読書リサーチレーダー投資フレームワーク
ログイン / 新規登録
すべて技術製品投資株式市場その他

読書

2026-11-0956 投稿

更新

EU VP: More AI Data Centers Needed but Grid Strain Must Be Avoided; US Local Ballots Restrict Development

トピック · AI数据中心社区阻力重大
2026-11-09 10:41 GMT+8

European Commission Executive Vice President Henna Virkkunen said in a Reuters interview on November 5 that Europe requires more data centers to build autonomous artificial intelligence (AI) computing capabilities, but construction must be managed carefully to avoid skyrocketing electricity bills for residents, power outages, or environmental damage. Her comments came amid coordinated protests by the 'Stop Dirty Data Centres' coalition across 16 European countries demanding moratoriums on large-scale AI data center development.

In the United States, opposition to data centers has translated into political action during the November 3 midterm elections. According to tracker website Data Centers on the Ballot, there were 47 local referendums related to data centers. Of the 45 counted as of November 8, restrictions on development won 44 times, while pro-development measures won only once. These issues covered bans on construction, permitting pauses, and land-use controls, indicating that public concern over the energy and environmental footprint of AI infrastructure is substantially hindering project deployment.

ソース:iThome

リサーチ

Semafor: Anti-Waymo PR Campaign Relied on Undisclosed AI-Generated Op-Eds

トピック · Waymo AI公关争议重大
2026-11-09 10:06 GMT+8

According to a report by technology media outlet Semafor published on November 8, 2026, opponents of autonomous driving company Waymo have been using artificial intelligence to generate opinion pieces for local US newspapers without disclosing the AI involvement to readers.

Analysis by the AI detection service Pangram showed that articles published in outlets such as the New York Daily News, The Dallas Morning News, and the San Francisco Chronicle exhibited significant characteristics of AI generation. For instance, two articles by lawyer Amy Witherite regarding autonomous vehicle accidents in The Dallas Morning News were detected as 95% and 71% AI-written, respectively. An article in the San Francisco Chronicle warning about potential autonomous vehicle paralysis during power outages was reported to be 88% generated by AI.

Several named authors responded to these findings. Ira Goldstein, former head of New York’s Black Car Fund, stated that the first draft of his op-ed in the New York Daily News was indeed prepared by a public relations firm, which he described as standard procedure when working with PR agencies. He noted that he reviewed and edited the piece and stands by its views. Goldstein also revealed that the PR firm in question later switched to working for Waymo.

David Margulies, a spokesperson for lawyer Amy Witherite, said her team sometimes uses AI tools to monitor media coverage, develop press lists, and prepare initial drafts, which are then reviewed, edited, fact-checked by staff, and approved by Witherite. He emphasized that all pieces published under her name reflect her personal views.

The incident has prompted media organizations to re-evaluate their policies on AI content. Ryan Sanders, commentary editor at The Dallas Morning News, stated that while limited use of AI for research or polishing does not automatically disqualify a submission, the paper does not accept pieces generated entirely by AI. Andrew Julien, executive editor at the Daily News, pointed out that given the publication's ongoing litigation against Microsoft and OpenAI, they aim to create stricter guardrails against AI, arguing that readers deserve original opinions rather than machine-generated content.

This investigation is part of a months-long battle between Waymo and its rivals over commercial operations in major US cities. The Washington Post reported earlier this month that the Coalition for Accountability and Road Safety, a group opposing Waymo, spent more than $2.6 million on ads in the DC area in 2026, exceeding the spending of the winning mayoral candidate. Waymo currently operates commercially in 15 US cities but faces opposition from labor unions, ride-hailing giants, and pedestrian safety advocates.

ソース:semafor.com

リサーチ

Microsoft Discloses AI-Driven Vulnerability Discovery, Fixing Over 1,100 CVEs in October Windows Update

トピック · 微软AI漏洞发现重大
2026-11-09 09:32 GMT+8

US technology company Microsoft disclosed on November 5 via its official blog a new strategy for managing Windows system vulnerabilities in the AI era. The company stated that in its monthly security update for October, it fixed more than 1,100 Common Vulnerabilities and Exposures (CVEs) for Windows systems. While this number is significantly higher than previous months, Microsoft emphasized that this does not indicate a decline in Windows security, but rather reflects the benefits of broader AI-assisted detection.

Microsoft noted that attackers are increasingly using AI to find unpatched vulnerabilities or analyze released patches to quickly generate malware. As a defensive measure, Microsoft has expanded AI-driven vulnerability exploration across the entire Windows codebase. This process utilizes an agentic code scanner named "MDASH" to assist engineers at various stages, from discovery to remediation and verification. Although AI has significantly increased scanning efficiency, Microsoft stated that final judgments and decisions remain the responsibility of human engineers.

As AI accelerates vulnerability discovery, the time window after a patch release is also shrinking. Microsoft mentioned that once a fix is public, attackers may parse it and develop exploit code within hours. However, diagnostic data from August showed that more than half of enterprise Windows devices had not applied updates two weeks after a patch was released. Microsoft urged users to stay on supported versions of Windows and apply security updates promptly to counter the increasingly complex threat landscape.

ソース:PC Watch

リサーチ

Japanese Chemical Giant Asahi Kasei Announces AI-Powered Smart Lab to Quadruple Semiconductor Material R&D Experiments

トピック · 旭化成智能实验室クイックテイク
検証済み 2026-11-09 09:33 GMT+8

Asahi Kasei Corporation, a major Japanese chemical and materials company, announced on November 5, 2026, that it will establish a "Smart Laboratory" at its branch in Fuji City, Shizuoka Prefecture. Scheduled to begin operations in January 2027, the facility will leverage artificial intelligence and automation technologies to accelerate the product development of "Pimer," its core photosensitive insulating material for semiconductors.

According to Asahi Kasei's official release, the new laboratory features a closed-loop system that automates the entire process from material design and experimental planning to raw material mixing, sample preparation, and performance evaluation simulating customer semiconductor packaging processes. The system automatically collects experimental data for machine learning, which then optimizes the next round of compositions and conditions. The company states that this autonomous experimental system can expand the number of development experiments (formulation combinations) conducted within a given period to approximately four times that of traditional methods.

This initiative responds to the increasing complexity and diversity of material performance requirements driven by advanced semiconductor innovations fueled by generative AI. In its "Medium-Term Business Plan 2027," Asahi Kasei has designated electronic materials, including Pimer, as a key growth business. The interlayer insulation film market is currently growing at an annual rate of about 8%, a trend expected to continue until around 2030. By adopting an AI-driven automated R&D platform, Asahi Kasei aims to overcome the efficiency bottlenecks of traditional manual research when facing vast variables, thereby responding more quickly to market demand for high-performance, highly reliable semiconductor packaging materials.

ソース:旭化成株式会社

リサーチ

Lenovo and HP Out of Stock on RTX 5090 Desktops as GPU Prices Triple

トピック · RTX 5090缺货涨价重大
2026-11-09 05:05 GMT+8

According to TechRadar, Nvidia's flagship consumer graphics card, the GeForce RTX 5090, is facing severe supply shortages in the United States. Both Lenovo and HP currently show no readily purchasable desktop configurations equipped with the RTX 5090 on their US online stores. High-end models like HP's OMEN MAX 45L, listed at $7,999.99, are marked as out of stock.

Availability of standalone graphics cards is equally critical. Major retailers such as Newegg and B&H Photo Video list multiple RTX 5090 models as unavailable or temporarily out of stock. On Amazon, the only available ASUS TUF Gaming RTX 5090 is priced at $7,679.80, approximately 284% higher than its original launch price of $1,999. Listed prices for MSI Suprim and ASUS TUF models on Newegg stand at $5,899.99 and $5,099.99 respectively, representing increases of over 150%.

While Dell's Alienware series and independent system builders like Skytech and CLX still offer complete systems containing the RTX 5090, prices generally range between $9,000 and $13,500. This shortage affects not only gamers but also professionals relying on the card's 32GB GDDR7 memory for local AI development and graphics rendering. Although Nvidia has not officially confirmed permanently ending RTX 5090 production, market observations indicate it is increasingly disappearing from mainstream retail channels.

ソース:TechRadar

リサーチ

Nvidia Board Launches Independent Investigation into CEO Jensen Huang's Ties with Suspected Chinese Spy

重大
2026-11-09 04:22 GMT+8

Nvidia's Board of Directors announced on November 8, 2026, that it has launched an independent investigation into Founder and CEO Jensen Huang. This move responds to recent reports alleging that Huang was the target of a foreign intelligence effort. According to The New York Times, U.S. intelligence agencies discovered that Alice Huang, an executive at Megaspeed, was recruited by China's Ministry of State Security as a spy and had developed a "close relationship" with Jensen Huang.

In a statement, the Nvidia Board said the investigation would be overseen by a committee of independent directors assisted by outside independent counsel. The Board emphasized that U.S. national and economic security are paramount to the interests of Nvidia and its shareholders. While the Board stated it has no reason to believe that either national security or Nvidia secrets were compromised, the inquiry will examine all available information. Jensen Huang fully supports the review and has committed to cooperating with it.

This internal governance action comes amid increasing external political pressure. Earlier this week, leaders of three Republican-led committees in the U.S. House of Representatives sent a letter to Huang requesting a closed-door briefing on his relationship with the suspected Chinese intelligence asset. Megaspeed, a customer of Nvidia, is currently subject to a Commerce Department investigation into whether it smuggled AI chips to China. An Nvidia spokesperson previously stated that regional cloud customers receive ordinary business information, not sensitive trade secrets or data relevant to national security.

ソース:The Verge

リサーチ

Mathematical Framework for Transformer Circuits: Residual Stream and Induction Head Mechanisms

検証済み 2026-11-09 00:07 GMT+8

長期読解 · 《A Mathematical Framework for Transformer Circuits》(2021)

The 2021 paper "A Mathematical Framework for Transformer Circuits" was written by the Anthropic team. It aims to reverse-engineer Transformer models using mechanistic interpretability methods. The authors compare this process to a programmer decompiling complex binary files into human-readable source code, attempting to understand the specific computational logic within the model.

To simplify analysis, the research focuses on "attention-only" toy models, excluding complex components such as MLP layers and layer normalization. This simplification allows researchers to clearly observe interaction patterns between attention heads, although the authors acknowledge this is one of the main weaknesses of the current work, as real large models contain numerous non-linear MLP structures.

One core concept is the "residual stream." In Transformers, each layer reads information from the residual stream via linear projections and adds the results back into the residual stream. The authors emphasize that the residual stream has a profound linear structure and can be viewed as a communication channel where all components read from and write to different subspaces. This perspective allows us to decompose the model into end-to-end interpretable functions, making it easier to reason about model behavior.

Another key finding is "induction heads." Research shows that single-layer attention models can only implement simple bigram or skip-trigram patterns, while models with two or more layers can form more complex algorithms through combinations of attention heads. Specifically, induction heads appear only in models with at least two layers of attention. They can explain the phenomenon of in-context learning in small models, i.e., how the model predicts subsequent content based on preceding prompts.

For modern AI developers, understanding these fundamental mechanisms helps in debugging model behavior. For example, when a model exhibits unexpected copying behavior, one can detect whether it is performing some form of "copying" by analyzing the eigenvalues of OV circuits. Although this framework currently applies mainly to small models, it provides preliminary mathematical tools and intuition for understanding larger, more realistic models, particularly in identifying safety issues and anomalous behaviors.

Engineers engaged in large model safety and interpretability research are advised to read "A Mathematical Framework for Transformer Circuits" (2021). Starting with the linear structure of the residual stream and the formation conditions of induction heads, they can build an intuitive understanding of the internal mechanisms of Transformers.

ソース:《A Mathematical Framework for Transformer Circuits》(2021)

リサーチ

"The Cost of Capital, Corporation Finance and the Theory of Investment": A Standard for Maximizing Market Value Under Uncertainty

検証済み 2026-11-09 00:07 GMT+8

長期読解 · 《The Cost of Capital, Corporation Finance and the Theory of Investment》(1958)

In 1958, Franco Modigliani and Merton H. Miller published this groundbreaking paper, challenging conventional views on capital costs and corporate finance at the time. Prior to this, economists often assumed that asset returns were certain, thereby simplifying the cost of capital to bond interest rates. However, the real world is full of uncertainty, and the traditional standard of "profit maximization" becomes difficult to operate in an uncertain environment because each decision corresponds to a distribution of possible outcomes rather than a single value.

The core mechanism proposed by the paper is a shift toward "market value maximization." The authors point out that under conditions of uncertainty, any investment project and its financing plan only need to pass one test: Will the project increase the market value of the company's stock? If the answer is yes, it is worth undertaking; otherwise, its return is below the company's marginal cost of capital. This standard is independent of the current owner's personal preferences because market prices reflect the preferences of all potential owners. Even if existing shareholders disagree with management's valuation, they can sell their shares and reinvest, yet still benefit from the capital appreciation brought about by management's decisions. This definition based on market value provides an operational concept of the cost of capital, resolving the normative and analytical defects of subjective utility approaches.

Furthermore, the paper explores the relationship between capital structure and firm value. Under a series of strict assumptions (such as atomistic competition and easy access to capital markets), the authors argue that a company's market value is unrelated to its capital structure. This means that simply changing the proportion of debt to equity will not change the total value of the company or its weighted average cost of capital. Although these assumptions do not fully hold in reality, this theoretical framework forces managers to focus on the profitability of the assets themselves, rather than merely manipulating financial statements through financial engineering. It reveals why, in some cases, increasing debt does not lower the overall cost of capital, but depends on whether the project itself can enhance market value.

For today's readers, understanding this theory helps answer a common investment puzzle: When a company announces a large-scale stock buyback or issues new debt, does this mean its intrinsic value has changed? According to the MM theory, in a frictionless perfect market, such financial restructuring does not create value on its own; value creation stems from improvements in operational efficiency. For example, suppose a mature technology company announces borrowing money to repurchase shares. Investors should first examine whether its cash flow generation capability has improved, rather than focusing solely on changes in leverage. If operations remain unchanged, fluctuations in the stock price more likely reflect market sentiment or short-term liquidity needs rather than fundamental shifts in the business. This perspective helps investors distinguish between "financial tricks" and "real value creation."

ソース:《The Cost of Capital, Corporation Finance and the Theory of Investment》(1958)

リサーチ

このビューで最新の記事まで読みました

ログイン / 新規登録
読書リサーチレーダー投資フレームワーク

Recommended reading: "The Cost of Capital, Corporation Finance and the Theory of Investment," published in 1958. Suitable for investors and managers who wish to deepen their understanding of the foundations of modern corporate finance and evaluate the impact of corporate financing decisions on long-term value. It is recommended to start with the introduction, focusing on discussions regarding decision-making standards under uncertainty, as well as the logical derivation of capital structure irrelevance in subsequent sections.