日讯，据CME“美联储观察”：美联储到9月维持利率不变的概率为13.8%，累计加息25个基点的概率为86.2%。美联储到10月维持利率不变的概率为6.9%，累计加息25个基点的概率为50.1%，累计加息50个基点的概率为42.9%。
Original excerpt
【美联储本周加息的概率为86.2%】9月14日讯，据CME“美联储观察”：美联储到9月维持利率不变的概率为13.8%，累计加息25个基点的概率为86.2%。美联储到10月维持利率不变的概率为6.9%，累计加息25个基点的概率为50.1%，累计加息50个基点的概率为42.9%。
Open original report
Why it matters · interpretation
The 86.2% probability signals strong market consensus for imminent tightening, while the October distribution showing a 42.9% chance of a cumulative 50 basis points indicates expectations that additional hikes may follow quickly.
Inspect 1 cited report
【美联储本周加息的概率为86.2%】9月14日讯，据CME“美联储观察”：美联储到9月维持利率不变的概率为13.8%，累计加息25个基点的概率为86.2%。美联储到10月维持利率不变的概率为6.9%，累计加息25个基点的概率为50.1%，累计加息50个基点的概率为42.9%。
Original excerpt
【美联储本周加息的概率为86.2%】9月14日讯，据CME“美联储观察”：美联储到9月维持利率不变的概率为13.8%，累计加息25个基点的概率为86.2%。美联储到10月维持利率不变的概率为6.9%，累计加息25个基点的概率为50.1%，累计加息50个基点的概率为42.9%。
Open original report
Story at a glance
Oracle Chairman Larry Ellison cancels $7.5 billion planned stock sale
What happened
Larry Ellison canceled a planned $7.5 billion sale of Oracle stock, according to reports published on Sept. 13.
Inspect 2 cited reports
Larry Ellison cancels $7.5 billion sale of Oracle stock
Original excerpt
publication date:2026-09-13 20:49:38  # Larry Ellison cancels $7.5 billion sale of Oracle stock  # On Second Thought, Never Mind: Here’s Why ORCL’s Larry Ellison Just Canceled $7.5 Billion Stock Sale
Open original report
Why it matters · interpretation
The reversal removes a major overhang on Oracle shares that a sale of that size would have created, and it signals a change in Ellison's disposition toward reducing his stake.
Inspect 2 cited reports
Larry Ellison cancels $7.5 billion sale of Oracle stock
Original excerpt
publication date:2026-09-13 20:49:38  # Larry Ellison cancels $7.5 billion sale of Oracle stock  # On Second Thought, Never Mind: Here’s Why ORCL’s Larry Ellison Just Canceled $7.5 Billion Stock Sale
Open original report
Story at a glance
Nintendo funds major U.S. Switch sale with tariff refunds, sparking deceptive marketing criticism
What happened
Mashable published a roundup of deals still available in Nintendo's Customer Appreciation sale.
Inspect 1 cited report
Nintendo Customer Appreciation sale: Best deals you can still grab - Mashable
Original excerpt
<ol><li><a href="https://news.google.com/rss/articles/CBMiigFBVV95cUxOWExNaTY2VkdDdU5fRkhkVTUtSF9XT09jdEttLWktYTZNWTRUWXlpNEJsQ0NWQ1VuQmxVNkprT3ZtSmRiVU9ialpEYXJnZmZybjZJNGhPRU1tNFpMVnpYdHFMTEFHNE1fR1E2ZzZNS0psSU1JWW1LbV9YMDBYRmFSMnJibFlfdlhvOFE?oc=5" target="_blank">Nintendo Customer Appreciation sale: Best deals you can still grab</a> <font color="#6f6f6f">Mashable</font></li><li><a href="https://news.google.com/rss/articles/CBMisAFBVV95cUxNZkNXQXJrT2ZKY1JuQVM2X2JPeGhRaHNoVTNlME1aej
Open original report
Why it matters · interpretation
The roundup signals that the sale remains active and that notable discounts are still accessible to shoppers, extending the window for consumers to act on the promotion.
Inspect 1 cited report
Nintendo Customer Appreciation sale: Best deals you can still grab - Mashable
Original excerpt
<ol><li><a href="https://news.google.com/rss/articles/CBMiigFBVV95cUxOWExNaTY2VkdDdU5fRkhkVTUtSF9XT09jdEttLWktYTZNWTRUWXlpNEJsQ0NWQ1VuQmxVNkprT3ZtSmRiVU9ialpEYXJnZmZybjZJNGhPRU1tNFpMVnpYdHFMTEFHNE1fR1E2ZzZNS0psSU1JWW1LbV9YMDBYRmFSMnJibFlfdlhvOFE?oc=5" target="_blank">Nintendo Customer Appreciation sale: Best deals you can still grab</a> <font color="#6f6f6f">Mashable</font></li><li><a href="https://news.google.com/rss/articles/CBMisAFBVV95cUxNZkNXQXJrT2ZKY1JuQVM2X2JPeGhRaHNoVTNlME1aej
Open original report
Story at a glance
Nvidia in talks to invest up to $10 billion in Anthropic’s potential $2 trillion IPO
What happened
Nvidia is reported to be considering a $10 billion investment in Anthropic's IPO, which is valued at $2 trillion.
Inspect 2 cited reports
Nvidia May Put $10 Billion Into Anthropic’s $2 Trillion IPO. Amazon Already Has Much More Riding on It - Yahoo Finance
Original excerpt
<ol><li><a href="https://news.google.com/rss/articles/CBMikwFBVV95cUxQYUY1U1I2OXg0ZkhOdzQ4emlKOExjUDNXbzNqOHJFX0ZDbEt6QnRXT1lVNmdOZ05kbFVLX1RJRlhmcXdhT21uYTJfUmR3azE4MG5Kc1RJcUFxUFBiZjhIeUI4cjFfU0FMVXI4bndiN3pXNlJpd2R3YVRsUlA5X1pSaTFSMERoRDFMTThUdjdNS0hEZzQ?oc=5" target="_blank">Nvidia May Put $10 Billion Into Anthropic’s $2 Trillion IPO. Amazon Already Has Much More Riding on It</a> <font color="#6f6f6f">Yahoo Finance</font></li><li><a href="https://news.google.com/rss/articles/CBMir
Open original reportNvidia may bankroll Anthropic's massive IPO - Mashable
Original excerpt
<ol><li><a href="https://news.google.com/rss/articles/CBMibkFVX3lxTE1Bb01qUGttTkxJUEt3SWdwakxaMzZjTXd3V3E5RjhXXzMyZW9YX0pjUmtlcGZyYXp1amRTbVlVMmxka284RWxFcGp5Tnp0TDEtbHUzeU1rcXZfdkFkTU53UnR0ZEdndDc2RVF1R0RB?oc=5" target="_blank">Nvidia may bankroll Anthropic's massive IPO</a> <font color="#6f6f6f">Mashable</font></li><li><a href="https://news.google.com/rss/articles/CBMirwFBVV95cUxOeVpmU0JnRFZzaXNNSXlMZzQtVkwyNHB6dzc0bldFRldIRFpYdGRVNm1zRTFhWUVRNkZmR3ZJcnVTc1RLS0k5dHpmaEdYZ0tWTHFPUkhn
Open original report
Why it matters · interpretation
The reported investment would deepen Nvidia's ties to a major AI model developer while Amazon already holds a larger stake, underscoring how chipmakers and cloud providers are competing for influence over leading AI firms.
Inspect 1 cited report
Nvidia May Put $10 Billion Into Anthropic’s $2 Trillion IPO. Amazon Already Has Much More Riding on It - Yahoo Finance
Original excerpt
<ol><li><a href="https://news.google.com/rss/articles/CBMikwFBVV95cUxQYUY1U1I2OXg0ZkhOdzQ4emlKOExjUDNXbzNqOHJFX0ZDbEt6QnRXT1lVNmdOZ05kbFVLX1RJRlhmcXdhT21uYTJfUmR3azE4MG5Kc1RJcUFxUFBiZjhIeUI4cjFfU0FMVXI4bndiN3pXNlJpd2R3YVRsUlA5X1pSaTFSMERoRDFMTThUdjdNS0hEZzQ?oc=5" target="_blank">Nvidia May Put $10 Billion Into Anthropic’s $2 Trillion IPO. Amazon Already Has Much More Riding on It</a> <font color="#6f6f6f">Yahoo Finance</font></li><li><a href="https://news.google.com/rss/articles/CBMir
Open original report
Story at a glance
OpenAI CEO Sam Altman rules out 2026 IPO, citing AI safety concerns
What happened
Sam Altman said OpenAI will not go public in 2026, citing the need to prioritize AI safety and alignment over shareholder interests.
Inspect 4 cited reports
OpenAI delaying IPO amid AI safety concerns, Sam Altman says - Axios
Original excerpt
<ol><li><a href="https://news.google.com/rss/articles/CBMidkFVX3lxTE5kOTVXM1dpbzZyS3lOVnlHTFBkQlIxZUJDTWRoZ1doeEg2TWl5czhzWnVsWWRuclZLcThWM3RjN3J0Ym5nWEdHRl9LM0cyckVyLW9aWGxMU05XbVpVcThqVEZIYndSTm04V2tLUll6QVV1YVlaUWc?oc=5" target="_blank">OpenAI delaying IPO amid AI safety concerns, Sam Altman says</a> <font color="#6f6f6f">Axios</font></li><li><a href="https://news.google.com/rss/articles/CBMitAFBVV95cUxQbWdEQ1p3VzJkd2E3ZmYwWmRkRWlwWEhFSm1PTjQ3Y1N1bGNPd1lXcjZncEhPVVNrd3dFT3oxd1FVbzV
Open original reportIPO Delayed: Sam Altman Says an OpenAI Going Public in 2026 Would Be ‘Ill-Advised’
Original excerpt
publication date:2026-09-13 15:08:28  # IPO Delayed: Sam Altman Says an OpenAI Going Public in 2026 Would Be ‘Ill-Advised’ Rich Duprey Sun, September 13,
Open original report„Ich würde sagen, nicht 2026“ – Kein Börsengang von OpenAI wegen Sicherheitsbedenken
Original excerpt
publication date:2026-09-13 05:41:35  Der ChatGPT-Entwickler OpenAI wird nach Angaben seines Chefs Sam Altman nicht mehr im Jahr 2026 an die Börse gehen.©Axel Springer Deutschland GmbH. Alle Rechte vorbehalten.
Open original reportOpenAI is keeping its IPO on hold as Altman argues the company needs room to put AI safety ahead of shareholder interests (double slow-down?) He told Fortune that OpenAI will not go public in 2026 and
Original excerpt
OpenAI is keeping its IPO on hold as Altman argues the company needs room to put AI safety ahead of shareholder interests (double slow-down?) He told Fortune that OpenAI will not go public in 2026 and has discussed pauses at new capability levels to allow more progress on safety and alignment. Via Fortune I wonder what this means for data center expansion and investments for 2027 if the handbrake is being applied everywhere.
Open original report
Why it matters · interpretation
The delay signals that OpenAI is willing to forgo near-term public market capital to maintain control over safety decisions, and Altman has also discussed pausing at new capability levels, which could affect the pace of AI development and related investment.
Inspect 2 cited reports
IPO Delayed: Sam Altman Says an OpenAI Going Public in 2026 Would Be ‘Ill-Advised’
Original excerpt
publication date:2026-09-13 15:08:28  # IPO Delayed: Sam Altman Says an OpenAI Going Public in 2026 Would Be ‘Ill-Advised’ Rich Duprey Sun, September 13,
Open original reportOpenAI is keeping its IPO on hold as Altman argues the company needs room to put AI safety ahead of shareholder interests (double slow-down?) He told Fortune that OpenAI will not go public in 2026 and
Original excerpt
OpenAI is keeping its IPO on hold as Altman argues the company needs room to put AI safety ahead of shareholder interests (double slow-down?) He told Fortune that OpenAI will not go public in 2026 and has discussed pauses at new capability levels to allow more progress on safety and alignment. Via Fortune I wonder what this means for data center expansion and investments for 2027 if the handbrake is being applied everywhere.
Open original report
Story at a glance
DeepSeek V4.1 Flash tops open-source leaderboard, processes 1 trillion tokens in 24 hours
What happened
DeepSeek V4.1 Flash was released with nearly double the backbone parameters of V4 Flash but roughly a quarter of its global KV cache, according to an analysis published by Whats_AI.
Inspect 1 cited report
DeepSeek V4.1 Flash has almost twice V4 Flash’s backbone parameters, but about a quarter of its global KV cache. For agents carrying long conversations, that is worth understanding. I broke down the changes, costs, limits and the
Original excerpt
DeepSeek V4.1 Flash has almost twice V4 Flash’s backbone parameters, but about a quarter of its global KV cache. For agents carrying long conversations, that is worth understanding. I broke down the changes, costs, limits and the results on our own benchmark: https://t.co/eCxy2yJRta
Open original report
Why it matters · interpretation
The parameter increase paired with a much smaller global KV cache changes the cost and memory profile for agents that maintain long conversations, making the trade-off relevant for developers evaluating model limits and benchmark performance.
Inspect 1 cited report
DeepSeek V4.1 Flash has almost twice V4 Flash’s backbone parameters, but about a quarter of its global KV cache. For agents carrying long conversations, that is worth understanding. I broke down the changes, costs, limits and the
Original excerpt
DeepSeek V4.1 Flash has almost twice V4 Flash’s backbone parameters, but about a quarter of its global KV cache. For agents carrying long conversations, that is worth understanding. I broke down the changes, costs, limits and the results on our own benchmark: https://t.co/eCxy2yJRta
Open original report
Story at a glance
Anthropic blocks state-linked misuse of Claude AI for bioweapons, missile development, and surveillance
What happened
Anthropic published its most detailed threat intelligence report to date, disclosing that it detected and disrupted attempts to misuse Claude for cyberattacks, influence operations, surveillance, biological work, and weapons development, including a case in which the Houthis used Claude to vibecode guided-missile software.
Inspect 2 cited reports
Anthropic：我们发布了迄今为止最详尽的威胁情报报告。报告涵盖了人们如何试图滥用Claude——用于网络攻击、影响力行动、监控、生物学和武器制造——以及我们如何发现并阻止了这些行为。我们瓦解了报告中的每一项行动，并从中汲取教训来加强我们的防护措施。在适当情况下，我们还将发现的情况分享给了当局和其他人工智能公司。这些案例并不典型：我们重点展示的是所见中一些最复杂的滥用行为。但它们尤其值得讨论，因为它们向我们展示了人工智能滥用的走向、我们的防护措施在哪里有效，以及在哪里需要改进。我们发布这份报告，是为了
Original excerpt
Anthropic：我们发布了迄今为止最详尽的威胁情报报告。报告涵盖了人们如何试图滥用Claude——用于网络攻击、影响力行动、监控、生物学和武器制造——以及我们如何发现并阻止了这些行为。我们瓦解了报告中的每一项行动，并从中汲取教训来加强我们的防护措施。在适当情况下，我们还将发现的情况分享给了当局和其他人工智能公司。这些案例并不典型：我们重点展示的是所见中一些最复杂的滥用行为。但它们尤其值得讨论，因为它们向我们展示了人工智能滥用的走向、我们的防护措施在哪里有效，以及在哪里需要改进。我们发布这份报告，是为了让其他人能在自己的平台上发现同样的活动，也为了让公众更清楚地了解新兴威胁是如何发展的。阅读报告：https://www.anthropic.com/threat-intelligence-report-september-2026 Anthropic：We're publishing our most detailed threat intelligence report to date. It covers how people tried to misuse Claude—for
Open original reportJUST IN: Anthropic reveals the Houthis used Claude to vibecode guided-missile software.
Original excerpt
JUST IN: Anthropic reveals the Houthis used Claude to vibecode guided-missile software.
Open original report
Why it matters · interpretation
The report signals that sophisticated misuse of frontier AI models has extended into real-world weapons development, and Anthropic's decision to share findings with authorities and other AI companies indicates the industry is treating these cases as a shared security problem rather than isolated incidents.
Inspect 2 cited reports
Anthropic：我们发布了迄今为止最详尽的威胁情报报告。报告涵盖了人们如何试图滥用Claude——用于网络攻击、影响力行动、监控、生物学和武器制造——以及我们如何发现并阻止了这些行为。我们瓦解了报告中的每一项行动，并从中汲取教训来加强我们的防护措施。在适当情况下，我们还将发现的情况分享给了当局和其他人工智能公司。这些案例并不典型：我们重点展示的是所见中一些最复杂的滥用行为。但它们尤其值得讨论，因为它们向我们展示了人工智能滥用的走向、我们的防护措施在哪里有效，以及在哪里需要改进。我们发布这份报告，是为了
Original excerpt
Anthropic：我们发布了迄今为止最详尽的威胁情报报告。报告涵盖了人们如何试图滥用Claude——用于网络攻击、影响力行动、监控、生物学和武器制造——以及我们如何发现并阻止了这些行为。我们瓦解了报告中的每一项行动，并从中汲取教训来加强我们的防护措施。在适当情况下，我们还将发现的情况分享给了当局和其他人工智能公司。这些案例并不典型：我们重点展示的是所见中一些最复杂的滥用行为。但它们尤其值得讨论，因为它们向我们展示了人工智能滥用的走向、我们的防护措施在哪里有效，以及在哪里需要改进。我们发布这份报告，是为了让其他人能在自己的平台上发现同样的活动，也为了让公众更清楚地了解新兴威胁是如何发展的。阅读报告：https://www.anthropic.com/threat-intelligence-report-september-2026 Anthropic：We're publishing our most detailed threat intelligence report to date. It covers how people tried to misuse Claude—for
Open original reportJUST IN: Anthropic reveals the Houthis used Claude to vibecode guided-missile software.
Original excerpt
JUST IN: Anthropic reveals the Houthis used Claude to vibecode guided-missile software.
Open original report
Keep following
Turn a quick read into understanding
Select a passage to ask Page AI about it, or use “Explain key terms” and “Compare sources” in the Page AI panel.