Search the archive

SHADOW

Shadow is a meeting assistant that captures insights and turns decisions into actionable tasks to streamline follow-ups.

NOTIONAPPS 2.0

NotionApps 2.0 is a no-code platform for building web apps, portals, and internal tools on top of Notion data.

THE CHAINSMOKERS' MANTIS VC CLOSES $100M THIRD FUND

Mantis VC, co-founded by The Chainsmokers, raised $100 million for its third fund, 25% more than its previous $80 million fund. Despite its celebrity status, Mantis focuses on B2B startups, with investments in companies like Chainguard and Rogo. It leverages its extensive network from private shows for Fortune 500 companies for valuable customer introductions.

HOW THE SYSTEM OF RECORD DIES

Everyone thinks AI agents will kill systems of record - why log in Salesforce when agents do everything? David Peterson thinks that's backwards. Agents need constraints and permissions, making systems of record essential. However, the pricing power shifts. Companies will pay for work, not data storage. When incumbents throttle APIs to maintain revenue, customers revolt. The winners won't be today's systems desperately locking down data - they'll be whoever builds ground-up "systems of intelligence" without legacy baggage.

AI IS TOO EXPENSIVE FOR UNLIMITED SUBSCRIPTIONS

Anthropic got backlash for adding usage limits to Claude Max, following similar issues with Cursor and Windsurf. It's always coding tools that face pricing revolts. The problem is simple - coding requires flow state, not paywalls interrupting a 5-hour feature build. Running out of credits in an SEO tool works because it's discrete tasks. However, limits mid-debug ruin the IDE experience. These companies aren't being deceptive, they're stuck. AI costs too much for unlimited subscriptions that developers want. Like expensive mobile data in 2003, this will look ridiculous once cheaper models make unlimited usage possible.

DEFAULTS MATTER

Your product's default settings determine success because AI agents won't customize anything - they use whatever works out-of-the-box.

WHY SEED ROUNDS ARE GROWING AS STARTUPS SHRINK

Writing $1M checks doesn't move the needle for a $500M fund - that's why smaller seed rounds are vanishing.

YOUR MOST PROFITABLE FEATURES ARE THE ONES YOU FORGOT ABOUT

80% of features in the average software product are never used, yet founders keep shipping them anyway.

ENDGAME GEAR MOUSE CONFIG TOOL INFECTED USERS WITH MALWARE

German peripherals maker Endgame Gear is warning users that malware was embedded in its configuration tool for the OP1w 4k v2 mouse between June 26 and July 9. The malicious version of the download included the XRed backdoor, which features keylogging capabilities, the ability to open a remote shell, and data exfiltration. Endgame Gear has announced that, going forward, it will discontinue hosting separate download pages and will add SHA hash verification and digital signing for all hosted files.

LOVENSE SEX TOY APP FLAW LEAKS PRIVATE USER EMAIL ADDRESSES

A group of security researchers discovered two vulnerabilities in the Lovense sex toy app that could allow an attacker to retrieve a user's private email address with just their username and perform an account takeover with just an email address. To exploit the flaw, an attacker sends a POST request to the `/api/wear/genGtoken` API endpoint, then forwards the encrypted payload to another endpoint, which returns a fake ID that the user can add to their contact list to reveal the real email address. Lovense claims that both flaws were fixed. However, researchers report that only the account takeover flaw was fixed.

TELECOM GIANT ORANGE WARNS OF DISRUPTION AMID ONGOING CYBERATTACK

French telecom giant Orange reported a cyberattack on July 25 that disrupted services for business clients and some public sector services in France. The company isolated affected systems and expects full restoration by Wednesday. Orange claims no customer or internal data was compromised and has notified the relevant authorities about the incident.

WIZ RESEARCH UNCOVERS CRITICAL VULNERABILITY IN AI VIBE CODING PLATFORM BASE44 ALLOWING UNAUTHORIZED ACCESS TO PRIVATE APPLICATIONS

Wiz Research discovered a critical vulnerability in Base44, a popular "vibe coding" platform that allows users to build apps through natural language prompts. The flaw allowed attackers to bypass authentication using only non-secret app IDs, potentially granting access to private enterprise applications and sensitive data.

FAKE CAPTCHA LED TO LUMMA

Attackers used a fake CAPTCHA page that employed the ClickFix social engineering technique to trick users into pressing Windows+R and pasting a PowerShell command that downloaded LUMMA stealer malware. The malware targeted browser credential stores (Login Data and Web Data) from Microsoft Edge and Google Chrome, establishing C2 communication within minutes of initial execution. The attack chain leveraged multiple evasion techniques, including heavily obfuscated PowerShell, AutoIt scripts, and mshta.exe proxy execution to bypass security controls.

BUILDING A CLOUD SECURITY ROADMAP: TOOLS BY LAYER AND WHEN YOU NEED THEM

This post provides an overview of the layers in cloud environments, the threats that target each layer, and the tools that are best suited to address them. The cloud infrastructure/control plane layer handles tasks such as resource provisioning, configuration management, and user permissions. The orchestration layer hosts platforms such as Kubernetes and related services live, whereas the platform layer runs more generalized workloads. Finally, the application layer includes software that end-users rely on, such as custom web applications, APIs, and SaaS apps.

Get the Toolkit

FIRECRAWL

Firecrawl is an API service that takes a URL, crawls it, and converts it into clean Markdown or structured data, which can then be processed by an LLM or other tool.

SEAL SECURITY (PRODUCT LAUNCH)

Seal is a supply chain security tool that differentiates itself by providing patches for vulnerable packages instead of pushing for version upgrades.

EBYTE-GO-MORPHER

Ebyte-Go-Morpher is a Go tool that parses, analyzes, and rewrites Go source code to apply multiple layers of obfuscation. It operates directly on the Go Abstract Syntax Tree (AST) and generates obfuscated source files along with runtime decryption logic.

SECURITY PROS ARE DROWNING IN THREAT-INTEL DATA AND IT'S MAKING EVERYTHING MORE DANGEROUS

A Google Cloud and Forrester study of 1,541 tech leaders found that 61% of security professionals are overwhelmed by threat intelligence data, while 60% lack skilled analysts. Top concerns include phishing (46%), ransomware (44%), and AI prompt injection attacks (34%).

PYPI USERS EMAIL PHISHING ATTACK

PyPI users are being targeted by phishing emails from "pypj[.]org" (note the "j"), which request email verification. The fake site mimics PyPI to steal credentials. Users should delete these emails, avoid clicking any links, and change their passwords if compromised.

NEW CHOICEJACKING ATTACK STEALS DATA FROM PHONES VIA PUBLIC CHARGERS

Choicejacking is a new attack that bypasses smartphone security prompts by spoofing USB or Bluetooth input devices to simulate user actions in under 133 milliseconds. The attack tricks phones into enabling data transfer mode without consent, allowing attackers to access photos, messages, and install malware through compromised public charging stations. Unlike traditional juice jacking attacks, Choicejacking exploits device communication protocols to simulate keyboard input and overflow input buffers, affecting both Android and iOS devices.

RTÉ INVESTIGATING POTENTIAL CYBER SECURITY INCIDENT

The UK's National Cyber Security Centre (NCSC) contacted the Irish broadcaster RTÉ regarding a potential cybersecurity threat, based on information that it had received.

MINNESOTA ACTIVATES NATIONAL GUARD AFTER ST. PAUL CYBERATTACK

Minnesota Governor Tim Walz activated the National Guard to provide cyber protection after a cyberattack on Friday crippled the City of Saint Paul's digital services and critical systems.

WE ARE SOCIAL: DIGITAL 2025 JULY GLOBAL STATSHOT REPORT

Search traffic is declining, but search engines are still the top method for discovering new brands, with one in three online adults introduced to brands through search. The share of users citing search as their primary discovery method has grown by 2.5 percentage points over two years, representing over 100 million more shoppers. Google's traffic has dropped by 4% over the past three years, now at 19 billion visits per week, and it is referring hundreds of millions fewer visitors to third-party sites. Despite these shifts and the rise of AI tools like ChatGPT, which has 500 million monthly users, marketers should not overhaul their search strategies yet.

TIKTOK DOUBLES DOWN ON SEARCH ADS AS COMMERCE AMBITIONS COOL

TikTok is ramping up investment in search ads. With over 120 open roles tied to search and rising adoption among brands, TikTok aims to compete with Google for performance marketing dollars. Early data shows promising results, including lower CPCs and higher CTRs.

HOW ADVERTISERS ARE APPROACHING ADS ON WHATSAPP

Advertisers are navigating WhatsApp's limited targeting by prioritizing creative messaging and contextual factors like time of day or device type. While user privacy concerns remain high, brands in retail, tech, and telecom are testing the waters. Early ad formats include low-cost Status ads and AI-powered agents for personalized engagement.

LOW-PRICED PRODUCTS SEEM LESS SUSTAINABLE

Consumers perceive higher priced sustainable products as more eco-friendly, regardless of actual quality. In experiments, lower priced sustainable backpacks were seen as 8.35% less eco-friendly than higher priced ones, and a more expensive detergent was seen as 6.1% more sustainable. This bias stems from the belief that sustainability increases production costs, so low prices trigger skepticism. The effect weakens if companies clearly explain how they keep prices low through methods like automation or bulk buying.

HOW TO FUTURE-PROOF YOUR EMAIL MARKETING WITH BIMI, AMP, AND AI

Email open rates rose 6% and click-to-conversion rates jumped 27.6% in 2024, showing email remains a high-performing channel despite privacy and security challenges. BIMI implementation can increase consumer confidence by 90% while interactive AMP features and AI optimization improve in-email actions and revenue per recipient. These innovations leverage first-party data to maintain deliverability and engagement as tracking cookies disappear.

Read the Delve case study.

MYLENS

MyLens converts raw materials into visuals. It can generate timelines, flowcharts, tables, or quadrant diagrams. Additionally, users can upload data to create line, bar, or doughnut charts.

GOOGLE CHROME ADDS AI-POWERED STORE SUMMARIES TO HELP US SHOPPERS

Google is adding AI-generated store summaries to Chrome for US desktop users. These provide quick insights on product quality, pricing, and service by aggregating reviews from sites like Trustpilot and Yotpo. The feature enhances shopping safety and efficiency while helping Google compete with Amazon and newer AI-powered browsers.

SOCIAL MEDIA MARKETING ETHICS IN THE AGE OF AI

AI is now deeply embedded in social media marketing, helping brands generate content, analyze trends, and automate interactions, but raising ethical concerns around transparency, privacy, and trust. Consumers are increasingly aware, with 41% most likely to call out unethical behavior over anything else, and regulators are responding with stricter disclosure requirements for both ads and AI-generated content. Brands must build trust by being open about data use and AI involvement, protecting privacy, and ensuring inclusive, respectful content. Transparency boosts credibility, as shown by a 2024 study where disclosing AI use in ads increased trust by 96%. Without clear ethical guardrails, brands risk losing customers, talent, and investor confidence.

JENNIFER CLARK (ZOOM) ON TURNING CONTENT FROM COST CENTER TO REVENUE POWERHOUSE

Zoom's five-person content team produced 300 assets last year while shifting content from a cost center to a revenue driver. They used a pod model that united product marketing, demand generation, and content teams to plan quarterly campaigns and measure success through demo bookings and sales calls. AI tools streamlined intake, transcription, and reporting, enabling the team to scale output and prove content's direct impact on pipeline.

'AI MODELS IN VOGUE?' GUESS AD SPARKS UPROAR

An AI-generated model in a Guess ad published in Vogue has led to widespread criticism from readers and influencers calling for boycotts of the magazine.

DIDN'T GO VIRAL? POST IT 7X

Screenshots showing how @lenamoriarty reposted the same video 7 times in a month (with small tweaks) before it went viral.

BRAND LOVE SCALES—SO STOP SETTLING FOR BEING LIKED

Many brands remain stuck in transactional engagement, chasing clicks and discounts without building trust with consumers.

Schedule a demo to learn more

PREDICTIVE CI: WHY SMART TEAMS ARE ALREADY MAKING THE SWITCH

AI coding agents are rapidly increasing code generation, shifting engineers' focus to reviewing AI-written code, which is more challenging than writing it. Current testing methods struggle to address the nuances and potential issues in AI-generated code. Predictive CI uses AI to proactively generate tests, identify regressions, and suggest fixes, which leads to better code reliability and reduces engineer workload.

BOOKMARKABLE BY DESIGN: URL-DRIVEN STATE IN HTMX

ENOUGH AI COPILOTS! WE NEED AI HUDS

The prevalent "AI copilot" model is old. Instead, design "AI HUDs" like the Head-Up Display in airplanes or spellcheck. These focus on human awareness and understanding through subtle, integrated information, offering new senses and insights.

ON THE EDGE OF COMPETENCE

The importance of understanding and expanding one's "circle of competence" is necessary for becoming a better engineer. The circle itself represents fluency in a particular domain, and maintaining a sharp circle happens through continuous learning, revisiting fundamentals, tracking decisions, and seeking feedback.

INTRODUCING STUDY MODE

OpenAI introduced "study mode" in ChatGPT, a new learning experience to guide users through problems step-by-step rather than providing immediate answers. This feature uses custom system instructions to encourage active participation, manage cognitive load, and promote self-reflection.

GLASS3D GENERATOR

A simple code generator for creating 3D glass effects using CSS.

VISUAL CSS

Visual CSS is an AI-powered styling tool that allows for real-time visual editing of any website's CSS.

A FULL CODE AGENT IN 200 LINES—HERE'S HOW

A minimal code agent, dubbed "nano-claude-code" in this article, can be built in just 200 lines of code using JavaScript and Bun. The agent uses tools like reading and writing files and executing bash commands to interact with the local filesystem. The core of the agent is a while loop that interacts with a large language model to plan and execute coding tasks based on user prompts and tool availability.