ChatGPT wants to be your morning briefing now. OpenAI just rolled out ChatGPT Pulse, a proactive daily digest that surfaces personalized updates before you even ask for them. Think of it like Google Now’s info cards resurrected, but powered by a chatty LLM that’s been quietly doing homework in the background.
Here’s the pitch: Pulse compiles a short, tailored summary each day based on “asynchronous research” ChatGPT performs on your behalf. You can steer it toward or away from certain topics, and it will also draw on your chat history and, if you opt in, data from Gmail and Google Calendar. In OpenAI’s examples, that translates into follow‑ups on things you’ve been discussing, quick dinner ideas, or incremental steps toward longer‑term goals—like nudging you through triathlon training instead of letting that impulse purchase of carbon shoes rot in the closet.
For now, Pulse is limited to OpenAI’s Pro tier, the same paywall where the company is parking its most advanced features. Engadget notes that Pro runs a steep $200 per month, a number that instantly reframes Pulse from “neat” to “this better save me time, money, or my job.” OpenAI says it will eventually roll this out more broadly.
Zooming out, Pulse is a clear move from a reactive chatbot to a proactive assistant. Google tried this a decade ago with Now, Microsoft has been doing a flavor of it with Copilot’s Briefing emails, and Apple keeps flirting with relevance via Siri Suggestions. The difference here is that ChatGPT’s generative core can synthesize, prioritize, and nudge across disparate inputs with fewer rigid templates. If it works, you get a lightweight chief of staff that preps your day—without having to spelunk through five apps and a thousand red badges.
Of course, the devil lives in the defaults. Proactive assistants win or lose on three things:
Trust: Connecting Gmail and Calendar is an opt-in, but people won’t hand over their crown jewels unless they understand what gets ingested, how long it’s retained, and where it flows. Google caught heat for Bard’s Gmail/Drive integrations; OpenAI will need boring, explicit data handling docs and easy kill switches.
Precision: Summaries can’t hallucinate. A daily brief that invents meetings or “helpfully” misreads an email thread is dead on arrival. Pulse will need ruthless source discipline and transparent citations if it’s going to graduate beyond novelty.
Control: The more a system guesses what matters to you, the more you need to steer and correct it. Topic tuning is table stakes; fine-grained controls over tone, frequency, and data scope will determine stickiness.
Strategically, this also solves a business problem for OpenAI: usage gravity. Right now, ChatGPT is something you launch when you need a thing. Pulse gives it a recurring slot in your morning routine—habit formation that competitors would kill for. It’s also a smart place to bundle adjacent upsells (agentic tasks, inbox triage, meeting prep) under that $200 banner for power users who can justify the spend because time literally is money.
The open questions are the same ones that have stalked every “intelligent” feed since 2012. Will it surface what you actually need versus what’s easy to find? Can it distinguish “interesting” from “actionable” on a Tuesday at 8:07 a.m.? And when it gets something wrong—as all models do—does the correction loop make it better tomorrow or just feel like you’re QA’ing a very expensive intern?
Verdict for now: Pulse is the right idea, finally shipping at the right moment. If OpenAI nails reliability and transparency, this could be the first widely adopted AI daily brief that doesn’t feel like a toy. If not, it’s another well-meaning assistant that spams you with vibes while your real work happens elsewhere. Either way, proactive AI is the game—and OpenAI just made its move.
No Comments