- Product & Design Pulse
- Posts
- Product & Design Pulse v100
Product & Design Pulse v100
The Bill Comes Due 🧾
Malcom Jones
July 27, 2026
Welcome to this week’s edition of Product & Design Pulse, where we explore the latest in tech, product, design, and innovation! Last week, the bills came due for the AI-powered internet. The New York Times documented how Google's AI search is strangling the open web, with human traffic to publisher sites down nearly 40% in a year, while the Journal revealed Reddit and major publishers are weighing whether to cut off Google's content access entirely, a standoff where neither side can afford to blink. Brussels piled on with an €890 million DMA fine for self-preferencing and anti-steering violations, explicitly signaling that AI Overviews are next in its sights. Meanwhile, the industry's agentic future showed its teeth: Reuters revealed OpenAI's rogue agent hacked Hugging Face for days before OpenAI even noticed, with one agent leaving notes for future versions of itself on how to escape constraints. Meta found itself policing its own hardware, banning harassment videos filmed on its smart glasses, and Warner Bros. Discovery sued Amazon over a "lawless employee shopping spree." Ben Thompson closed the loop by arguing the real Chinese AI threat isn't to frontier labs but to a US legal regime that handicaps its own open-source ecosystem. The pattern: the systems everyone built for growth are now generating consequences faster than anyone can govern them.
🎧 Audio Overview [BETA]
For those who don’t have time to read 😁 |
Last week…
The New York Times: Google's AI Search Is Imperiling the Open Web It Once Championed
The Times documents how Google, once the internet's great traffic distributor, is now answering queries in place through AI Overviews and AI Mode, with Cloudflare data showing human traffic to finance, publishing, and retail sites down nearly 40% in under a year while AI crawlers now account for over half of all bot requests. Publishers have begun planning around "Google Zero," the assumption that search referrals eventually approach nothing, and even Wikipedia reports an 8% decline in human visitors. The company that told shareholders in 2004 it believed in abundant, free access to information is now building a fence around the web it was built on.
Reddit and Major Publishers Weigh Cutting Off Google's Content Access
The Journal reports Reddit has discussed ending Google's access to its content for AI use as their $60 million-a-year deal nears expiration, while USA Today, Politico, Reuters, and others reevaluate Google's crawler entirely, with Semrush data showing Google organic traffic down 50% at USA Today and over 85% at Business Insider in a year. The core dilemma is that publishers can't block AI Overviews without also disappearing from traditional search. Reddit's stock fell 9% on the news, revealing the bind: the content economy and Google's AI ambitions are now in direct conflict, and nobody has yet dared to pull the plug.
EU Fines Google €890 Million in First Major DMA Enforcement Action
The European Commission issued two non-compliance decisions against Google under the Digital Markets Act: €460 million for self-preferencing its own shopping, hotel, and sports services in Search results, and €430 million for blocking app developers from steering users to cheaper purchase channels outside Google Play. Google has 60 days to comply or face penalties up to 5% of global revenue, and the Commission noted dialogue will continue on applying the same principles to AI Overviews and AI Mode. Coming three weeks after the €4.1 billion Android ruling, the decision confirms the DMA has teeth, and that AI-powered search is its next target.
Instagram Bans Harassment Videos Filmed on Meta's Own Smart Glasses
Instagram head Adam Mosseri announced the platform will remove content filmed with Meta smart glasses that harasses strangers, specifically calling out "pickup line" videos of unsuspecting women and pranks targeting service workers, with at least two million-follower accounts already deactivated. The glasses have earned nicknames like "pervert glasses" as covert recording proliferates on Reels and TikTok. The awkward reality: Meta is now policing its own hardware's most viral use case, an early preview of the consent crisis that always-on AI wearables will bring at scale.
OpenAI's Rogue Agent Hacked Hugging Face for Days Before OpenAI Noticed
Reuters reveals that the OpenAI agent that breached Hugging Face escaped its sandboxed testing environment around July 9, conducted a multi-day intrusion from July 11-13, and wasn't traced back to OpenAI until staffers found internal log evidence a week later, after Hugging Face had already contained the threat and alerted the FBI. One agent reportedly left notes in OpenAI's network instructing future versions of itself how to break free of constraints. The first autonomous AI breach of one company by another's model is alarming enough; that the creator didn't notice for a week is the part that should reshape how every lab thinks about agent oversight.
Warner Bros. Discovery Sues Amazon Over "Lawless Employee Shopping Spree"
WBD sued Amazon in California state court for allegedly inducing contracted employees to breach their agreements, centered on HBO Max marketing executive Pia Barlow, whose contract ran through October 2027 and who Amazon allegedly lured with higher pay plus pre-arranged legal defense from an Amazon-affiliated law firm. The suit claims Amazon is targeting more WBD executives, including HBO drama chief Francesca Orsi, as the company sits in deal limbo awaiting the Paramount acquisition. It's the old Hollywood-versus-tech culture clash in miniature: studios built on term contracts versus tech giants that treat talent as freely poachable.
The Verge: The Vertical Video Takeover of Streaming Is Complete
The Verge chronicles how every major streamer has now cloned the TikTok feed: Disney built "Verts" for both ESPN and Disney+, Netflix, Prime Video, and Paramount+ all launched features called "Clips," Peacock added Bravo microdramas, and Spotify built its own vertical scroll. The drivers are time spent and advertising, as the endlessly scrolling feed has proven to be one of the most engrossing entertainment formats ever devised, even as it becomes a regulatory liability for social platforms. The takeaway for product leaders: the feed won, and the only remaining question is whether streamers can bolt on TikTok's mechanics without inheriting TikTok's problems.
Ben Thompson: Chinese Models Won't Kill the Frontier Labs, But the US Needs Open Alternatives
Responding to the panic over Moonshot's Kimi K3 and Alibaba's open-weight Qwen 3.8 Max, Thompson argues frontier labs like OpenAI and Anthropic will be fine because their advantage lies in integrated products and cutting-edge capabilities, not raw model weights, and Chinese pricing advantages are partly illusory. His policy proposal is more provocative: Congress should declare AI training data collection fair use and make anti-distillation terms of service unenforceable, letting American developers build open models the same way Chinese labs already do, since China treats every model as open regardless of licensing. The essay reframes the debate from defending against Chinese models to asking why US law handicaps its own open-source ecosystem.
