Alvaro Lopez Ortega / 2025-11-03 Briefing

Created Mon, 03 Nov 2025 22:14:53 +0000 Modified Tue, 04 Nov 2025 03:15:03 +0000
5544 Words

Cloudflare launches Oxy, a high-performance Rust-based proxy for next-gen traffic routing, supporting multiple protocols and extensive customization. Meanwhile, Balancer suffers a $128 million exploit, prompting chain halts and recovery plans. OpenAI signs a $38 billion cloud deal with AWS, expanding AI infrastructure amid rising investments and industry shifts.

▶️ Internet Infrastructure

Cloudflare Unveils Oxy: A Rust-Based Proxy for Next-Gen Traffic Routing

Cloudflare launched Oxy, a Rust-based proxy framework enabling programmable, high-performance traffic routing across protocols, supporting extensive customization and deployment at scale.

  • Cloudflare introduced Oxy, a Rust-based next-generation proxy framework, on March 2, 2023
  • Oxy supports multiple protocols, including HTTP/1/2/3, TCP, UDP, ICMP, with extensive traffic analysis and decapsulation capabilities
  • It is used as the foundation for projects like Zero Trust Gateway, iCloud Private Relay second hop proxy, and internal egress routing service

Balancer Loses $128M in Exploit; Berachain Halts Chain for Recovery

Balancer lost $128 million from a rounding error exploit in V2 pools, prompting Berachain to halt its chain and prepare a hard fork to recover funds.

  • Balancer suffered a $128 million exploit across Ethereum, Arbitrum, Base, and other chains, affecting V2 liquidity pools.
  • The attack was caused by a rounding error in Balancer V2 pools, exploited via multiple swaps in a single transaction, undervaluing Balancer Pool Tokens.
  • Balancer confirmed the issue is isolated to V2 pools; its token dropped over 11%, and the attacker appears to have ceased further withdrawals.
  • Berachain halted its blockchain and plans a hard fork to recover approximately $12.86 million, due to its codebase vulnerability similar to Balancer V2.
  • The chain halt is contentious, echoing past blockchain rollback debates, notably Ethereum’s 2016 DAO fork.

Android’s Hidden Flag Inflates Signal Strength Reports to Boost Coverage Perception

Android’s CarrierConfig contains an undocumented flag that allows operators to falsely report signal strength as higher, impacting user perception and network trust.

  • Android CarrierConfig includes a hidden flag (KEY_INFLATE_SIGNAL_STRENGTH_BOOL) to artificially report signal strength as one bar higher
  • The flag is undocumented in Android official docs but present in source code, enabling operators like AT&T and Verizon to activate it
  • This manipulation can inflate perceived network coverage, undermining trust and complicating network performance assessments

Microsoft invests $9.7B in AI capacity with IREN and Dell hardware

Microsoft commits $9.7 billion over five years to acquire AI computing capacity from IREN, including Nvidia systems and $5.8 billion in GPU hardware from Dell Technologies.

  • Microsoft signed a $9.7 billion five-year deal to purchase AI computing capacity from IREN Ltd.
  • The agreement grants Microsoft access to Nvidia systems in Texas optimized for AI workloads.
  • IREN will buy $5.8 billion worth of advanced graphics processing units and related equipment from Dell Technologies.

Aurora’s Autonomous Trucks Reach 100,000 Miles and Expand to New Routes

Aurora Innovation’s autonomous trucks have driven 100,000 miles safely; expansion to new routes and advanced lidar hardware confirm vehicle autonomy has arrived.

  • Aurora’s driverless trucks have completed 100,000 miles with perfect safety and on-time records
  • Expansion to a 600-mile Fort Worth-El Paso route, reducing transit time from 9-10 hours to autonomous operation
  • Aurora’s next-generation hardware features proprietary FirstLight lidar capable of detecting objects up to one kilometer away

Nvidia’s $5 Trillion Valuation Fueled by AI Data Center GPU Dominance

Nvidia’s valuation at $5 trillion is driven by its dominant role in AI data center GPU spending, which accounts for 39% of $35 billion per gigawatt, fueling the AI infrastructure boom.

  • AI data centers are now measured in gigawatts of power, with 1 GW roughly equivalent to a nuclear reactor output.
  • Bernstein Research estimates each 1 GW AI data center costs approximately $35 billion, encompassing semiconductors, networking, power, and construction.
  • Nvidia dominates AI data center spending, capturing nearly 30% profit margin from 39% of total costs allocated to GPUs, primarily from GB200 and Rubin chips.

Ghostty and Kitty Lead in Unicode Support Among Terminal Emulators

Terminal emulators’ Unicode support varies; Ghostty and Kitty lead, with Ghostty developed in Zig and supporting advanced Unicode features, while performance issues persist in slow, complex terminals.

  • The ucs-detect tool and blessed library have been extended to detect support for DEC Private Modes, sixel graphics, pixel size, and software version.
  • Ghostty, developed in Zig by Mitchell Hashimoto, scored highest in Unicode support tests and supports correct grapheme cluster handling, including Variation Selector 15.
  • Kitty, by Kovid Goyal, scored similarly high, with a text-splitting algorithm aligned with Python wcwidth standards, and is the only terminal supporting Variation Selector 15.

OpenAI Signs $38 Billion Cloud Deal with Amazon to Boost AI Development

OpenAI has secured a $38 billion, seven-year cloud services deal with Amazon to support AI development, after renegotiating its Microsoft partnership and expanding its cloud infrastructure investments.

  • OpenAI signed a $38 billion, seven-year cloud computing services agreement with Amazon.
  • The deal expands OpenAI’s access to cloud infrastructure for building and deploying AI technologies, including ChatGPT.
  • This follows OpenAI’s previous contracts with Microsoft, which were renegotiated in October 2025 to allow multi-cloud purchasing; OpenAI also collaborates with Oracle, SoftBank, and others for data center development.

Azure Outage Sparks Calls for Cloud Diversification After Systemic Risk Exposure

Azure’s October 29, 2025, outage, caused by a tenant configuration error in Azure Front Door, highlights the systemic risk of cloud dependency, prompting calls for diversification and sovereign cloud strategies.

  • Microsoft Azure experienced a hours-long outage on October 29, 2025, caused by an inadvertent tenant configuration change within Azure Front Door (AFD)
  • The incident impacted services and customer applications, affecting UK retailers Asda and Marks & Spencer, Dutch Railways, and their online systems
  • Microsoft confirmed the outage began at 15:45 UTC and was mitigated by midnight, with disruption lasting hours despite minimized timing impact

IBM Quantum Outage Affects 156-Qubit System Amid Ongoing Restorations

IBM experienced a quantum computer outage affecting the “ibm_aachen” system with 156 qubits, while maintaining over 15 quantum machines with 100+ qubits each, citing ongoing efforts to restore service.

  • IBM reported a temporary outage of its quantum computer “ibm_aachen” within the Qiskit Runtime service on October 30, 2025
  • The outage affected the Aachen system with 156 qubits of the Heron r2 processor; IBM is actively working to restore service
  • IBM’s fleet includes over 15 quantum computers, each with 100+ qubits, claiming industry-leading uptime and reliability

Google and Meta’s AI Innovations Drive Record Q3 2025 Revenue

Google and Meta’s AI-driven innovations boosted Q3 2025 revenue, with Alphabet reaching $102.34 billion and Meta $51.25 billion, as both plan increased AI infrastructure investment.

  • Google’s parent company Alphabet reported Q3 2025 revenue of $102.34 billion, a 16% year-over-year increase, with advertising revenue rising 12% to $74.2 billion.
  • Google Search’s AI-powered update enhances query results as prose, driving incremental query growth and revenue; Google Cloud revenue grew 33% YoY to $15.15 billion.
  • Meta’s Q3 revenue was $51.25 billion, up 26% YoY, with AI-powered recommendation systems increasing user engagement and the company’s ad revenue run-rate surpassing $60 billion; Zuckerberg announced increased infrastructure spending to develop “superintelligence.”

Amazon Q3 2025 Earnings Rise with Strong AWS Growth and Heavy AI Investment

Amazon Q3 2025 earnings show continued revenue growth, especially in AWS, with $33 billion in sales, $14.8 billion free cash flow, and $125 billion capex, amid heavy AI infrastructure investment.

  • Amazon Q3 2025 earnings reported with a 20% revenue increase in AWS, totaling $33 billion, and stock up over 13% after hours
  • AWS cloud revenue growth outpaced competitors, with a 20% increase; Google Cloud reported $15.2 billion (34% YoY), Microsoft Intelligent Cloud $30.9 billion (28% YoY)
  • Amazon’s free cash flow declined to $14.8 billion from $47.8 billion in Q3 2024, driven by $50.9 billion in property and equipment purchases; capex for 2025 projected at $125 billion, up from earlier estimates of $100 billion
  • Amazon announced a 14,000-job cut, unrelated to AI, aimed at cultural normalization; job cuts cost $1.8 billion in severance
  • Amazon invested heavily in AI infrastructure, with $34.2 billion in capital expenditures in Q3, and plans for higher 2026 capex; added over 3.8 GW of power capacity in the past year to support AWS AI workloads
  • Amazon’s AI investments include a $9.5 billion pre-tax gain from Anthropic, which it initially invested $8 billion in 2023; launched Project Rainer with 500,000 Trainium2 chips for Anthropic’s Claude models

Alaska Airlines Hires Accenture After Major IT Outage Despite Increased Spending

Alaska Airlines called in Accenture to investigate a late October IT meltdown that grounded flights for two days, despite 80% increased IT spending and redundant datacenters, highlighting ongoing cloud and infrastructure vulnerabilities.

  • Alaska Airlines experienced a major IT outage from October 23 to October 25, grounding flights and affecting nearly 50,000 customers
  • The airline increased IT investments by nearly 80% since 2019, including redundant datacenters and cloud migration, though the primary datacenter failure caused the outage
  • Accenture has been hired to audit Alaska Airlines’ IT infrastructure and improve system resilience, amid ongoing cloud dependencies on Microsoft Azure which suffered a recent outage

Europe Builds Open Source Cybersecurity Solutions to Protect Power Grids from Rising Threats

Europe is developing open source and standardized cybersecurity solutions, like SOARCA, to improve power grid resilience against cyberattacks and cascading failures, amid rising threats and outdated infrastructure.

  • The 2025 Spain power outage affected Spain, Portugal, and southwestern France, lasting 10 hours for power restoration, deemed Europe’s most severe blackout in two decades
  • The outage was caused by cascading failures, not cyberattack, involving simultaneous disconnections and overvoltages; human error likely involved
  • The 2015 Ukraine cyberattack on the power grid was the first attributed to online attackers, linked to Russian actors; cyber threats against critical infrastructure are increasing globally
  • Power grid IT infrastructure is complex, with outdated hardware, insecure protocols like DNP3, and multiple operating systems including Windows XP, NT4, BeOS, increasing vulnerability
  • The European Commission funds projects like the eFort framework and SOARCA, an open source security orchestration platform for power plants, demoed in Ukraine
  • SOARCA integrates layered security systems to detect anomalies, isolate incidents, prevent lateral movement, and coordinate response across substation, control room, and enterprise layers
  • Experts highlight the need for a unified European incident response standard, improved regulation, and shared communication protocols to enhance crisis management
  • Vendors like Claroty assist in inventorying and monitoring power plant cybersecurity, advocating for renegotiated contracts for oversight; regulators are pushing for baseline cybersecurity standards
  • The [Network Code on Cybersecurity (NCCS)] aims to enforce risk assessments across Europe; standards like CACAO Playbooks facilitate threat detection and response automation
  • Industry leaders emphasize the importance of standardization, information sharing, and proactive threat intelligence to defend critical infrastructure against increasing cyber and physical threats

xFusion Unveils Ultra-Efficient Cooling and Power Solutions for AI Data Centers

xFusion promotes a hardware-centric approach to mitigate thermodynamic and energy challenges in AI datacenters, delivering ultra-efficient cooling, power, and connectivity solutions.

  • xFusion advocates a holistic hardware strategy addressing thermal management, power consumption, and data connectivity in datacenters
  • Introduces “Black Technology” with proprietary materials, microchannel cold plates, and submersible liquid-cooled power supplies
  • FusionPoD liquid-cooled server cabinet delivers 1500W cooling capacity, achieves pPUE <1.06, below global average of 1.56, supporting high-power GPUs
  • High-speed interconnects based on PCIe 5.0/6.0 with proprietary PCBs and open standards, offering 10% higher IO density and double power in smaller form factor
  • 3KW Titanium PSU with 96.2% efficiency and modular liquid-cooled hot-swappable units support up to 36 units, reducing energy use
  • Demonstrated in harsh environments, e.g., ENAGEO’s data center in Sahara at 55°C, reducing downtime by two-thirds
  • Focuses on sustainable AI infrastructure amid rising energy demands, environmental challenges, and geopolitical considerations

Cisco Unveils Compact “Unified Edge” Chassis for AI and Security at the Edge

Cisco’s “Unified Edge” offers a compact, integrated edge infrastructure chassis with shrunken servers, security, and management tools, targeting multi-location deployments with AI and security workloads.

  • Cisco introduced the “Unified Edge,” a 3U chassis (UCS XE9305) with five bays, supporting shrunken servers, a Xeon 6 processor, and Catalyst 8200 router with firewall, connected via 25G backplane
  • The chassis is 18 inches deep, designed for small edge locations, with rear fans operable against walls and optional vertical standing orientation
  • Managed through Cisco’s Intersight SaaS platform, supporting remote maintenance and hot-swappable servers with up to four NVMe drives and GPU options for AI workloads; product ships December 2025

Windows 11 Fixes Long-Standing Shutdown Bug in October 2025 Update

Windows 11 25H2 update resolves a decades-old bug causing “Update and shut down” to restart PCs, with Microsoft confirming a fix in the October 2025 optional update and a broader fix coming in November.

  • Windows 11 25H2 Build 26200.7019 (or 26100.7019 on 24H2) fixes the “Update and shut down” bug preventing proper shutdown after updates.
  • The issue caused PCs to restart instead of powering off when selecting “Update and shut down,” affecting Windows 10 and 11.
  • Microsoft confirmed the October 2025 optional update (KB5067036) addressed the underlying cause, with a fix expected in the November 11 Patch Tuesday.

OpenAI Signs $38 Billion Cloud Deal with AWS to Power AI Growth

OpenAI signed a $38 billion, seven-year cloud computing deal with AWS, enabling access to Nvidia-powered data centers to support its AI growth amid competition from Microsoft and Google.

  • OpenAI agreed to pay $38 billion to Amazon Web Services (AWS) for multiyear cloud computing capacity, marking its first partnership with AWS.
  • The deal aims to meet OpenAI’s increasing computational demands, with capacity expected to be fully available by the end of 2026, utilizing Nvidia chips in Amazon’s data centers.
  • The seven-year agreement allows OpenAI to train AI models, process ChatGPT queries, and use Amazon’s CPUs for autonomous agentic AI tasks; the deal is smaller than OpenAI’s $300 billion Oracle and $250 billion Microsoft commitments.

▶️ Open Source

Python PEP 810 Aims to Clarify Lazy Imports and Prevent Eager Loading

Discussions on PEP 810 focus on preventing unintended eager imports, verifying lazy module states, and implementing runtime checks, with suggestions for sys.modules assertions and importlib functions.

  • The discussion centers on Python Enhancement Proposal (PEP) 810, which introduces explicit lazy imports in Python.
  • Key concerns include preventing accidental eager imports that undermine lazy import mechanisms and verifying lazy module states at runtime.
  • Proposed solutions include asserting module laziness with sys.modules checks and potentially adding functions in importlib to determine if modules are eagerly imported.

htmx 4.0 to Overhaul AJAX with Fetch and New Features in 2026

htmx 4.0, planned for early 2026, will overhaul internals by adopting fetch(), introduce explicit attribute inheritance, replace DOM snapshots with network requests for history, and add streaming, morphing, <partial>, view transition queuing, and improved extension and scripting support.

  • htmx 4.0 will replace XMLHttpRequest with fetch() as the core AJAX infrastructure, changing the event model accordingly
  • Attribute inheritance will be explicit via the :inherited modifier, ending implicit inheritance
  • History support will no longer snapshot DOM; instead, it will issue network requests for content restoration, with an opt-in extension for previous behavior
  • Major internal rebuild based on lessons from fixi.js and five+ years of support
  • New features include streaming responses and Server Sent Events (SSE) support in core, DOM morphing with idiomorph, <partial> tag support, improved view transitions with queuing, stabilized event ordering with standardized naming, enhanced extension API, and refined hx-on syntax

Maxima Now Runs in Browser Using WebAssembly for Math Visualizations

Marius Gerbershagen used ECL to compile Maxima to WebAssembly, enabling full browser-based functionality with TeX formulas and gnuplot graphics.

  • Marius Gerbershagen used ECL to compile Maxima to WebAssembly (WASM) for browser execution
  • Implementation includes TeX formula display and gnuplot graphics via WASM
  • Demonstration available at maxima-on-wasm.pages.dev

VimGraph: Visualizing Vim Navigation Movements with Custom Patterns

VimGraph generates a graph of Vim-style text navigation movements, supporting customizable key patterns and shortest path analysis, requiring Wolfram Language 13.0+ and contributed by Pavel Hajek.

  • VimGraph constructs a graph of text movements based on Vim editor’s modal navigation keys
  • Supports movements: h/l (left/right), j/k (up/down), w/b (word start), e (word end), ^/$ (line start/end)
  • Allows options “IncludedVimMovements”, “StringPattern”, “CustomPatterns” for customization; integrates with Graph options

System76 to Launch Pop!_OS 24.04 LTS and COSMIC 1.0 with Sync Features in December 2025

System76 will release Pop!_OS 24.04 LTS and COSMIC v 1.0 on December 11, 2025, featuring a modular desktop environment with sync capabilities, but delaying HDR and advanced workspace features.

  • System76 announced the December 11, 2025, release date for Pop!_OS 24.04 LTS and COSMIC v 1.0 at the Ubuntu Summit
  • COSMIC is a customizable desktop environment built from composable components, supporting branding, with options for minimal or Windows/macOS-like layouts
  • Version 1.0 will include COSMIC Sync, enabling end-to-end encrypted synchronization of system settings, data files, and applications across multiple devices; HDR support and advanced workspace features are delayed

Google Removes Gemma AI Model Over Hallucination and Defamation Concerns

Google withdrew Gemma from AI Studio after allegations of hallucinated false criminal claims about a US senator, emphasizing hallucination challenges in open LLMs and restricting public use.

  • Google removed Gemma from AI Studio following a lawsuit and complaints from US Senator Marsha Blackburn and conservative activist Robby Starbuck over hallucinated false criminal allegations.
  • Gemma, a family of open large language models (LLMs) released in 2024, is designed for research, not general-purpose chat, and remains accessible via API.
  • The removal was prompted by the model’s fabrication of defamatory claims when prompted with sensitive questions, highlighting hallucination issues in smaller open models.

OpenAI Extends Cloud Reach with $38B AWS Deal and Nvidia GPU Clusters

OpenAI’s $38 billion, seven-year partnership with AWS broadens its cloud infrastructure, involving Nvidia GPU clusters and scaling to tens of millions of CPUs, complementing its Azure commitments.

  • OpenAI signed a seven-year, $38 billion agreement with AWS on November 3, 2025, expanding beyond Microsoft Azure for AI compute resources
  • The deal provides immediate access to AWS EC2 UltraServers, scaling to tens of millions of CPUs over seven years for agentic workloads
  • AWS is deploying custom infrastructure with Nvidia Blackwell GB200 and GB300 GPUs clustered on EC2 UltraServers, supporting AI efficiency and performance

▶️ Management and Leadership

NYC Mack Super Pumper: The Firefighting Beast of the 1960s and 70s

The NYC Mack Super Pumper, powered by a 2,400hp Napier-Deltic engine, could pump over 8,800 GPM at 350 psi, responding to over 2,200 calls from 1965-1980s.

  • The Mack Super Pumper system was built in response to the 1963 Staten Island fire, costing $875,000 and operated from 1965 to early 1980s.
  • It comprised five trucks: a locomotive-engined central pump capable of 10,000 GPM at low pressure and 8,800 GPM at 350 psi, a tender truck, and three satellite units.
  • Powered by a 2,400hp Napier-Deltic two-stroke diesel engine with three crankshafts, capable of consuming 137 gallons of diesel per hour, and producing over 10,000 gallons per minute flow.

Skald Migrates Backend from Python to Node.js for 3x Throughput Boost

Skald migrated its backend from Python (Django) to Node.js (Express + MikroORM) in three days to improve scalability, achieving ~3x throughput and better concurrency, due to Python async limitations.

  • Migration from Python to Node.js completed within three days, one week after launch
  • Benchmarks show ~3x increase in throughput with Node.js, enabling more concurrent processing
  • Migration involved switching from Django to Express + MikroORM, replacing Python’s async ecosystem and ORM

Innovative Fish Feeds to Protect Marine Ecosystems and Secure Seafood Supply

Innovative, sustainable fish feed alternatives—such as plant-based blends, algae oils, insects, and microbes—aim to reduce forage fish dependence, protect ecosystems, and ensure global seafood security.

  • The anchoveta population collapsed in 2016 and 2023, due to overfishing and ocean temperature spikes, impacting seabirds, marine mammals, and fishing livelihoods
  • Ninety percent of forage fish caught are ground into fishmeal and fish oil for farmed seafood, threatening marine ecosystems and food security
  • The F3 Challenge incentivizes development of plant-based, algae-derived, insect, and microbial feeds to replace wild forage fish in aquaculture

How Identity and Emotion Drive Costly Programming Language Biases

Decisions on programming languages are driven by identity and emotion, not logic, causing costly biases; reframing choices as economic decisions can improve long-term outcomes.

  • The article analyzes how identity, emotion, and ego influence programming language decisions, often leading to multi-million dollar technical debt.
  • A case study from Takkle illustrates a $300K/month rebuild delay caused by switching from PHP to Perl based on leadership bias.
  • Scientific research shows that challenges to core identity activate threat responses in the brain, hindering objective evaluation of technical choices.

Google Suspends SSLMate Cloud Account Causing Service Disruptions

Google suspended SSLMate’s Google Cloud account three times between 2024-2025, disrupting customer integrations; current solutions involve insecure long-lived keys or complex OIDC configurations.

  • Google suspended SSLMate’s Google Cloud account three times in 2024 and 2025 without notification, causing service disruptions.
  • The suspensions involved account disablement, service restrictions, and a violation notice for Terms of Service, with no clear reasons provided.
  • SSLMate’s current workaround involves impersonating service accounts with long-lived keys or using complex OIDC setup, which is less secure and overly complicated.

Amazon lays off 14,000 employees with early texts to prevent badge-in issues

Amazon laid off 14,000 employees on October 28, 2025, with early-morning texts sent to prevent badge-in issues, citing AI-driven rapid changes despite strong company performance.

  • Amazon laid off 14,000 employees on October 28, 2025, as part of a strategic effort to streamline operations and accelerate innovation.
  • Some affected employees received early-morning text messages to inform them of job elimination, preventing badge-in issues at office entry.
  • The texts instructed employees to check email or call a help desk if no email was received, aiming to avoid employees arriving at work unable to badge in.

Palantir CEO Highlights Disciplined Hiring Amid Industry Layoffs

Palantir reports earnings surpassing estimates; CEO Alex Karp highlights disciplined hiring to sustain software development amid industry layoffs and economic uncertainty.

  • Palantir CEO Alex Karp emphasizes “disciplined” hiring practices amid widespread tech layoffs
  • Company maintained lean headcount despite significant growth, avoiding casual expansion
  • Other tech giants like Microsoft, Amazon, and Target have laid off thousands of employees to streamline operations

Michelle Lim Secures $5M Seed Funding for AI Startup Flint

Michelle Lim, founder and CEO of Flint, raised $5 million seed funding leveraging her startup experience at Warp, mentorship from Ali Partovi, and investor trust built through her track record.

  • Michelle Lim, 28, raised nearly $5 million in seed funding for her AI startup Flint from investors including Accel, Neo, and Sheryl Sandberg’s family office.
  • She previously worked at Warp, a coding terminal startup, where she gained experience in engineering, marketing, product, and sales, presenting to investors like Dylan Field.
  • Lim credits her early startup experience and mentorship from Ali Partovi at Neo as critical to her fundraising success and entrepreneurial confidence.

Amazon Starts Global Blocking of Pirate Fire TV Stick Apps to Fight Streaming Piracy

Amazon begins global blocking of unauthorized Fire TV Stick apps from today, targeting piracy by disabling side-loaded apps that provide illegal access to premium content, supporting ACE efforts.

  • Amazon will block ‘dodgy’ Fire TV Stick apps globally starting from today to combat piracy
  • The crackdown targets side-loaded unofficial apps providing access to pirated content, including premium movies and live sports
  • The initiative is led by Amazon in partnership with the Alliance for Creativity and Entertainment (ACE), aiming to prevent illegal streaming and protect users from malware

ServiceNow Raises Full-Year Guidance Amid Strong Q3 Revenue Growth

ServiceNow raised full-year guidance after $3.4 billion Q3 revenue (+22% YoY), with strong subscription growth, but warned federal shutdown may impact future government contracts and demand.

  • ServiceNow reported $3.4 billion revenue in Q3 2025, up 22% YoY, with a 33.5% operating margin, exceeding expectations
  • Subscription revenue outperformed analyst forecasts; guidance for full-year revenue increased
  • CFO Gina Mastantuono cautioned that US government shutdown effects are not yet reflected in results, despite strong demand from federal agencies

Canonical Aims to Make Core Desktop Default in Ubuntu Within a Decade

Canonical plans to make Core Desktop the default Ubuntu experience within 5-10 years, focusing on modular architecture, snapd permissions prompts, and hardware support, while progress on immutable distros remains slow.

  • Canonical’s VP of Engineering, Jon Seager, predicts Ubuntu’s default will shift to a Core Desktop image within 5-10 years, with current plans for a non-immutable, modular architecture.
  • Core Desktop aims to replace traditional Ubuntu Desktop, with ongoing development of snapd permissions prompting, systemd integration, and support for hardware like cameras and speakers via Pipewire as a Snap in 26.04.
  • Progress on immutable distros like NixOS and OStree is slow; Canonical favors a flexible architecture that allows swapping components like gdm and desktop environments, emphasizing security and long-term support.

AWS, Nvidia, CrowdStrike Launch Third Cybersecurity Startup Accelerator

AWS, Nvidia, and CrowdStrike launched the third annual Cybersecurity Startup Accelerator, offering global applicants access to resources and investor exposure, focusing on AI, cloud, and data security innovations.

  • AWS, Nvidia, and CrowdStrike are running a cybersecurity startup accelerator open for applications until November 15, 2025
  • The program targets early-stage startups developing cloud, application, identity, agentic security, and data security technologies
  • Participants gain access to cloud, compute, threat intelligence resources, and opportunities to pitch at AWS Demo Day on March 24, 2026; 59 startups have graduated, raising over $730 million collectively

Debian to Mandate Rust Dependencies in APT from 2026, Dropping Legacy Ports

Debian will enforce a “hard requirement” for Rust dependencies in APT from May 2026, impacting legacy ports like Alpha, PA-RISC, m68k, and SH4, which lack official Rust support.

  • Debian’s APT package manager will require Rust or rust in peace starting May 2026, extending to the Rust compiler, standard library, and Sequoia ecosystem
  • Debian supports seven official architectures and 11 maintained ports; Rust officially supports three Tier 1 architectures (Arm64, i686, x86-64) and Tier 2 for Loongson and SPARC64
  • Architectures like Alpha, PA-RISC, m68k, and SH4 will face removal due to lack of Rust support, with potential sunset within the next Debian release cycle

IEEE Survey Shows Decline in AI Software Development Skills Amid Rapid AI Innovation

IEEE survey indicates declining demand for AI software development skills (down 8 points to 32%) as agentic AI advances, with increased focus on AI ethics, data analysis, and machine learning in 2026.

  • IEEE survey of 400 CIOs, CTOs, and IT directors across six countries finds an 8 percentage point decline in demand for software development skills in AI roles, dropping to 32% in 2026.
  • Demand for AI ethics expertise increased by 9 points to 44%, data analysis skills rose by 4 points to 38%, and machine learning capabilities grew by 6 points.
  • Nearly all professionals expect AI innovation to accelerate rapidly in 2026, with industries like software (52%), banking (42%), healthcare (37%), and automotive (32%) expecting significant transformation.

Rockstar Fires Over 30 Employees Amid Union Busting Allegations

Rockstar Games fired over 30 employees, including union members on Discord, citing gross misconduct, amid accusations of union busting; the company previously mandated in-office work and faced criticism for crunch culture.

  • Rockstar Games laid off over 30 coders and graphic designers, accused by IWGB of “blatant and ruthless union busting.”
  • The dismissed staff were members of a private trade union Discord channel discussing workplace improvements; all accounts in that channel were terminated after management became aware.
  • Take-Two Interactive cited “gross misconduct” as the reason for termination, supporting Rockstar’s actions; UK law requires at least 10% union membership for statutory recognition, possibly just met.

Palantir Reports Record $1.2B Revenue Amid US Growth and CEO Praise

Palantir achieved $1.2 billion Q3 2025 revenue (+63% YoY), with $476 million profit, driven by US government growth; CEO Karp praised growth and criticized skeptics.

  • Palantir reported $1.2 billion in Q3 2025 revenue, a 63% year-over-year increase, with $476 million profit.
  • US government business grew 52% YoY to $486 million; company beat earnings estimates and raised FY2025 revenue guidance to 53% YoY.
  • CEO Alex Karp criticized critics, highlighted record growth, and emphasized Palantir’s role in enabling retail investors’ high returns through substantive growth.

▶️ Technology

Apple Launches Web App Store with Redesigned “Today” Tab and Media Features

Apple introduced a comprehensive web version of the App Store with a full “Today” tab, media-rich app pages, and platform-specific sections, enhancing browsing and discovery capabilities.

  • Apple launched a new web interface for the App Store on November 3, 2025, accessible via apps.apple.com/us/iphone/today
  • The redesign includes a full replica of the “Today” tab, top charts, editorial content, and dedicated pages for iPhone, iPad, Mac, Vision, Watch, and TV libraries
  • App product pages now feature media-rich interfaces with new iconography for categories, awards, and events; a new search interface allows quick app discovery

New Insights into LLM Security: Limiting Agents and Overcoming Defenses

Meta’s “Agents Rule of Two” advises restricting AI agents to two of three properties to improve security, while “The Attacker Moves Second” shows adaptive attacks defeat most prompt injection defenses, highlighting ongoing vulnerabilities.

  • Two new papers on LLM security: “Agents Rule of Two” (Meta AI, Oct 31, 2025) and “The Attacker Moves Second” (arXiv, Oct 10, 2025)
  • “Agents Rule of Two” proposes limiting agents to two of three properties—processing untrustworthy inputs, access to sensitive data, and changing state—to mitigate prompt injection risks
  • “The Attacker Moves Second” demonstrates that adaptive attacks bypass 12 defenses with over 90% success, with human red-teaming achieving 100% success

Skyfall-GS: Real-Time 3D Urban Scene Generation from Satellite Images

Skyfall-GS creates immersive 3D urban scenes from satellite images via a two-stage process, utilizing diffusion models for high-quality textures and iterative refinement for geometric accuracy, enabling real-time exploration.

  • Skyfall-GS synthesizes 3D urban scenes from satellite imagery using diffusion models, enabling real-time, explorable city-block scale environments without costly 3D annotations
  • The framework employs a two-stage process: initial 3D reconstruction with pseudo-camera depth supervision and appearance modeling, followed by iterative dataset refinement using prompt-to-prompt editing with a pre-trained T2I diffusion model
  • Extensive experiments show improved cross-view geometry consistency and more realistic textures compared to state-of-the-art methods; datasets and code are available here

YASA Unveils 750 kW Axial Flux Motor Powering High-Performance EVs

YASA’s 28-pound axial flux electric motor delivers 750 kW (1,005 hp), outperforming previous models by 40%, with scalable design for high-performance EVs and existing applications in luxury sports cars.

  • UK-based YASA developed a 28-pound axial flux electric motor producing over 1,000 horsepower (750 kW), equivalent to 1,005 hp
  • The motor exceeds previous YASA record (550 kW, 737 hp) by 40%, with continuous power output between 350-400 kW (469–536 hp)
  • The motor’s design uses no exotic materials, enabling scalability; already supplies motors for high-performance vehicles like Mercedes-AMG GT XX and Ferrari 296 GTB

US Expands AI Hardware Exports with Nvidia Chips to UAE

The US authorized Microsoft to export Nvidia AI chips to UAE for the first time, expanding US AI hardware export licenses and enabling UAE access to advanced AI computing technology.

  • US permits Microsoft to ship Nvidia AI chips to UAE for the first time
  • This marks a significant expansion of US export licenses for AI hardware
  • The license was granted by the US Bureau of Industry and Security, enabling Nvidia to supply AI chips to UAE-based customers

Do Large Language Models Truly Understand or Just Recognize Patterns?

The article explores whether AI models like ChatGPT genuinely understand or simulate thinking, highlighting their brain-like properties, limitations, and implications for understanding human intelligence.

  • The article examines whether large language models (LLMs) like ChatGPT truly understand or are merely sophisticated pattern recognizers.
  • It discusses the evolution of AI from neural network simulations in the 1980s to current models trained on internet data via next-token prediction.
  • Neuroscientists and cognitive scientists see parallels between LLMs and human brain functions, suggesting models may radically demystify thinking and cognition.

Pat Gelsinger Launches Gloo to Infuse AI with Christian Values and Spiritual Impact

Pat Gelsinger launched Gloo, a faith-based AI platform, to infuse LLMs with Christian values, aiming to influence spiritual health and reform, amid debates on AI’s truthfulness and theological integrity.

  • Pat Gelsinger, former Intel CEO, founded Gloo, a platform connecting the faith ecosystem, emphasizing Christianity and training Large Language Models (LLMs) infused with Jesus.
  • Gloo aims to QA AI models to monitor their impact on users’ spiritual health and ensure acceptability across Christian denominations, avoiding theology and politics.
  • Gelsinger compares AI to Gutenberg’s printing press, highlighting its potential for reform and dissemination of religious messages, while acknowledging challenges with LLMs’ truthfulness and theological implications.

Microsoft invests $7.9B in UAE AI infrastructure as Alphabet and Meta raise billions for AI growth

Microsoft plans to spend over $7.9 billion on UAE AI infrastructure (2026–2029), including GPU shipments equivalent to 60,400 A100 chips, while Alphabet and Meta raise billions via bonds for AI expansion.

  • Microsoft will invest over $7.9 billion in AI infrastructure in the UAE from 2026 to 2029, including $5.5 billion in capital expenses and $2.4 billion in operating costs
  • Microsoft secured export licenses in September to ship GPUs equivalent to 60,400 A100 chips, including GB300 GPUs
  • Alphabet is raising at least €3 billion ($3.5 billion) via euro-denominated bonds and up to $15 billion in US bonds to fund AI expansion; Meta also issued $30 billion in bonds for AI infrastructure

AI Boom 2025 Mirrors 1995 Internet Rise in Infrastructure and Impact

The 2025 AI boom mirrors the 1995 internet rise, with infrastructure growth, industry-specific employment effects, and potential bubble risks, driven by unmet demand and automation.

  • The article compares the current AI era to the 1995 dial-up internet phase, highlighting parallels in infrastructure buildup and hype cycles.
  • Employment impacts vary by industry; radiology has seen increased demand despite AI advancements due to Jevons Paradox, while industries like journalism have shifted from employment growth to content creator roles.
  • Long-term infrastructure investments by hyperscalers (Microsoft, Google, Meta, Amazon) totaling nearly $500 billion are foundational for future AI development, despite current bubble concerns.