ReadingResearchRadarInvestment framework
Sign in / Sign up
AllTechnologyProductsInvestingPublic marketsOther

Reading

2026-10-29198 posts

Refresh

Gartner Report: Only 3% of Organizations Hit AI Productivity and Savings Goals

Material
2026-10-28 22:05 GMT+8

Research and advisory firm Gartner has released a new report indicating that only 3% of organizations say they have successfully met their targets for productivity, efficiency, and cost reduction through the use of artificial intelligence (AI). This figure contrasts sharply with widespread corporate enthusiasm: approximately 75% of leaders expect AI to have a significant impact on their organization, and 85% plan to increase AI spending in 2027, even though 38% are already over budget.

"Ultimately, organizations will not generate cost savings from AI at this point," said CV Viverito, Senior Director Analyst for Gartner's HR practice, speaking at the firm's HR Symposium. He suggested that the greater opportunity lies in using AI to do things that weren't possible before, such as shifting employee time into new capabilities, outcomes, and sources of growth. Viverito noted that the struggle for organizations is not creating capacity—which is happening as work is redesigned—but rather making that capacity usable and redirecting it toward valuable ends.

The report advises businesses to focus their AI efforts where the technology can make a tangible effect in key areas and to spread AI knowledge across the entire business, emphasizing that C-suite executives are as important as entry-level workers in identifying new use cases.

Sources:TechRadar

Research

Postgres Core Committers Number Only About 50; pg_lake Developer Craig Kerstiens Says AI-Generated Code Is Exhausting Review Capacity

Topic · PostgresMaterial
2026-10-28 23:00 GMT+8

Postgres's success rests on a seemingly contradictory design choice: the core engine is deliberately kept "boring," with new features pushed primarily into the extension ecosystem. This strategy allowed Postgres to grow from an academic project into the most widely deployed open-source relational database globally, but it also means there are very few guardians of the core code. In Episode 688 of The Changelog podcast, Craig Kerstiens, developer of the pg_lake project, warns that this structure is facing unprecedented pressure in the AI era.

Kerstiens describes a specific numerical relationship: fewer than 100 people directly participate in Postgres core development, and only about 50 actually hold commit permissions. Meanwhile, the ratio of application developers using Postgres to database professionals has risen from 10:1 twenty years ago to 100:1 today, and is approaching 1000:1. This means the vast majority of Postgres users lack the ability to review core changes, while the team of maintainers who possess this capability is extremely small.

The influx of AI-generated code exacerbates this imbalance. Kerstiens argues that AI-generated C code is exhausting the review capacity of these roughly 50 committers. Postgres's contribution process still relies on mailing lists, without a pull request mechanism; every line of code entering the core requires human committers to review it line-by-line and understand its design decisions. When AI tools batch-generate patches that look reasonable but lack architectural understanding, the review burden does not grow linearly but expands in ways far beyond the maintainers' processing capacity.

This judgment is not isolated evidence. Tom Lane, a Postgres core committer, independently described the same phenomenon in a thirty-year anniversary interview published by Snowflake: "We are receiving a huge number of AI-generated bug reports and security reports... a significant portion of which is slop because the tools do not understand basic architectural concepts." Lane also revealed that the project recently convened over twenty core members to specifically discuss whether to accept AI-generated submissions. The consensus reached so far is that a human must still be accountable for every line entering the codebase and able to explain every decision within it—but this has not yet become formal project policy.

Kerstiens's argument reveals a broader open-source governance dilemma: Postgres's "boring core" strategy was a shrewd engineering choice in the pre-AI era—it lowered the barrier to contribution, attracted many extension developers, and maintained core stability and auditability. But when AI drastically reduced the cost of generating code that "looks like a reasonable contribution," the review bottleneck of this strategy was exposed. The prosperity of the extension ecosystem does not depend on the bandwidth of core committers, but any change touching the core must still pass through the review of those ~50 people.

It should be noted that this episode has not yet released a transcript; the specific numbers and arguments above come from secondary summaries and show notes of this episode by The Podcast Summary. Kerstiens's 1000:1 ratio is his personal observation rather than strictly statistically measured data. The figure of approximately 50 committers aligns with the email-driven, self-selecting small-team governance model described in Tom Lane's interview, but the exact number fluctuates over time.

Sources:changelog.com

Research

Etisalat Posts Record Net Profit, Sells Full Vodafone Stake and Pivots to AI Infrastructure

Material
2026-10-29 00:58 GMT+8

UAE telecommunications giant Etisalat released its financial results for the third quarter and the first nine months of 2026 on October 28, highlighting a significant boost in profitability. Consolidated net profit for January through September reached AED 17.2 billion (approximately $4.68 billion), a 45.9% increase from the same period last year. Consolidated revenue grew by 10.0% to AED 57.6 billion. This strong performance was partly driven by major adjustments in the company's investment portfolio.

The most notable financial event this quarter was Etisalat's completion of the sale of its entire stake in UK telecoms group Vodafone. The transaction, including final dividends, generated total proceeds of AED 21.9 billion and realized a gain of approximately AED 8 billion. This substantial inflow directly boosted quarterly net profits and significantly enhanced the company's financial flexibility, providing ample cash reserves for future strategic investments. Excluding one-off gains from Vodafone, adjusted net profit for the first nine months stood at AED 9.2 billion, with earnings per share of AED 1.06, indicating robust growth in core operations.

Alongside the earnings report, Etisalat announced a brand refresh to mark its 50th anniversary, adopting the unified name 'etisalat' and unveiling a new corporate vision: to become a leading telco powered by AI and supporting AI applications. CEO Masood M. Sharif Mahmood stated that as AI becomes deeply integrated into enterprise and government workflows, the importance of connectivity, computing power, and security infrastructure grows. The company plans to concentrate its accumulated capabilities in digital services, cloud computing, cybersecurity, and fintech into an AI-centric communications infrastructure.

To achieve this strategic goal, Etisalat outlined specific infrastructure expansion plans. Aligning with the UAE's national objective to become a global AI hub, the company aims to expand international communication capacity to over 500 terabits per second (Tbps) by 2030, more than 25 times its current level. Additionally, it will develop diverse communication paths to improve efficiency and stability, aiming to provide low-latency performance to support a knowledge-based economy. This move signals traditional telecom operators' attempt to secure a more central position in the AI value chain by controlling underlying compute networks and data transmission channels.

Sources:prnewswire.com

Research

ServiceNow Q3 Earnings: Subscription Revenue Grows 21%, AI Annual Contract Value Target Raised to $1.7 Billion

Topic · ServiceNowMaterial
Verified 2026-10-29 06:06 GMT+8

US enterprise software company ServiceNow (NYSE: NOW) released its third-quarter 2026 financial results on October 28, 2026, reporting subscription revenues of $3.988 billion, representing 21% year-over-year growth (20.5% in constant currency). Total revenues reached $4.106 billion, growing 20.5% year-over-year. The company raised its full-year AI annual contract value (ACV) target again, increasing it from $1 billion at the start of the year to $1.7 billion, a 70% increase from the initial goal.

As of September 30, 2026, current remaining performance obligations (cRPO) stood at $13.65 billion, up 20.5% year-over-year, while total remaining performance obligations (RPO) reached $29.7 billion, up 22%. This metric reflects the level of future service demand locked in by customers. During the quarter, ServiceNow closed 19 transactions with net new ACV exceeding $5 million, representing over 135% year-over-year growth. The number of customers with more than $5 million in ACV grew approximately 23% year-over-year to 680.

ServiceNow Chairman and CEO Bill McDermott stated that this quarter was not just a "beat-and-raise" but an inflection point for the company as the "Super Intelligence Control Tower." CFO Gina Mastantuono noted that AI net new ACV exceeded expectations, with agentic AI usage in production surging more than 13x year-over-year. Despite reducing capital expenditures, the company raised its full-year free cash flow margin target by 100 basis points to 36%.

In terms of ecosystem partnerships, ServiceNow announced an expanded partnership with OpenAI, launching a ChatGPT plugin enabled by Action Fabric and planning to develop security solutions powered by OpenAI cyber models. Additionally, the company unveiled a new AI Passport solution to verify first- and third-party AI agents’ scopes in real-time, enhancing enterprise AI governance.

Sources:sec.gov

Research

Vercel AI Gateway Adds Particle-1 for Efficient Software Engineering Tasks

Topic · Particle-1模型Quick take
Verified 2026-10-29 06:05 GMT+8

Vercel has announced that its AI Gateway now supports Particle-1, a model from Particle AI. Post-trained on DeepSeek V4.1 Flash, the model is designed to complete software engineering tasks such as debugging, planning code changes, and breaking down multi-step problems using fewer tokens.

Particle-1 features a 1 million-token context window, suitable for working with large codebases and extended task histories. Developers can access it by setting the AI_GATEWAY_API_KEY environment variable and using particle/particle-1 as the model name. It is also compatible with coding agents like Claude Code, Codex, and Cursor via the latest Vercel CLI setup.

Vercel states that AI Gateway provides a unified API for calling models, tracking usage, and configuring retries without charging a platform fee or markup on inference. The claimed gains in shorter reasoning and reduced trial-and-error rounds are based on vendor descriptions, with no independent benchmark results provided.

Sources:Vercel

Research

WRITER Study: AI Medical Q&A Almost Never Cites Pharmaceutical Company Content, Suggests Shifting Marketing Budget to Academic Publications

Topic · 药企AI学术发表策略Quick take
2026-10-21 23:13 GMT+8

Product pages and display ads carefully optimized by pharmaceutical companies are almost entirely absent from AI medical Q&A systems. This is the core finding of a citation analysis released in October 2026 by WRITER. The study utilized the WRITER Agent tool to submit 40 health-related queries to four consumer-grade AI engines: Gemini/Google, ChatGPT, Perplexity, and Claude. These queries covered four therapeutic areas—oncology, obesity, immunology, and neurology—with half oriented toward patients and half toward physicians. All 777 cited sources were then categorized and sorted by type.

The results showed that when AI engines answered questions from physicians, 19.5% of citations originated from academic journals—a frequency four times higher than in patient-oriented queries. For physician-facing responses, the most frequently cited domain was PubMed/PMC (cited 50 times), whereas for patients, it was GoodRx (cited 15 times). The only domain appearing in both top lists was fda.gov. Commercial and manufacturer content remained virtually invisible across both query types. WRITER summarized this pattern as follows: AI search synthesizes medical answers from high-trust third-party sources such as academic literature, institutional websites, and reference platforms, while deprioritizing commercial and manufacturer content.

Based on this discovery, WRITER proposed strategic recommendations for pharmaceutical companies: shift a portion of digital marketing budgets away from traditional SEO, display advertising, and sponsored content toward evidence generation, specifically by increasing the volume of peer-reviewed publications. The article emphasizes the importance of publication volume—when an AI engine synthesizes answers from hundreds of sources, having only a dozen papers in a therapeutic area will not attract attention. WRITER recommends a journal portfolio strategy: continue targeting top-tier journals such as The New England Journal of Medicine, The Lancet, and JAMA, while simultaneously submitting to mid-tier journals with shorter acceptance cycles to accelerate the indexing of content. Conference abstracts and posters, which are also indexed in PubMed, can serve as additional citable content sources.

WRITER also recommended using AI-assisted writing to expand publication capacity, arguing that manual medical writing teams cannot produce the volume of publications required to influence AI-generated answers. The article acknowledges that peer-reviewed research has explicitly listed the risks associated with AI participation in writing—including hallucinated references, bias amplification, and exposure of confidentiality—but cites the same body of research confirming that AI can accelerate manuscript preparation in stages such as literature discovery, citation validation, and editorial refinement, provided human experts remain involved at every stage.

It is important to note that this study has significant limitations. The sample size of 40 queries is small, the full methodology is marked as available upon request but not publicly disclosed, and WRITER itself is a commercial supplier of AI writing tools; thus, its recommendation that AI is necessary to publish at this scale aligns directly with its own product positioning. The article cites research from DataForSEO as independent corroboration, noting that health content is particularly suitable for citation and that large language models tend to favor medically authoritative sources such as the CDC, Mayo Clinic, and peer-reviewed publications. However, DataForSEO’s specific methods and sample sizes are not detailed in this article. The core mechanistic finding—that AI engines prefer academic and institutional sources while deprioritizing commercial content—is consistent with broader discussions on AI search visibility. WRITER’s contribution lies in providing specialized data for the healthcare sector and concrete operational pathways for medical affairs teams.

Sources:WRITER

Research

Anthropic Expands 'AI for Science' Program to 20 Foundations and Research Institutions

Material
2026-10-28 08:00 GMT+8

On October 28, 2026, Anthropic announced it is expanding its "AI for Science" program from supporting individual scientists to partnering with 20 scientific foundations, research institutions, and specific research projects worldwide. The program provides free Claude credits to researchers working on high-impact scientific projects.

New partners include major science funders such as the Alfred P. Sloan Foundation, founded in 1934; the Simons Foundation, which advances mathematics and basic sciences through grants and its Flatiron Institute; Schmidt Sciences, founded in 2024 by Eric and Wendy Schmidt; and the Open Source for Science Fund, a program of Renaissance Philanthropy. Notably, the list includes the US National Science Foundation (NSF), a federal agency accounting for about 25% of federal support for basic research at US colleges and universities, as well as CAS, a division of the American Chemical Society.

Leading research institutions joining the program include the Space Telescope Science Institute (STScI), which runs science operations for NASA’s Hubble, James Webb, and Roman space telescopes; the Wellcome Sanger Institute in the UK; the Francis Crick Institute in London; Imperial College London’s Department of Life Sciences; the European Molecular Biology Laboratory (EMBL); and Gladstone Institutes in San Francisco.

Anthropic is also directly supporting specific initiatives, such as FlyWire at Princeton University (neuroscience), LEAP at Columbia University (Earth system modeling), AstroPath at Johns Hopkins University (cancer research), gnomAD at the Broad Institute (genomics), Tau Ceti (mathematics autoformalization in Lean), and the Centre for Population Genomics in Australia.

The company stated that it has previously partnered with HHMI Janelia Research Campus to develop the Model Hardware Standard (MHS), a specification allowing AI agents to safely operate lab instruments, and supports the Allen Institute in building AI agent teams for data analysis. Anthropic noted that many of these new collaborations aim to produce public goods for the wider scientific community, including open datasets, open-source software, open standards, and shared libraries.

Sources:anthropic.com

Research

AllianceBernstein CAIO Andrew Chin discusses AI governance and portfolio convergence with CFA Institute

Material
2026-10-29 00:00 GMT+8

CFA Institute has published a new episode titled "AI and the Future of Investment Decision-Making," featuring Andrew Chin, the Chief Artificial Intelligence Officer at AllianceBernstein. The conversation is hosted by Frank J. Fabozzi and addresses how artificial intelligence is reshaping asset management practices.

According to the publisher's description, the episode explores the integration of large language models into research, scenario analysis, and portfolio management processes. A central theme is the necessity of maintaining human accountability for investment decisions despite increasing automation.

Chin also examines specific industry challenges such as portfolio convergence and automation bias. The discussion highlights the importance of governance, explainability, and critical thinking as firms scale their AI capabilities. Additionally, it touches upon the fundamental law of active management and the skills required for investment professionals to effectively combine AI tools with domain expertise.

Sources:youtube.com

Research

Microsoft Reports Q1 Revenue of $92.3 Billion as Azure Grows 48%

Material
Verified 2026-10-29 05:38 GMT+8

U.S. technology giant Microsoft announced its financial results for the first quarter of fiscal year 2027, ended September 30, 2026, on October 28. The company reported total revenue of $92.3 billion, an increase of 19% year-over-year. Operating income rose 21% to $45.9 billion, while net income increased 32% on a GAAP basis to $36.6 billion.

In terms of cloud performance, Microsoft Cloud revenue reached $63.9 billion, growing 30% year-over-year. Notably, Azure revenue accelerated by 48%, reflecting sustained and increasing demand for artificial intelligence infrastructure. Commercial remaining performance obligation (RPO) also surged 73% to $677 billion, indicating strong visibility into future contracted revenue.

It is worth noting that the 32% increase in GAAP net income was partially driven by changes in the valuation of Microsoft's investments in OpenAI. In this quarter, net gains from these investments added $207 million to net income, whereas in the same period last year, they had resulted in a $3.1 billion loss. Excluding this impact, non-GAAP net income was $36.4 billion, up 18% year-over-year.

Sources:microsoft.com

Research

Big Data Firm Cloudera Launches Preconfigured AI Blueprints and Expands Partnerships with Nvidia, Neo4j, and Mistral AI

Topic · ClouderaQuick take
2026-10-29 03:21 GMT+8

U.S.-based big data platform company Cloudera Inc. announced on October 28, 2026, at its Evolve26 conference in New York that it is expanding its enterprise artificial intelligence ecosystem. The centerpiece of this update is "Cloudera Blueprints," a portfolio of preconfigured AI application packages that combine Nvidia Corp.'s AI reference workflows with Cloudera's data management, security, and governance capabilities.

Wim Stoop, Senior Director of Product Marketing at Cloudera, stated that these initiatives are designed to help enterprises move AI projects into production while maintaining control over their data, infrastructure, and deployment costs. "We'll bring AI to where the data is and where the data lives," Stoop said. Built using the Nvidia Agent Toolkit and deployed through the Cloudera AI development platform, the blueprints aim to reduce the integration work required to turn experimental AI applications into production systems.

In addition to the new blueprints, Cloudera expanded its partnerships with graph database company Neo4j Inc. and French AI model firm Mistral AI SAS. Neo4j's knowledge graph technology provides semantic context to help AI agents understand relationships within enterprise data, while the partnership with Mistral AI emphasizes a "sovereign AI" strategy, allowing customers to use open-weight models within their own environments to meet data privacy and geopolitical compliance requirements, particularly in Europe.

The initial blueprint portfolio includes applications for financial transaction analysis, automated clinical documentation, telecommunications network configuration, and synthetic video detection. Cloudera noted that as AI applications scale, inference computing costs can grow substantially, making flexible deployment options across public clouds, private clouds, and on-premises infrastructure critical for managing economics.

Sources:SiliconANGLE

Research

RAND Experts Analyze US ASI Strategy: Why 'Freedom of Action' Beats Single Strategies?

Topic · 兰德ASI战略研究Material
2026-10-29 02:57 GMT+8

In this episode of The Cognitive Revolution, host Nathan Labenz interviews Joel Predd (Director), Ben Boudreaux (Policy Researcher), and Matt Chessen (Resident Technical Expert) from the RAND Corporation's Center for the Geopolitics of AGI. They discuss their recent report, 'A U.S. Strategy to Secure Geopolitical Advantage on an Uncertain Path to Superintelligence: Maintaining Freedom of Action.'

The guests argue that current AI strategy debates often fall into binary oppositions, such as accelerationism versus pausing development. To address this, they propose a taxonomy of seven strategic archetypes: Dominance, Co-development, Preparedness, Moratorium, Deterrence, Continuity of Society, and Acceleration. Each archetype has a coherent win state but also a specific pathology or risk.

To determine which strategy is most appropriate, the team identifies five 'pivot factor' questions: Is danger close? Is coexistence feasible? Is restraint feasible? Is a decisive strategic advantage reachable? Is suppression feasible? Since evidence cannot currently resolve these questions with consensus, and the timeline is tight, committing prematurely to any single strategy risks foreclosing necessary options.

Consequently, they recommend a 'Freedom of Action' strategy. This approach focuses on building long-lead capabilities—such as verification technologies, evaluation systems, and crisis communication infrastructure—to maintain the flexibility to switch strategies as the situation evolves. Drawing parallels to nuclear history, they suggest that preserving optionality is crucial when facing profound uncertainty.

Sources:cognitiverevolution.ai

Research

Cisco Joins ServiceNow Control Tower Alliance as Founding Member

Quick take
2026-10-29 04:09 GMT+8

US networking company Cisco announced on October 28, 2026 that it has joined the ServiceNow Control Tower Alliance as a founding member. The alliance brings together companies across AI, cloud infrastructure, security, and governance to help customers adopt AI safely at scale through ServiceNow's Super Intelligence Control Tower. Two Cisco AI Defense integrations—importing discovered AI assets into the Control Tower inventory and sending security detections to ServiceNow AI Security Exposure Management—have been available on the ServiceNow Store since April 2025. This membership formalizes existing capability under a named multi-vendor framework. The alliance's other members and governance structure have not been disclosed.

Sources:Cisco Blogs

Research
Next →
Sign in / Sign up
ReadingResearchRadarInvestment framework

The practical implication for readers is: if you are evaluating the impact of AI-assisted programming on open-source infrastructure, Postgres provides a concrete stress-test case. It demonstrates that the true bottleneck for AI-generated code is not the ability to write syntactically correct patches, but the ability to pass through deep architectural review by a tiny team. This differs from the broad discussion of "whether AI can replace programmers"; it is a quantitative issue regarding review bandwidth and governance structures.