Skip to content
noisebrief.

Today's tech noise. Briefly.

One email. Every morning.

Past brief

Friday, August 28

AI Power Plays, Courts, and Chaos

A federal judge has ruled that the Trump administration illegally blacklisted Anthropic (opens in a new tab) from government supply chains, a decision that hands the Claude-maker a significant legal victory and exposes the politicization of federal AI procurement. The ruling arrives in the same week that OpenAI, Anthropic, Google, and more than 100 other companies (opens in a new tab) jointly called for coordinated action to defend against rogue AI systems — a rare show of industry unity that reframes AI safety as an urgent governance problem rather than a distant hypothetical. Meanwhile, Barret Zoph (opens in a new tab), the Thinking Machines co-founder who briefly passed through OpenAI, has landed at Google, continuing the talent carousel that makes the frontier AI race as much about poaching as it is about research. The cumulative picture is an industry simultaneously fighting the government, courting it, and racing to staff up faster than rivals can respond.

Rockstar's GTA VI extended look (opens in a new tab), now streaming on YouTube and Netflix, makes one thing undeniable: the $2 billion-plus production has erased the visual line between game and prestige television, with cinematics that critics are already calling film-quality. The release strategy itself is a statement — distributing a game preview across streaming platforms treats the audience like TV viewers, not gamers waiting at a console. Rockstar is essentially running a blockbuster movie marketing playbook for interactive software, and given that GTA V generated over $8 billion (opens in a new tab) in revenue, the stakes for getting this launch right are enormous.

Meta's $18 billion settlement (opens in a new tab) over children's privacy contains a buried clause granting the company a legal pass on future kids' data claims — a provision that consumer advocates say effectively buys regulatory immunity at scale. Separately, Meta is now running a marketing campaign to rehabilitate its Ray-Ban smart glasses (opens in a new tab) after the 'pervert glasses' label stuck following a privacy loophole that allowed wearers to identify strangers in real time using facial recognition. Bluesky, meanwhile, is moving in the opposite direction, adding an algorithmic opt-out (opens in a new tab) feature that lets users choose to stay out of viral recommendation loops entirely — a direct repudiation of the engagement-maximization model that Meta built its empire on. The contrast between the two companies' approaches to user control could not be more deliberate.

Sources · 45