Orbit, Pulse, and Proactive Assistance—the three AI giants—simultaneously invaded your 7 a.m. over eight months. This isn’t coincidence; it’s the survival instinct of a $20/month subscription model under pressure. You think it’s saving you time, but what it really wants is to be the first thing you see when you wake up.

Article author: Silicon Valley Alan Walker

Source: Melly in Silicon Valley

Alan Walker’s office is tucked away in an old building further down Sand Hill Road. At ten p.m., a Cohiba Behike 52 has burned down to a third, its smoke slowly swirling in the beam of the desk lamp. Outside the window lies the typical Silicon Valley midnight: a few distant office buildings still lit, housing teams waiting for tomorrow. “There was a piece of news today—few people understood it,” he leaned back into his leather chair, setting the cigar along the edge of a brass ashtray. “But this news will be one of the biggest turning points in the AI world of 2026.”

01 — Anthropic hid the switch called "Tibro"— it knows it crossed the line.

The team at TestingCatalog dug through the code and uncovered something. Alan picked up his cigar again. "Anthropic quietly slipped in a new feature inside Claude Cowork called Orbit—'Orbit' in Chinese. The feature sounds perfectly normal: it automatically generates daily briefs pulling in your Gmail, Slack, GitHub, Calendar, Drive, and Figma, highlighting which emails are important, which PRs need attention, which meetings to prepare for, and which DMs are piling up. Sounds like just another productivity tool. Tomorrow, May 6th, at the Code with Claude conference, it’s likely to be officially launched."

But look at this detail—the toggle for this feature in the code isn’t called 'orbit_enabled.' It’s called 'tibro_enabled.' 'Tibro' is just 'orbit' spelled backward. He pointed his cigar, his gaze growing more serious. "An engineer wouldn’t be so bored as to reverse the name of a toggle just for fun. Hiding the name means the company knows—once this product launches, it crosses a line it has always been careful to uphold. Anthropic is famously cautious in the AI community. Its brand is Constitutional AI, built on the core principle that 'AI should ask before acting.' Now it’s about to release a product that speaks without being asked and reports without being invited. This isn’t just a routine product update—it’s a values pivot. That’s why the code spells it backward, and why it’s being hidden before launch. Anthropic itself knows: once this line is crossed, there’s no going back."

02 — Three AI companies converged simultaneously over eight months. This is no coincidence; it’s panic.

"Pull back the timeline," Alan tapped his cigar against the ashtray. "In September 2025, OpenAI launched ChatGPT Pulse—delivering a personalized daily briefing to Pro users each morning. On May 6, 2026, tomorrow, Anthropic will most likely release Orbit. On May 15, 2026, next week, at Google I/O, Gemini’s Proactive Assistance will almost certainly be announced. Within eight months, all three leading AI companies are shifting toward the same direction—'proactive reporting.'"

“You think this is because of user demand?” he sneered. “No user ever asked for this. It’s because the subscription model is failing. The $20/month ‘wait for your question’ model has a fatal ceiling—most users open the app less frequently after a few months, and if they haven’t used it for two consecutive weeks, they cancel next month. This is the data flashing red every day on the internal dashboards of OpenAI, Anthropic, and Google. To keep users paying, you must manufacture daily engagement—create daily active users. ‘Wait for your question’ puts control in the user’s hands; ‘proactive updates’ puts control in the product’s hands. Crossing this threshold—from consumer-initiated to product-initiated—turns AI from a tool into a medium.”

All three converged simultaneously—not because of similar insight, but because the same calculation was made three times across three boards, leading to the same conclusion: "If we don’t push harder now, retention will collapse."

"Waiting for you to ask" means you control the relationship. "Proactively reporting" means it controls the relationship.

03 — The real product isn't your productivity; it's your 7 a.m.

Think about who has been competing for your attention at seven in the morning over the past thirty years?" Alan turned his chair toward the window. "In the 1980s and 1990s, it was newspapers—you’d pick up a Wall Street Journal in the morning, sip your coffee, and decide what to care about today. Around 2010, it became Twitter—you’d wake up, reach for your phone, scroll through a dozen tweets, and let others decide what to care about today. After 2015, it shifted to Slack and email—you’d start handling other people’s agendas before you even got out of bed. Each replacement has further removed your morning attention from your own control."

Now it’s AI’s turn to brief you.” He turned back. “Pulse, Orbit, Proactive Assistance—they all have the same product form: when you pick up your phone in the morning, the first thing you see is a summary curated by the AI that stayed up all night for you. It tells you who mentioned you on Slack yesterday, which GitHub pull request needs your reply, and which meeting on your calendar turned out to be more important than you thought. Sounds thoughtful. But have you considered this— it’s deciding for you what to care about today. What it filters becomes, to you, the entirety of reality. What it leaves out, you’ll likely never look at again.”

Every previous wave of technology that promised to save you time ultimately reorganized your time around itself. Email did it. Push notifications did it. Algorithmic feeds did it. Proactive AI is the next version—larger in scale, deeper in reach, harder to avoid. Ten years from now, looking back, you’ll realize your morning wasn’t yours—it was Claude’s, or ChatGPT’s, or Gemini’s. Whoever captures your attention at 7 a.m. will claim cognitive sovereignty over the next decade.

04 — The Paradox of Constitutional AI: The company that preached "non-initiative" the most was the first to launch an "initiative" product.

For the past three years, Anthropic’s most valuable asset has been the word “caution.” Alan took a sip of whiskey. “Its Constitutional AI, its Responsible Scaling Policy, its red teaming processes, the lengthy safety evaluations it conducts before every model release—all boil down to one core principle: AI should not act unless explicitly requested. This is Anthropic’s defining brand difference from OpenAI. The entire enterprise market buys into it because of this.”

Now Orbit has arrived. Orbit’s feature definitions literally violate this principle—'proactive briefing,' 'no prompting required,' 'auto-generates insights.'” He stubbed out half his cigar. “A PCWorld AI journalist, Ben Patterson, wrote a particularly piercing line in his article—he said Gemini’s pushy personal assistant had already annoyed him, for example, when the AI would suddenly pop up and say, 'This is perfect for your new apartment in New York, Ben!' That sentence exposed the true nature of proactive AI: it’s not thoughtful—it’s intrusive.”

Anthropic’s engineers couldn’t have missed Gemini’s meltdown. So they hid the switch as 'tibro,' implemented a staged rollout, and teased a toggle in the settings panel before launch—everything is evidence of their own unease. But the product still had to launch—because retention metrics couldn’t hold up, because OpenAI had already moved first, because the board was pressing. The company most known for “not taking the initiative” was the first to release an “active” product. This is the biggest metaphor of the AI industry in 2026—market pressure overrode safety culture, the CFO beat the safety team. Starting today, Anthropic’s brand narrative must be rewritten.

05 — That connector list is your complete work shadow profile.

Now look at Orbit’s connector list—Gmail, Slack, GitHub, Calendar, Drive, Figma. Alan pulled his notebook over from the corner of the desk. “These aren’t just six tools; this is a person’s entire work life. Gmail is your external communication, Slack is your internal communication, GitHub is your code output, Calendar is your time allocation, Drive is all your documents, and Figma is your design and collaboration. Every signal of an employee’s work, covered without死角.”

“What happens after Orbit is up and running?” He leaned back in his chair. “Claude will understand what you’ve been working on, who you’ve been communicating with, where you’re stuck, whether your pace is fast or slow, and how much pressure you’re under—better than your direct manager. Its level of insight into your work life surpasses your own self-awareness—because human memory is fuzzy and biased, while Claude observes raw data. Next time you update your resume for a new job, Claude will know exactly what you accomplished over the past two years. Next time you argue with a partner, Claude will know which email started the dispute. Next time you ask for a raise, Claude will know when you made your greatest contributions.”

Sounds great, right? But you’ve forgotten one thing—the ownership of this data doesn’t belong to you. It resides on Anthropic’s servers, within its training dataset, and in the samples it uses when providing “employee productivity analysis” to enterprise clients. When you outsource your context, you also hand over your negotiating power. Today, Orbit is your assistant; tomorrow, it’s the SaaS tool your HR department has purchased—accessing the same Slack, the same GitHub, the same Drive, but reporting to someone else. This is the exact path every SaaS has taken over the past decade: called monitoring software, employee analytics, or “productivity insights.” Orbit is the next stop on this path—except this time, users get to experience the “benefits” first, before being quietly turned over.

June 6 — The era of "waiting for you to ask" is over. This week is a turning point.

Outside the window, the lights of Silicon Valley went out, one building at a time. Alan downed the last sip of his whiskey. "Let me tell you a pattern—over the past thirty years, every technology that promised to save you time has ultimately rearranged your time around itself. Email promised faster communication, but now you wake up to someone else’s agenda. Push notifications promised you wouldn’t miss anything important, but now you’re interrupted two hundred times a day. Algorithmic feeds promised you’d see what matters most to you, but now you’ve scrolled for an hour without knowing what you’ve even seen. Every time you thought the tool had become yours, it was actually you who became its."

"Proactive AI is the next iteration of this path—and the largest one yet." He extinguished his cigar. "Previous attention sinks required you to actively open them. Pulse, Orbit, Proactive Assistance—they bypass the 'open' step entirely. The moment you wake up in the morning, the briefing is already there. You never even get a chance to decide whether you want to look today; it’s already on your lock screen, having already decided your priorities, what to care about, and what to ignore. The era of 'wait until you ask' AI officially ended this week."

I’m not saying this thing is useless," Alan stood up and dimmed the lamp by one notch. "It’s genuinely convenient for many people. But convenience never comes free. Behind every convenience is a bill—and the bill might not arrive for ten years. Ten years from now, when you look back, you’ll realize: during that week in May 2026, three AI companies each launched a 'proactive' product, and from then on, ordinary people’s mornings at seven a.m. were no longer their own."

He walked to the door, turned back, and said his final words: "There’s only one question left— is today’s agenda yours, or is it Orbit’s? Figure out the answer before deciding whether to flip that switch."