Daily Briefing — June 9, 2026

Apple finally did something about Siri. At WWDC 2026, the company announced a full overhaul of its long-struggling voice assistant, including a standalone Siri app, chatbot-style conversations with memory, and the ability to pull context from whatever is on your screen. Apple is also partnering with Google Gemini to power the underlying model — a notable move for a company that has historically treated its tech stack like a private club.

Siri has been an embarrassment in slow motion. While ChatGPT, Gemini, and Claude were changing how people work, Siri was still struggling to set timers correctly. The gap got so obvious that Apple could not pretend the old approach was fine. This announcement is essentially Apple admitting it lost a few years and is now trying to buy back credibility through a partnership it probably did not want to make.

What sets this version apart from past Siri promises is the personal data angle. An assistant that knows your context, your screen, your habits, and your history is a fundamentally different tool than one that just queries the internet. Whether Apple can execute that without becoming a privacy story is the real question to watch as this rolls out later this year.

S&P Dow Jones Indices just told SpaceX no. After a monthlong consultation about whether to waive or bend its entry requirements for so called MegaCap companies, the index committee held the line. SpaceX had reportedly made accelerated inclusion in major indexes a condition of its IPO, which would have funnelled passive fund money automatically toward the company the moment it listed. The committee said thanks but no thanks.

This matters more than just one company's stock market debut. If S&P had carved out an exception here, it would have opened the door for OpenAI, Anthropic, and every other highly valued but not yet profitable AI company to fast track into indexes right after their own IPOs. Your retirement fund, if it tracks the S&P 500, would have started absorbing that exposure without you ever clicking a single button. The committee essentially blocked that from happening automatically.

AI companies right now are burning through capital, struggling to build data centers, and shifting costs onto customers through usage-based pricing. The S&P's profitability requirement exists because index inclusion was never supposed to be a reward for ambition — it was supposed to reflect actual, sustained business results. Someone on that committee remembered that.

Meta shipped face recognition code to over 50 million phones while publicly saying it was still "thinking through" whether to use face recognition. The feature, internally called NameTag, has been baked into the Meta AI app since at least January. It is not active yet, but the scaffolding is already sitting on your device, waiting for a server update to flip the switch.

The gap between what Meta said publicly in April and what Wired found in the actual code is significant. This is not a leaked internal prototype or a rogue engineer. This is production code, distributed at scale, to millions of people who use it to run their Ray Ban and Oakley smart glasses. When it gets activated, it will convert faces captured by those glasses into biometric faceprints and check them against a locally stored database that Meta controls and updates remotely.

The deeper implication is a trust architecture story. Whether this technology is legal or whether it works is almost beside the point. The more telling thing is that a company can pre-position capabilities on your device before disclosing them, then activate those capabilities on their own timeline. That is a design choice, and it tells you something about how the people building these tools think about consent.