Search the archive

DEPUTY

Deputy aims to enable dependency management at scale by providing core dependency management primitives along with a unified toolchain

RISKY BULLETIN: SLOVAKIA FINDS RUSSIAN BACKDOOR IN TRAFFIC SPEED CAMERAS

Slovakia's national security service (NBU) discovered that recently installed traffic cameras contain a backdoor granting shell and network access via SMS commands from hardcoded Russian phone numbers. The compromised NERO R-ONE cameras, allegedly purchased through a Cyprus shell company, were identified as rebadged units manufactured by a St. Petersburg-based firm. In addition to the SMS backdoor, the NBU technical report revealed that the devices shipped with SecureBoot disabled and exposed live video streams without authentication.

OWASP FLAGS TOP AI SKILL RISKS IN NEW SECURITY BLUEPRINT

OWASP released its Agentic Skills Top 10 and Universal Agentic Skill Format v1.0. Malicious skills and supply-chain compromise rank first and second. A Paperclip impersonation campaign used Trojanized packages and skills. The skills exceeded 300,000 installs, and it shows that the skill format should record provenance, permissions, dependencies, hashes, and changes. Security teams should inventory agent skills and quickly disable flagged ones.

WHY AI LEARNED TO ATTACK BEFORE IT LEARNED TO DEFEND

Offensive AI is outpacing defensive AI due to a structural asymmetry: offensive exploits provide a clear verification "oracle" (the exploit either lands or fails), allowing reinforcement learning loops to train attackers much faster than defenders. The industry is actively experiencing this imbalance: Anthropic's Project Glasswing recently surfaced over 10,000 high- or critical-severity flaws, far outpacing the two-week average patch time, while OpenAI's Daybreak Red achieved a roughly 95% success rate on advanced offensive security tasks. This rapid offensive acceleration, illustrated by the Hugging Face breach in July and the rise of highly capable open-weight models, highlights a growing threat landscape where AI substitutes human effort on offense but only assists on defense.

PRIVATE EQUITY FIRM APOLLO CONFIRMS DATA BREACH AMID HACKING WAVE TARGETING FINANCIAL GIANTS

Apollo disclosed that attackers accessed its cloud environment from July 6 to 10 through social engineering.

TATA'S B2B PLATFORM RETURNED OTPS IN API RESPONSES

Tata nexarc's login API returned the OTP directly in its response body, letting researcher Eaton take over any account, including Tata Steel's own, just by knowing the target's phone number, until CERT-IN confirmed a fix on July 31.

OPENAI REGAINS GROUND IN THE CORPORATE AI MARKET

OpenAI appears to be regaining ground against Anthropic among US corporate customers, according to spending data from Ramp covering more than 70,000 organisations. Anthropic led in May and July, but third-quarter trends suggest faster growth for OpenAI, while increasing overall spending on paid AI services indicates that both providers may expand even as corporate loyalty remains fluid.

NVIDIA PUTS GROQ 3 LPX INFERENCE CHIP INTO PRODUCTION

Nvidia's Groq 3 LPX accelerator is entering production as part of its Vera Rubin platform, targeting ultra-fast token generation for agentic AI. Nebius is the first cloud provider committed to deploying it.

IBM BUILDS CHIP THAT RUNS ARM AND Z WORKLOADS ON THE SAME CORES

IBM is developing a dual-architecture processor that can natively execute both Arm and IBM Z instructions. The design could let enterprises run modern AI and cloud-native software directly alongside critical mainframe workloads.

WHEN NOT TO USE STREAM PROCESSING: MESSAGING AND APIS ARE OFTEN ENOUGH

Stream processing is valuable for continuous event-driven computation but is often the wrong architecture for request-response interactions, direct state queries and corrections, or work requiring locks. As evidence, Kestra rebuilt its messaging layer around plain queueing, and Confluent moved Control Center's metrics pipeline from Kafka Streams to Prometheus.

AUTONOMY AND INNOVATION

AI cybersecurity capabilities are inherently dual-use: the same ability to discover and exploit vulnerabilities is needed to defend infrastructure, with intent shaped by incentives and whoever controls the system. Defenders must automate the full loop from vulnerability detection through patching, deployment, and rollback.

See Zoom Phone plans →

ATLASSIAN CODE CONTEXT GIVES AI AGENTS EYES ACROSS YOUR ENTIRE CODEBASE

Atlassian Code Context is a Teamwork Graph capability that lets developers and AI coding agents search connected GitHub and Bitbucket repositories alongside Jira, Confluence, Loom, and other organizational data. Atlassian says internally benchmarked agents using this context produced 44% more accurate results with 48% fewer tokens, while access remains controlled by existing repository permissions.

ANTHROPIC GIVES ITS SLACK AGENT FULL CONVERSATION CONTEXT

Claude Tag can now follow entire Slack conversations and proactively join discussions instead of waiting to be explicitly invoked. Anthropic describes the approach as “multiplayer AI,” where agents participate directly in team workflows.

THOMSON REUTERS LAUNCHES ITS OWN AI MODEL FOR LEGAL WORK

Thomson is Thomson Reuters' first proprietary large language model. It is designed around professional legal workflows and the company's specialized data.

ENSURING RELIABLE OPENTELEMETRY INGESTION AT SCALE

ClickHouse describes how its LogHouse observability platform evolved to handle OpenTelemetry ingestion at 50 million events per second without losing telemetry during ClickHouse backpressure. After finding in-memory queues and local collector write-ahead logs inadequate, the company designed a priority-based failover pipeline that keeps ClickHouse in the hot path and uses blob storage as a durable overflow buffer.

CAN NEOCLOUDS CORNER AI COMPUTE?

AI-focused clouds such as CoreWeave, Lambda, and Nebius offer cheaper, highly optimized GPU infrastructure, but hyperscalers still hold advantages in security, managed services, and enterprise integration. Gartner expects neoclouds could capture 20% of the AI cloud market by 2030.

GOOGLE BRINGS ANTIGRAVITY UNDER GEMINI ENTERPRISE

Google is adding Antigravity to Gemini Enterprise with centralized budgets, quotas, and overage controls. The move gives IT teams more visibility over rapidly growing AI usage and spend.

BUILDING A CANONICAL MTTR TIMELINE ACROSS INCIDENT TOOLS

A design for building a source-neutral, continuously maintained MTTR timeline across PagerDuty, Jira, Slack, observability systems, and other incident tools.

CURSOR EXPLAINS HOW IT SCALED BEYOND GIT

Cursor uses S3 as its source of truth and local NVMe repositories for latency-sensitive operations to overcome Git scaling limitations.

Watch the on-demand webinar

DOUBLE-DIGIT ECOMMERCE GROWTH RETURNS

US ecommerce sales grew 12.2% YoY to $340.2B in Q2 2026, marking a second straight quarter of double-digit growth and outpacing overall retail growth of 6.7%. Ecommerce now accounts for 17.1% of retail sales. Growth varied sharply by category, with general merchandise up 21.6% while clothing rose just 3.8%. The acceleration suggests renewed ecommerce momentum, though inflation and Amazon's earlier Prime Day may have boosted Q2 figures.

IS THE ERA OF THE 15,000 PLUS MARKETING TECHNOLOGY LANDSCAPE COMING TO AN END?

Consolidating a martech stack is less about cutting applications and more about whether each tool provides unique value a broader platform cannot. AI is making many single-purpose tools redundant by combining functions that once required separate systems. However, over-consolidation risks vendor lock-in and losing specialized capabilities. The likely outcome is a smaller core of shared platforms, with specialized tools kept only when they add clear value.

HOW LONG SHOULD BRANDS TEST A CREATOR PARTNERSHIP? 28 CREATOR ECONOMY EXPERTS WEIGH IN

Brands should test creator partnerships broadly, then give creators 3-6 months or at least three activations to prove repeatability. A single post measures content performance, not the partnership, while short attribution windows can undercount conversions by 40-60%. Match evaluation to the goal and sales cycle, then shift budget toward creators showing brand fit and long-term potential.

YOUR BEST CUSTOMERS MAY NOT BE YOUR MOST ENGAGED

More engagement doesn't always mean more value. The right lifecycle strategy starts by defining what healthy usage looks like for each product. A dating app may succeed when users leave, while a utility product may only need monthly use. Customer replies can also reveal intent that clicks and opens miss. As automation makes it easier to build and launch more campaigns, teams need stronger QA and guardrails to manage the added complexity.

META'S AI-FIRST AD BUYING IS FORCING A NEW OPERATING MODEL FOR PAID SOCIAL TEAMS

Meta's AI-first ad buying is reducing the need for manual campaign optimization and shifting paid social teams toward managing the inputs that guide the system. Clean conversion signals become critical because they shape how Meta optimizes campaigns. Creative production also becomes a bigger performance lever as targeting gets automated. Teams will need to rethink workflows and governance as fewer campaign decisions remain under direct human control.

SAAS DATA MOATS: WHY PROPRIETARY DATA IS THE STRONGEST COMPETITIVE ADVANTAGE IN 2026

As AI makes software features easier to replicate, proprietary data is becoming a stronger competitive moat. The advantage comes from data that is unique and improves the product with each customer interaction. A data flywheel can make this advantage compound as more usage creates better insights. However, data only becomes a moat at sufficient scale and with strong governance. Otherwise, it can become an expensive liability rather than a defensible advantage.

GTM ENGINEER, APPLIED AI AT TLDR ($175-205K BASE + $40-60K BONUS, FULLY REMOTE)

TLDR is hiring a GTM Engineer to join our Applied AI team and own our AI-native GTM stack. We're looking for someone comfortable building AI agents and working with HubSpot. Click here to learn more [14]!

SERVED VS SEEN: WHICH BRAND ARE YOU?

Brands are optimizing for impressions and media efficiency while business growth declines. Being served an ad doesn't mean it was seen or remembered. Test creative organically on social platforms before investing in paid media, then put spend behind content that has already proven it can earn attention.

MARKETERS, THE FREE TOKEN LUNCH IS OVER. IS YOUR STACK READY?

AI costs are changing as models use more tokens for reasoning and companies deploy them across more tasks. The biggest model isn't always the most economical choice. Simple decisions can use cheaper models, while complex problems may justify higher-cost reasoning models. The key metric is shifting from token price to the cost of getting the right answer. AI stacks need to match each task with the right model, rather than defaulting to one model for everything.

“THANKS FOR YOUR QUICK RESPONSE”

Quick responses show thoughtfulness, helping brands stand out beyond price.

COOKIE REJECTION ≠ INVISIBLE USER

Enable Advanced Consent Mode and Google Signals in GA4 so that Google can model the behavior of users who declined cookies based on the patterns of those who accepted.

BEHIND JACK IN THE BOX'S PUSH INTO COLLECTIBLES

Collectible merch has become its own acquisition channel for quick-service brands.

GET THE GUIDE

AUTONOMY AND INNOVATION

Incumbents will adopt AI cautiously because they have more to lose from mistakes. Startups can embrace autonomy more aggressively, giving them a natural advantage in turning AI into disruption.

INSIDE PM AT STRIPE: WHAT COMES AFTER "BECOME A BUILDER?" (61 MINUTE VIDEO)

AI is pushing every product-building role upstream rather than simply eliminating functions. PMs will increasingly create leverage through shared context, judgment, strategy, persuasion, and domain expertise while AI and adjacent functions absorb more execution.

THE DEADLINE THAT WAS ALREADY MISSED

Bad roadmaps often fail from missing dependencies, not bad estimates. Working backwards exposes what must be true before execution begins.

A GOOD MOVE IS ENABLING

Good product moves create value now while opening better options for what comes next. Teams should avoid short-term wins that lead to strategic dead ends.

START YOUR NEXT FLOW IN FIGR →

PRODUCT OWNER INTERVIEW QUESTIONS FOR THE AGE OF AI

As AI lowers the cost of building software, product judgment matters more than tool fluency. Great Product Owners know what to build, what to stop, and where AI should assist rather than decide.

TO INVENT WAYMO, THEY HAD TO REINVENT PM (60 MINUTE VIDEO)

Waymo shows how PM must evolve for problems without clear metrics or playbooks. The key is defining what good looks like, setting concrete checkpoints, and building teams capable of learning their way forward.

BIG-BANG REDESIGNS VS. LEGACY MODERNIZATION

Big-bang legacy redesigns often fail because critical dependencies are easy to miss. Modernizing gradually, with users involved and workflows replaced in small slices, reduces risk and delivers value sooner.

THE ORACLE PROBLEM: AN INVISIBLE BOTTLENECK TO AI AND MEDICINE

Healthcare AI benchmarks can measure test performance without proving clinical effectiveness. Better evaluations must measure whether AI improves real-world health outcomes.

DEVELOPERS MUST BECOME MINI-PMS

AI is pushing engineers beyond code and toward greater product ownership.

ENTERPRISE AI'S BIGGEST OPPORTUNITIES

These opportunities lie between the model and the workflow.