Search the archive

BUILD SCALABLE APPS & WEBSITES WITH AI

AppWizzy is a vibe‑coding platform for building scalable web apps and websites.

PATTERN MAPPING FOR ADHD & NEURODIVERGENT MINDS

Unloop is a visual pattern mapping tool that helps you see the triggers, thoughts, emotions, and behaviors that keep you stuck.

FROM SKETCHES TO POLISHED AI ART

Compose your ideas, add a prompt, and generate stunning visuals, all in one creative workspace.

I USED NANO BANANA PRO TO CREATE 20 SOCIAL MEDIA POSTS, THE RESPONSE THEY GOT WAS SHOCKING

A journalist tested Google's Nano Banana Pro image generator by creating over 20 social media graphics, resulting in a 57% increase in LinkedIn impressions. The tool produced professional-looking explainer graphics, inspirational posters, and controversial images within seconds, though it frequently hallucinated information. While the AI-generated images boosted engagement and looked polished, they lacked personality and an authentic human touch, with attempts to create viral content falling flat despite the tool's technical competence.

LOWER THE SURPRISE: APPLYING THE FREE ENERGY PRINCIPLE TO UX

The Free Energy Principle describes how the brain aims to minimize the mismatch between expectations and sensory input. It helps explain why design principles like feedback, Gestalt grouping, and clear states feel intuitive while broken interactions cause stress and frustration. By using abductive reasoning—forming hypotheses about user expectations and emotional context—designers can go beyond rules and create solutions that better align perception with expectation and reduce user uncertainty.

FRONTIFY PRESENTS: HOW GRETEL AND PHILADELPHIA ART MUSEUM NAVIGATED CHANGE

Gretel partnered with the Philadelphia Museum of Art on a bold rebrand aimed at making the institution feel more open, local, and welcoming while preserving its authority, grounding the work deeply in Philadelphia's culture, history, and vernacular. Through close collaboration, they evolved the museum's name, redrew the historic griffin, created a custom typeface, and built a flexible brand system that balances heritage with accessibility and contemporary relevance.

FIELD KALLOP MEDITATES ON UNIVERSAL PATTERNS THROUGH BOLD CHROMATIC COMPOSITIONS

Field Kallop creates abstract paintings exploring universal patterns connecting the micro and macro, drawing inspiration from astronomy, physics, and spiritual iconography.

WHY MCDONALD'S AI CHRISTMAS AD BACKFIRED SO BADLY

McDonald's Netherlands pulled its AI-generated Christmas ad after a strong backlash over its uncanny visuals and bleak tone, echoing criticism recently faced by Coca-Cola's AI holiday campaign.

Dive in and start building

HOW WE PWNED X (TWITTER), VERCEL, CURSOR, DISCORD, AND HUNDREDS OF COMPANIES THROUGH A SUPPLY-CHAIN ATTACK

Daniel, a 16-year-old hacker, along with friends, uncovered supply-chain vulnerabilities in Mintlify, an AI documentation platform used by many top companies. Daniel specifically found a cross-site scripting (XSS) vulnerability that allowed malicious scripts to be injected into documentation through SVG files, exploiting Mintlify's internal file fetching. This flaw had a widespread impact, affecting major customers like Discord, X (Twitter), Vercel, and Cursor, but Mintlify quickly fixed the issue once the hackers notified it.

WHAT IS EC2 INSTANCE ATTESTATION

EC2 instance attestation makes it easier to validate that only trusted software is running on EC2 instances. It is now available in all AWS Commercial Regions, including the AWS GovCloud (US) Regions. EC2 Instance Attestation can unlock more use cases for trusted execution environments, like LLM use, since it allows the utilization of GPU features. This article provides a demo that shows how a service provider can build and deploy a service that can be verified from source code, and how a service consumer can verify those claims by the provider.

YOUR JOB IS TO DELIVER CODE YOU HAVE PROVEN TO WORK

A software developer's core responsibility is to deliver code that has been proven to work. The practice of submitting large, untested pull requests is a dereliction of duty that unfairly burdens reviewers. Proving code works involves two essential steps: thorough manual testing to personally verify functionality and demonstrate it, and bundling automated tests that validate the change.

ONE WEIRD TRICK TO MANAGE ENGINEERING CRISES; STAKEHOLDERS LOVE IT

Tactical wins address a specific symptom with a specific fix. These wins matter because they demonstrate traction. Strategic investments don't fix individual problems, but instead eliminate entire classes of problems. The trick is to sequence tactical wins and strategic investments to reinforce each other. Transparency will help stakeholders understand the tradeoffs you are making.

AI-connected infrastructure

INTRODUCING GPT-5.2-CODEX

OpenAI has launched GPT-5.2-Codex. This new model improves performance on long-horizon tasks, large code changes, and in Windows environments. It sets new benchmarks in agentic coding. The model's capabilities have already led to the discovery of real-world vulnerabilities.

PILOTING CLAUDE IN CHROME

Claude in Chrome has an integration with Claude Code, where Claude Code can test code directly in the browser to validate its work. Claude can also see client-side errors via console logs.

PICKNPLACE.JS

picknplace.js is a simpler alternative to the usual drag-and-drop experience, especially on mobile devices. It streamlines the process into a three-step interaction: pick, scroll, then place. When an item is picked, an interactive duplicate of the list appears, allowing users to easily position their selection.

NVME IS NOT A HARD DISK

NVMe (nonvolatile memory express) is a storage access and transport protocol for flash and next-generation solid-state drives that delivers high throughput and fast response times for all types of enterprise workloads. NVME over fabric over TCP allows you to make storage available even if it is in the wrong box. Distributed storage may be complicated, but it has a number of very attractive use cases.

THE ANATOMY OF LCP – A DEEP DIVE INTO SUB-PARTS

Largest Contentful Paint (LCP) is a Core Web Vitals metric that measures when the most meaningful piece of content on a page becomes visible. Optimizing it matters because users should see this primary content as early as possible. There's more to optimizing than just reducing image file sizes. This article breaks down LCP into sub-parts (Time to First Byte, Resource Load Delay, Resource Load Duration, and Element Render Delay) to explain exactly where time is being spent, so readers know how to diagnose slow LCP on sites.

out the research preview today.

MINIFICATION ISN'T OBFUSCATION - CLAUDE CODE PROVES IT

The minification process in JavaScript doesn't actually secure anything, it just makes it hard to read.

A BETTER WAY TO WORK WITH CLAUDE CODE

Nimbalyst is a local WYSIWYG editor and session manager that lets users iterate with Claude Code with full context.

ONE AGENT ISN'T ENOUGH

Agentic coding often struggles with variance, as the stochastic nature of LLMs means a single agent run might miss the optimal solution.

GITHUB WALKS BACK PLAN TO CHARGE FOR SELF-HOSTED RUNNERS

GitHub reversed its decision to charge for self-hosted Actions runners, postponing the billing change to re-evaluate its approach.

CHECK IF YOU'RE ELIGIBLE

STATE OF CONSUMER AI 2025

a16z has released a ranking of the top 100 consumer AI apps. The ecosystem is stabilizing - there were only 11 new names on the web list, compared to 17 last time. ChatGPT still leads by a wide margin, but Google's Gemini and xAI's Grok are closing the gap. The more interesting finding is what's working. Vibe coding startups Lovable and Replit both debuted on the main list. 22 of the top 50 mobile apps were developed in China, but only 3 are primarily used there. Only 9% of consumers pay for more than one AI subscription.

THE END OF INTERCHANGE

The interchange fees that make up much of the cost of the credit card system are eventually going to go away. The interchange fee model is vulnerable to disruption from stablecoins, bank-to-bank direct payments, Pay by Bank platforms, and cryptographic identity layers in ACH. It is still an open question whether stablecoins or retrofitted ACH/bank rails will win, how fraud and reversibility concerns will be addressed, and what the timeline is for meaningful interchange compression.

THE TOP REASONS I SAW CROS FLAME OUT THIS YEAR

CROs can't just be managers with playbooks. To be successful, they need to be closers who learn, leaders who show up, and executives who stay when it gets hard. They need to personally close deals and be familiar with products, as well as understand their partner channels. A good CRO visits their major teams in person because they know that culture gets built in person and trust is earned face-to-face.

THE LAWYER MOST B2B STARTUPS DON'T REALIZE THEY NEED UNTIL FAR TOO LATE

Startups that bring on contract counsel generally see their overall sales cycle reduced. Founders signing at least two enterprise contracts a quarter should consider bringing on contract counsel. It will save the startup money and often help close deals faster. Having counsel will help startups avoid ending up with terms that let customers take advantage of them or create problems later down the line, like during an acquisition.

THE 13 TACTICS BEHIND LOVABLE'S INSANE GROWTH

Lovable went from $0 to $100M ARR in 8 months with 45 employees and almost no paid acquisition. That's $2.2M revenue per employee - the industry benchmark is $200k. The playbook was running 12+ growth channels simultaneously while competitors struggled with one or two. A Product Hunt launch drove GitHub stars. GitHub credibility led to podcast invitations. The flywheel required no paid advertising to maintain.

GOWATCH

Create monitors that watch the web for you.

NOVA

Nova automates resume screening, pre-screening calls, and first-round interviews.

AMPCODE

Amp is a frontier coding agent that lets you wield the full power of leading models.

DESIGNERS OF PROBLEMS WORTH SOLVING

The best founders aren't solving obvious problems - they're designing problems worth solving. The distinction matters because finding meaningful problems may be more important than solving them. The approach involves reframing situations through expansion and elevation until the problem aligns the founder's purpose with how society and technology are changing. This lens helps identify founders who have the agency for entrepreneurship, even if their ideas seem "unfundable" by consensus standards.

SOFTWARE APPLICATIONS FACE A NEW INTERMEDIARY

Many years ago, Amazon tried to build its own device to escape Apple's control. AI companies now see personal devices as a way to escape the App Store tax and to own the user relationship. These companies don't want to live at the mercy of operating systems they do not control. Creating their own devices will allow them to own their own destiny rather than ask Cupertino for permission.

SOME THOUGHTS ON "HEAD OF NARRATIVE"

This essay explores why tech companies are suddenly paying $300k+ for "heads of storytelling" and whether it actually works. Anthropic is offering up to $400k for a "Head of GTM Narrative." The author's skepticism comes down to one thing: companies are extensions of their founders. Public perception of Anthropic is downstream of Dario, regardless of what a narrative lead says.

all free

BACKCHANNELING IS BECOMING A CRUTCH

Executives who can actually move a company aren't always universally liked - you have to talk to them before letting someone else's context make the call.

DIGITAL ADVERTISING'S "BITTER LESSON" MOMENT

Explicitly labeling an ad as 'AI-generated' or 'AI-edited' can cause CTR to plummet significantly.

I SURVIVED 2025

A founder's review of this year, having survived another year of self-employment.

VALVE: THE REVERSE APPLE

Valve created an ecosystem first before creating its own first-party hardware that leverages all that work.

TWO CHROME FLAWS COULD BE TRIGGERED BY SIMPLY BROWSING THE WEB: UPDATE NOW

Google patched two Chrome vulnerabilities exploitable via malicious web pages: a use-after-free in WebGPU (CVE-2025-14765) that enabled heap corruption, and an out-of-bounds read/write in the V8 JavaScript engine (CVE-2025-14766). Both flaws affect Chrome's 3.4 billion users and require no user interaction beyond visiting a crafted site. Update to version 143.0.7499.146 or later immediately.

113,000 IMPACTED BY DATA BREACH AT VIRGINIA MENTAL HEALTH AUTHORITY

Richmond Behavioral Health Authority in Virginia was hit by a ransomware attack in late September that encrypted parts of its network and revealed data on more than 113,000 individuals, including names, Social Security numbers, passport information, and financial and health details. Victims are advised to monitor their accounts and credit reports closely, as the ransomware group Qilin claims to have leaked 192 GB of stolen data.