Skip to main content

ChatGPT Pulse

Ref: https://openai.com/index/introducing-chatgpt-pulse

OpenAI has introduced a new feature for ChatGPT called Pulse, which is currently available as a preview for Pro users on mobile devices. Pulse transforms ChatGPT from a reactive tool to a proactive assistant by delivering personalized updates.

Here's how it works:

  • Proactive Research: Every night, Pulse analyzes your chat history, memory, and direct feedback to identify topics relevant to you.

  • Personalized Updates: The next day, it presents these findings in the form of visually appealing, easy-to-scan cards. These updates can include follow-ups on previous conversations, suggestions for daily activities, or progress on long-term goals.

  • Enhanced Personalization: You can optionally connect your Gmail and Google Calendar to further personalize the updates.

  • User Control: You can provide feedback on the updates and directly request topics for research, which helps improve the relevance of the information provided.

Pulse is still in its early stages and may not always be accurate, but it's the first step towards an AI that proactively assists with research, planning, and other helpful actions. Future plans include integration with more applications and the ability to deliver timely information throughout the day.

Labels:

Comments

Post a Comment

Popular posts from this blog

A Wider Perspective on Chips, Code, and AI

Ever since I was a kid, I've been fascinated by the magic of technology. Not the stage-magic kind, but the real, tangible magic that happens when you etch impossibly small patterns onto a slice of silicon, or when you write lines of code that spring to life, learning and creating in ways we're only beginning to understand. For me, taking apart an old radio wasn't just about seeing the components; it was about trying to grasp the invisible logic that connected them. That curiosity never left. It led me down a rabbit hole that became a career, a path that has placed me right at the intersection of hardware and software engineering. I've spent years with one foot in the cleanroom, marveling at the physics of semiconductor fabrication, and the other foot in the command line, crafting the software that breathes intelligence into that silicon. I’ve come to see these two worlds not as separate disciplines, but as two sides of the same revolutionary coin. And that's why I...
1 comment
Read more

LLM as of now

Ref: https://blog.arcbjorn.com/state-of-llms-2025 In the late-2025 landscape of Large Language Models (LLMs), the era of a single, dominant AI model has conclusively ended. Instead, we are now in a specialized ecosystem where different models excel at specific tasks. This shift, as detailed in the blog post "State of LLMs in Late 2025," has led to a more diverse and competitive market, with users needing to understand the key technical differentiators between models to select the right tool for their needs. These differentiators include the model's architecture, the data it was trained on, and its fine-tuning methods. The major players in the LLM space of late 2025 each have their own strengths and ideal use cases. OpenAI's GPT-5 is noted for its unified intelligence system, making it a versatile and powerful option for a wide range of applications. Anthropic's Claude Sonnet 4.5 has established itself as the leader in coding and autonomous tasks, a go-to for devel...
Post a Comment
Read more