When an AI therapy chatbot got banned in the UK, 54% of its users didn't switch to a regulated alternative - they went straight to ChatGPT.

This week we explore why the health sector is being swallowed by the same playbook that gave us Uber and Netflix: optimise it, make it frictionless, inflate the value, charge a subscription.

From looksmaxxing apps running fake AI to one-person GLP-1 startups doing glorified doctor shopping, we ask whether tech is solving health problems or just selling patches for a flat tire nobody wants to fix.

We put Moda's $7.5M "taste" pitch and Google Stitch to the test with a deliberately chaotic brief. The results say a lot about where AI design actually is, and isn't. We unpack why Silicon Valley keeps claiming "taste" as a competitive moat, what that means for designers' careers, and whether AI homogenisation might accidentally make great design more valuable.

A Melbourne startup is powering computers with human brain cells grown from stem cells. They call it "wetware" instead of software. These neurons are learning, adapting, and playing Doom - using 99% less power than traditional chips.

Have you ever received an email from your boss and thought "did they actually write this, or was it ChatGPT?"

This week we explore what happens when we lose the "micro-conversations" that build real connection, why AI-generated communication feels like a trust violation even when it's efficient, and whether we're sleepwalking into a Wall-E future where nobody actually talks to each other anymore.

Topics: AI at work, ChatGPT communication, workplace trust, algorithmic personalisation, human connection, future dystopia

We dig into China's stunning leap in AI video generation, what their approach to training data means for the global AI race.

We also explore the liar's dividend, the idea that the more you understand deepfakes and AI-generated media, the less you're able to trust what you see. Awareness doesn't seem to protect you. It might actually make things worse.

From AI video generation and disinformation to the collapse of visual evidence in journalism and politics, we explore why 2026 might be the year "seeing is believing" stopped being true.

We dig into China's stunning leap in AI video generation, what their approach to training data means for the global AI race.

We also explore the liar's dividend, the idea that the more you understand deepfakes and AI-generated media, the less you're able to trust what you see. Awareness doesn't seem to protect you. It might actually make things worse.

From AI video generation and disinformation to the collapse of visual evidence in journalism and politics, we explore why 2026 might be the year "seeing is believing" stopped being true.

Late last year, an open-source AI agent went from zero to 60,000 GitHub stars, got acquired by OpenAI, and spawned a social media platform where 1.5 million AI bots started posting, forming social hierarchies, and having existential crises. One bot just kept asking "does anyone else feel this?"

WDTFW unpack the OpenClaw phenomenon, from the hype explosion to the security nightmares to the uncomfortable question of what it means when AI starts acting as lost and confused as we are online. What is real and what is hype?

Dario Amodei, CEO of Anthropic (the company behind Claude AI), just published an essay about AI that reads like a warning letter to civilisation. He predicts massive job displacement, military risks, and potential civilisational collapse from the technology his company is building.

This week we break down:

• Dario's main arguments about AI risks and timeline

• The contradiction between dire warnings and aggressive growth
• What Silicon Valley leaders gain by predicting doom while accelerating

• How seriously should we take these warnings

How does AI learn ethics? This episode breaks down Anthropic's Constitutional AI approach - the system that gives Claude AI its values, personality, and ethical boundaries.

In part one of our 2026 predictions series we explore: The "offline luxury" phenomenon and digital class divides, Australia's under-16 social media ban, the psychology of doomscrolling and self-judgment, how platform algorithms keep us cycling between apps, the death of creator culture and rise of passive consumption.

The data sources we reference in this episode come from: datareportal.com, theharrispoll.com and partnercentric.com