LeituraPesquisaRadarFramework de investimento
Entrar / Cadastrar
TudoTecnologiaProdutosInvestimentoBolsaOutros

Leitura

2026-11-1392 posts

Atualizar

South Korea's LG Corp Reports 21.8% Drop in Q3 Operating Profit

Material
2026-11-12 14:58 GMT+8

South Korea's LG Corporation reported its consolidated financial results for the third quarter of 2026 on November 12. The holding company posted an operating profit of 327.7 billion won, marking a 21.8% decrease compared to the same period last year. Despite a 2.2% increase in revenue to 2.404 trillion won, the decline in profitability was attributed to weak performance in the chemical business segment.

LG Group explained that the sharp reduction in equity method income from its chemical affiliates was the primary driver behind the drop in operating profit. Equity method gains fell to 107.7 billion won in the quarter, a 51.8% plunge from 223.3 billion won in the previous year. Net profit attributable to controlling interests also decreased by 34.4% to 206.8 billion won.

Among consolidated subsidiaries, IT services firm LG CNS showed resilience, with cumulative revenue for the first three quarters reaching 4.5 trillion won, up 7% year-on-year, driven by cloud service expansion. Its cumulative operating profit rose 6% to 360.3 billion won, though this growth was insufficient to offset the broader drag from the chemical division.

Fontes:미래를 보는 창 - 전자신문

Pesquisa

Higress Integrates Qwen3Guard for Zero-Code AI Safety Gateway

Material
Verificado 2026-11-13 03:24 GMT+8

Alibaba Cloud's open-source cloud-native API gateway, Higress, has recently integrated the Tongyi Qianwen safety model Qwen3Guard-Gen as a Wasm plugin to provide AI content security capabilities. This solution allows enterprises to implement unified input and output safety moderation at the gateway layer without modifying application code or upstream model services.

In traditional AI application development, security policies are often scattered across individual application SDKs, making it difficult to manage risk thresholds and refusal logic centrally. Each model upgrade or policy adjustment requires redeploying applications. Higress places content extraction, safety calls, risk decisions, and streaming buffering logic within the data plane path, achieving decoupling and reuse of security capabilities.

The plugin defaults to calling the Qwen/Qwen3Guard-Gen-4B model and supports real-time moderation of non-streaming JSON responses and SSE streaming outputs. For streaming responses, the plugin uses a segmented buffering mechanism, triggering a moderation check every time a certain number of characters accumulate (default 1000 Unicode characters), balancing latency and security. If a risk is detected above the threshold, the gateway can directly return a refusal response, preventing the request from consuming backend inference resources.

Qwen3Guard is a family of safety moderation models launched by Alibaba's Tongyi Lab, trained on the Qwen3 architecture, covering risks such as violence, illegal acts, and privacy leaks. This integration emphasizes "self-hosting" features, allowing enterprises to independently deploy and scale safety models without relying on cloud-based content safety services, thereby better controlling data privacy and risk policies.

Fontes:Alibaba Cloud Community

Pesquisa

Study: AI Chatbots Recommend Fake Retail Sites in Up to 15% of Shopping Answers

Material
2026-11-13 01:00 GMT+8

Shoppers using AI chatbots to find products may be directed to imposters, according to a new study by IP protection firm Podqi shared exclusively with The Verge. The research tested ChatGPT, Google’s Gemini, and Perplexity with shopping-related questions for 151 brands, including Lululemon and Le Creuset.

Across 6,795 answers, the study found that all three engines cited risky sites—defined as fake impersonators, reregistered old domains, or dead links. Perplexity performed the worst, displaying risky sites in 15% of its answers. ChatGPT followed at 8%, while Gemini showed them least often at 4%. Smaller brands were found to be at higher risk than larger companies.

Podqi CEO Trevor West emphasized that consumers should verify sources before entering personal or payment information. Representatives from Google, OpenAI, and Perplexity acknowledged the findings, stating they are working to improve safeguards and information quality.

Fontes:The Verge

Pesquisa

Mercedes-Benz Halts Investments in German Sites and Weighs Plant Closures

Material
2026-11-13 02:44 GMT+8

Mercedes-Benz Production Board Member Michael Schiebe stated in an interview published on the company's intranet that the group has decided to "stop all further investments in the German locations" because negotiations with the works council and the IG Metall union over cost reductions have failed to yield progress. This move aims to pressure the employee side into agreeing to lower labor costs.

Schiebe noted that no "solution corridor" is currently visible and that investing in Germany is not feasible under the existing cost structures. He revealed that hourly production costs in Germany are up to five times higher than in Eastern European locations. If no agreement is reached, new models will be produced exclusively abroad.

Additionally, Human Resources Board Member Britta Seeger confirmed that management is examining "concretely and in detail" which German plants the group can still afford to maintain. While the company aims to avoid closures at all costs, it must now pursue this option more seriously. Previously, Schiebe had mentioned the possibility of closing two German plants, involving one assembly plant and one powertrain plant, if costs were not reduced.

The core of the conflict lies in Mercedes-Benz's "productivity offensive." Facing a significant drop in passenger car profits, management is demanding longer working hours without pay increases, proposing a shift from a 35-hour to a 40-hour work week. However, the 35-hour week is viewed as a historic achievement by the German metal industry union, which considers it a red line. Simultaneously, the union is seeking a 5% wage increase in the upcoming collective bargaining round, which management argues would further exacerbate Germany's high labor cost disadvantage.

Fontes:tagesschau.de

Pesquisa

Nvidia Ousts CEO Jensen Huang; CFO Colette Kress Named Interim Leader

Estrutural
2026-11-12 22:14 GMT+8

U.S. chipmaker Nvidia announced on Thursday that it has ousted Chief Executive Officer and Chairman Jensen Huang. This decision marks a dramatic leadership change at one of the world's most valuable companies, triggering severe concerns about the stability of the artificial intelligence infrastructure supply chain.

According to reports, Nvidia's Chief Financial Officer, Colette Kress, will serve as interim CEO while the board conducts a search for both internal and external permanent successors. Huang will also leave the board, with his chairmanship replaced by Harvey Jones, a director since Nvidia's founding in 1993 who previously ran chip design company Synopsys.

The ouster came four days after Huang admitted to an affair with a suspected Chinese spy, following a New York Times report of another affair with a subordinate. He is now subject to investigations by the board, Congressional leaders, and the U.S. Department of Justice. Reports suggest these probes involve allegations that Nvidia sold advanced chips to banned Chinese buyers through a Singaporean front company, intensifying Washington's scrutiny of export control enforcement.

Market reaction was sharp. Nvidia shares fell as much as 7% on Thursday, extending losses since the scandal broke. Neocloud firms like Nebius and CoreWeave, which rely on Nvidia for AI compute leasing, saw shares drop 19% and 9% respectively, while competitor Alphabet rose slightly. Analysts note that Huang's personal credibility underpinned billions in customer financing and AI investments, creating significant 'key-man risk' upon his sudden departure.

Fontes:semafor.com

Pesquisa

Financial Automation Company H6.ai: Freezing Real Invoice Status with Database Branches, Multi-Model Judge Regression Testing for AI Pipelines

Material
2026-11-13 00:56 GMT+8

In AI-driven accounts payable automation, the most thorny testing problem is not the code logic itself, but the fact that the data is constantly moving. In a Databricks official blog post, two H6.ai engineers, Benji Borg and Mauro Pirrone, described this dilemma: an invoice at 10:00 AM might match a purchase order with remaining quantity, but by afternoon that order could be partially consumed, modified, or closed. The same invoice, processed with the same code, yields different results. When regression tests show discrepancies, the team cannot determine whether the difference was caused by a code change or natural changes in customer data.

Their solution is not synthetic data or static fixtures, but preserving the real state. H6.ai runs an independent Lakebase Postgres database for each customer, storing purchase orders, suppliers, charts of accounts, historical transactions, and custom policies synced from the ERP. When the team discovers an invoice has been processed incorrectly, they create a copy-on-write branch of that customer's database as a snapshot. The branch does not duplicate any data; it shares storage with the parent database and only records new changes made on the branch. In the snapshot, that purchase order remains open, with its remaining quantity and price unchanged, even if it has long since been closed in the ERP.

The test run consists of four steps. First, migrate the snapshot to the current schema: the team applies the latest database migrations to the snapshot branch, then reuses this migrated branch as the starting point for all runs—migrate once, fork many times, while the snapshot itself never changes. Second, fork a short-lived run branch from the already-migrated branch for each run. Third, the candidate version of the pipeline reads and writes to this run branch just as it would a customer's live database, fully executing extraction, matching, classification, and validation. Fourth, after judging the result, delete the run branch while keeping the snapshot intact for the next run.

The judgment phase is the most notable part of this description. H6.ai does not merely check whether the invoice was processed, but uses two mechanisms to evaluate whether the result is correct. Deterministic assertions check facts that can be stated precisely: the final status of the document, which purchase order it matched, and the total amount. An AI judge panel spanning different model providers reviews the semantic consistency between the full output and the expected result, returning pass, fail, or inconclusive, along with reasoning. Together, these two mechanisms answer whether the outcome was right, not just whether the process finished.

The portability of this pattern lies in the fact that any AI pipeline relying on continuously changing real data for extraction, matching, or classification decisions faces the same regression testing dilemma. The idea of freezing state plus disposable branches shifts the problem from reconstructing conditions to preserving them. H6.ai also turns every anomalous invoice into a permanent test case—after fixing the issue, they retain its database snapshot, invoice file, and expected result. As customers and invoice formats increase, test coverage naturally expands.

Fontes:Databricks

Pesquisa

Multiple Analysts Cut Nvidia Ratings and Price Targets Amid CEO Uncertainty

Material
2026-11-12 19:13 GMT+8

A wave of analyst downgrades and price target cuts hit Nvidia this week, driven by growing uncertainty regarding the future of CEO Jensen Huang. Firms including William Blair, Oppenheimer, Stifel, Bernstein, Wells Fargo, Robert W. Baird, and Citi revised their outlooks, signaling that governance risks have become a primary factor in the stock's valuation.

On November 12, William Blair downgraded Nvidia from 'Outperform' to 'Market Perform,' warning that the potential departure of Huang and increased regulatory scrutiny could act as significant near-term headwinds. Oppenheimer also downgraded the stock to 'Perform' and removed its price target entirely, arguing that leadership changes could complicate execution and undermine investor confidence. Stifel maintained a 'Buy' rating but sharply reduced its price target from $315 to $205, now considering Huang's departure as a base-case scenario. Bernstein analyst Stacy Rasgon cut the target from $400 to $200, identifying the viability of Huang remaining in his position as the biggest risk.

Other revisions included Wells Fargo lowering its target from $315 to $300, Robert W. Baird cutting from $500 to $200, and Citi reducing from $285 to $225. While most analysts continue to recognize Nvidia's strong fundamentals in AI infrastructure, these moves highlight that management uncertainty has superseded underlying business performance as the key driver of valuation concerns.

Nvidia is scheduled to release its third-quarter earnings on November 17. The recent downgrades follow reports that the board is preparing to remove Huang, a move that could trigger independent investigations and briefings before the U.S. House of Representatives.

Fontes:Finbold

Pesquisa

BKFi Co-founder AJ Ayers: The Core Strategy for Irregular Income is to Pretend You Have a Salary

Visão rápida
2026-11-13 00:30 GMT+8

The core dilemma for those with irregular income is not that they earn too little, but that their earnings are unpredictable. AJ Ayers, co-founder of BKFi, pointed out on the "At the Money" podcast that traditional financial advice assumes a fixed paycheck arrives every two weeks, whereas freelancers, consultants, and creators have entirely different income patterns. She argues that the solution is not to wait for income to become regular, but to actively create regularity: the essence of financial planning is tricking yourself into believing you have a salary.

The specific approach is to treat income as if you are running a business. Ayers recommends setting up your own checking account and paying yourself a fixed amount every two weeks, prioritizing savings before spending. She observes that when a large sum of money suddenly arrives, the instinctive reaction is to spend it immediately—on a house, luxury items, or the next project. In reality, this money may be an advance on your earnings for the next three years, so it must be allocated strategically.

For windfalls, Ayers provides a clear sequence in her book: Step one, spend one percent on something that brings you joy; step two, do nothing and let the money sit quietly for a while; step three, pay taxes; step four, invest. She emphasizes that step three is often skipped—many creators receive a large check in their first year, fail to set aside taxes, spend the entire amount, and then face a massive tax bill, plunging them into debt. Her rule of thumb is: when in doubt, save 50% of every dollar earned for taxes; if that sounds excessive, save at least 30%.

Regarding long-term habits, Ayers uses the acronym ABS: Always Be Saving. She believes the key lies not in the amount saved, but in the habit itself. Even if you only save one dollar a week during lean times, establishing this habit allows you to quickly increase your savings rate when income rises or windfalls occur. Once you stop saving completely, you become entirely disconnected from financial success.

Ayers explicitly states that this method is not limited to artists and musicians. She notes in her book that consultants, freelancers, and professionals in many industries face the same irregular income challenges. The value of this framework is that it does not rely on external institutions to distribute income regularly, but instead empowers individuals to act as their own payroll department. It should be noted that these are recommendations based on Ayers' personal practice and consulting experience, not independently verified investment return strategies; tax withholding rates also vary significantly depending on specific tax regimes and income structures.

Fontes:The Big Picture

Pesquisa

Indian Public Library Faces Anti-Corruption Probe Over Alleged AI-Generated Books

Material
2026-11-13 00:37 GMT+8

The Periyar Arivulagam library in the Indian state of Tamil Nadu has become the subject of an anti-corruption investigation just days after its opening. The probe was triggered by accusations from internationally recognized Tamil novelist Perumal Murugan, who claimed that up to 1,800 books purchased by the library may have been partially or wholly generated using artificial intelligence.

According to TechRadar, citing BBC reporting, the books were purchased from a single publisher at unusually high prices. Many reportedly contained false information, suspicious author names, and questionable International Standard Book Numbers (ISBN). The library, built at a cost of 3 billion rupees ($34 million), opened on October 16.

Tamil Nadu's School Education Minister A Rajmohan stated that if the accusations are true, it could be "India's first AI scam." The state government has blamed the previous administration for the purchases, an allegation denied by the Dravida Munnetra Kazhagam party. Over 160 writers, researchers, artists, and publishers have signed a statement calling for an independent inquiry and a public release of procurement records.

Fontes:TechRadar

Pesquisa

US VC Firm Norwest Raises $4 Billion Fund to Expand Investments in Physical AI, Voice AI, and Agentic Cybersecurity

Tópico · Norwest风投基金Material
2026-11-12 23:18 GMT+8

According to TechCrunch, Norwest Venture Partners, a US multi-stage venture capital firm backed solely by Wells Fargo, has raised a $4 billion fund. This marks the largest fund in the firm's 65-year history and its eighteenth overall. The new vehicle aims to deepen its exposure to artificial intelligence, with planned investments across various sectors and countries, notably the United States, India, and Israel.

Jeff Crowe, a senior managing partner at Norwest, told TechCrunch that the firm is taking a broad approach to AI because "its impact is broad." He noted that AI is not limited by industry, geography, or application. Just as digital transformation did, AI transformation has the potential to touch most layers of enterprise, consumer, and healthcare solutions, including networking, enterprise IT, and applications. Adapting to an agentic environment means applying AI broadly across products and services that solve problems.

Crowe added that the firm is particularly excited about early-stage companies building AI-native applications and infrastructure, physical AI, domain-specific models, and cybersecurity solutions. Recent investments reflect this strategy, such as backing Coval, which builds tools to test and evaluate AI agents, and Generalist AI, a startup developing AI foundation models for robots.

Since closing its seventeenth fund, a $3 billion raise in 2024, Norwest has made more than 87 investments. The scale of this new fund indicates a significant shift in capital allocation towards AI and its related infrastructure and vertical applications for a firm traditionally known for a diversified portfolio.

Fontes:TechCrunch

Pesquisa

Baidu Launches ERNIE 5.5 Flash Preview and Token Exchange, Discloses Low-Cost API Pricing and Surge in Internal Agent Usage

Material
2026-11-12 22:08 GMT+8

Chinese technology company Baidu unveiled its latest foundation model, ERNIE 5.5 Flash Preview, and launched a new marketplace called "Token Exchange" at its annual World Conference on November 12, 2026. The announcements included major upgrades to its agent product line and a vision for the "intelligent organization" outlined by Co-founder and CEO Robin Li.

ERNIE 5.5 Flash Preview is a Mixture-of-Experts (MoE) model with 280 billion total parameters and 14 billion active parameters per token, supporting a 1 million token context window. Baidu claims strong performance in coding, 3D modeling, and content creation, optimized by its proprietary recursive self-improvement technology, ERNIE Evolve. Notably, Baidu disclosed specific API pricing: RMB 1 per million input tokens and RMB 4 per million output tokens, with cached input at just RMB 0.04. Model weights will be open-sourced by November 25. This aggressive pricing strategy suggests Baidu aims to expand its market share by lowering barriers to entry for developers.

Beyond the model itself, Baidu's Token Exchange serves as a neutral marketplace connecting AI model providers with customers. Users can compare model services side-by-side based on price and performance, accessing them through a single interface. According to Baidu, the platform has onboarded over 100 model resources in its first month, handling an average of more than 50 million calls and 700 billion tokens daily. This initiative addresses the difficulty of discovering high-quality model services and may help utilize underused compute resources.

Regarding internal applications, Robin Li disclosed concrete progress in Baidu's transformation into an "intelligent organization." Over the past year, the number of daily active agents (DAA) within Baidu surged from 5,000 to 37,000, an increase of over 600%. Weekly tasks completed by agents rose from 84,000 to more than 1 million. These figures provide investors with a quantifiable view of how Baidu is converting AI technology into operational efficiency. Additionally, Baidu upgraded its general-purpose agent DuMate and digital human agent Baidu Yijing, and presented a new roadmap for its Kunlunxin chips.

Fontes:prnewswire.com

Pesquisa

Logan Lawler, Head of AI Strategy for Dell Pro Precision Workstations: Practical Guide to Selecting Models for Agents — Benchmark Shortlists, Real-World Comparisons, and Keeping the Cheapest Option

Tópico · 戴尔智能体模型选型策略Material
2026-11-12 22:00 GMT+8

A small model was asked to install a tool, and it replied: 'You have a brain and I don’t.' Logan Lawler, Head of AI Strategy for Dell Pro Precision Workstations at Dell Technologies, recalled this testing experience in a blog post—the toolchain had fed the model all the information needed to complete the installation, yet the model completely failed to understand the task. This anecdote leads to his core judgment: in the formula where an agent equals a model plus a tool harness, the model is the brain, and even with all necessary conditions met, a wrong brain will fail.

His method consists of three steps. First, build a shortlist from benchmark websites. He recommends four: Artificial Analysis compares quality, price, and speed using independent tests, making it suitable for filtering by budget; Arena (formerly LMArena) allows real people to anonymously compare answers from two models and vote for the better one, reflecting actual preferences in everyday writing and Q&A; the Berkeley Function Calling Leaderboard tests how accurately a model uses tools (such as web search and file access)—since agents perform work through tools, this leaderboard is the most critical for agent scenarios; and SWE-bench tests whether a model can fix real bugs in open-source software, which is ideal for agents that need to write code. Second, run comparisons on real tasks. Lawler suggests using an OpenRouter account (which provides access to models from multiple providers via a single account) or LM Studio (which loads models locally) to test one frontier model and one or two smaller models, running the same task two to three times. Third, score each run across four dimensions: completion, correctness, amount of feedback needed, and cost. Keep the cheapest model that gets the job done right.

After scoring comes routing. Lawler cites specific data: when LangChain tested NVIDIA’s NeMo Switchyard routing library on 145 agent tasks, only 7% of the steps required a frontier model. This means that the vast majority of daily steps—calling tools, summarizing results—can be handled by small models. His personal production setup uses frontier models (via OpenRouter and his Claude plan) for designing agents and writing skills, because that represents 'hard thinking that requires payment'; once skills are built, they run daily on DeepSeek V4 Flash, deployed on a Dell Pro Precision GB10 workstation.

Lawler also defines boundaries. If comparison testing shows that only a frontier model can get the task done correctly, or if the task falls into the hardest categories of coding, reasoning, or research, do not force the use of open-source models. His exact words were: 'Don’t force an open-source model into a race it cannot win yet.' He also emphasized that he is constantly testing and advised readers to ask him again in a quarter.

Contextual note: This article was published on the official Dell Technologies blog. The author’s role is to lead the AI strategy for Dell Pro Precision workstations, and the NVIDIA Nemotron series models, NeMo Switchyard routing library, and Dell GB10 workstation recommended in the article are directly related to his commercial interests. The 7% figure from LangChain comes from its testing of the NVIDIA Switchyard routing library and is not an independent third-party assessment. While the methodology itself—building a shortlist, running comparisons, scoring across four dimensions, and keeping the cheapest option—is transferable, the specific tool recommendations should be viewed as suggestions carrying vendor bias.

Fontes:Dell

Pesquisa
Próxima →
Entrar / Cadastrar
LeituraPesquisaRadarFramework de investimento

Boundaries to preserve: This is a customer engineering blog published by Databricks as the vendor of the Lakebase product, where the H6.ai team describes its own practices. Architectural details are tied to the specific capabilities of the Lakebase product, and the consistency and accuracy of the multi-model judge panel's determinations are not provided with quantitative data. The overall effect is self-reported rather than independently benchmarked.

But the contribution at the methodological level is clear: when the correctness of an AI system depends on its handling of real context rather than isolated code logic, the testing environment needs to preserve the context itself. The copy-on-write feature of database branches reduces the cost of this from copying the entire production database to a single metadata operation, making it an engineeringly feasible daily practice for every run to start from the same frozen state.