As unions confront the prospect of humanoid robots on factory floors, Raycast co-founder Thomas Paul Mann argues that AI is also making software creation radically more personal. The conversation traces a shift from bloated, one-size-fits-all tools toward custom Mac apps, disposable software, and workplace systems assembled for the way people actually work.

Casey Newton and David Pierce sort the genuinely useful AI aids from the new wave of productivity theater, arguing that simple systems and reliable capture matter more than ever. Along the way, they weigh where AI actually saves time, from transcription to summarizing source material, and where it mostly adds friction, false confidence, and more work.

Casey Newton tests Raycast’s Glaze by building a Nightwing-themed to-do list, a custom Platformer archive search tool, and a half-finished source tracker, using the experience to argue that AI-made software is getting more visual, more personal, and more immediately useful. The appeal is less technical novelty than the thrill of instantly reshaping the tools that used to make users live with their compromises.

AWS chief Matt Garman argues that AI will reorder white-collar work rather than erase it, even as Amazon automates more tasks and trims parts of its own workforce. The conversation ranges from college AI degrees and shaky labor data to enterprise adoption, data center backlash, and the question of whether efficiency gains create new jobs fast enough.

At Platformer’s first live show, Casey Newton and guests take stock of AI’s contradictory role in the workplace: a source of anxiety, a generator of strange new job titles, and a tool that may make software creation radically more personal. The conversation moves from Atlassian’s enterprise “context graph” to Eugenia Kuyda’s vision of companions and bespoke apps, tracing how automation may reshape both office work and the products people use every day.

Brookings fellow Molly Kinder argues that the real danger of AI is not an instant jobs apocalypse but a long, destabilizing period of selective white-collar displacement for which government and industry have no credible plan. The conversation weighs retraining, safety nets, and broader wealth-sharing proposals against a future in which automation chips away at the skills and careers workers once thought were secure.

Labor economist Catherine Ann Edwards argues that the real danger is not an AI apocalypse but a government safety net too weak to handle ordinary job loss, recessions and worker disempowerment. As chip workers at Samsung win huge bonuses through union pressure, she makes the case for stronger unemployment insurance, labor power and tax policy instead of waiting for a technological crisis.

Anthropic’s Boris Cherny argues that AI coding tools are already blurring the boundaries between engineer, manager and designer, even as their labor-market effects remain unsettled. Around that debate, the conversation traces how companies are pushing workers to adopt AI, rewarding token usage unevenly and fumbling toward a broader social response to automation.

A Google executive and longtime labor economist argues that AI is more likely to reshape work than erase it, even as new forecasting data suggests a small but meaningful chance of far more severe disruption. The conversation pits Silicon Valley’s job-loss rhetoric against a slower, messier reality in which tasks automate faster than occupations, and policy failures may matter more than the technology itself.

Casey Newton and Ella Marciano open a new Platformer series with survey data showing AI adoption is highest among managers and top earners, while junior workers and administrative staff remain more skeptical of its benefits. A later conversation with Box CEO Aaron Levie argues that enterprise software will be remade by agents, but that most jobs will shift rather than disappear as human work moves to the harder, higher-value last mile.