Are we entering the post-attention era?
When I was working, the first thing I’d do after arriving at the office was open HackerNews and browse through the front page, post by post. I really enjoyed that feeling of consuming high-quality, useful information. Over time, I realized that this habit not only helped me break free from the “information bubble,” but also sparked ideas later on—moments when something would suddenly remind me of a post I’d seen on HackerNews.
However, this habit had one big drawback: it required a lot of focused time. Back then, it would often take up an entire morning.
Now, with the advent of AI—especially large language models (LLMs)—I’ve realized that LLM-based technology can: 1.Help me watch things, like YouTube videos or podcasts. 2.Help me read. 3.Help me think. 4.Store my memories.
This made me realize that I could completely automate this habit. I could save time by simply reviewing the AI-generated highlights instead of manually going through everything. To make this happen, I developed an iOS app based on HackerNews. It not only helps me browse HackerNews but also has a powerful memory feature. Anything I see or browse in the app is stored in a local vector memory. When I need to recall a specific piece of news or even a meme, I can search for it naturally using plain language.
While developing this app, I discovered something interesting: my mental flow could also be "augmented" by this tool. What do I mean? Let’s say I read news item A on one day and news item G on another day. Over time, I might forget both. But they still linger in my subconscious. Then, one day, while walking, A and G might suddenly connect and spark a new insight. To support this kind of serendipity, we built a “Flow” module into the app to help externalize and augment parts of my thought process. Of course, this is still an area I’m actively exploring and developing.
Recently, we also added a “Listening” feature to the app. I believe that listening—whether during conversations or by having the app listen passively—can inspire countless new memes and ideas.
So far, I’ve released 33 versions of this app, and I’ve been privately using it to great effect. It’s now available on the Apple App Store: Echostream - Mind Copilot:https://apps.apple.com/us/app/echostream-mind-copilot/id6476086097. I welcome you to give it a try!
Privacy is something I take very seriously. It’s incredibly important, so we’ve fine-tuned a model to run locally within the app. This means most of the processing happens on your device and doesn’t sync to any server.
As we move into the era of AGI, I’ve been wondering: Are we entering a “post-attention” age? Over the past 30 years, so much of our online attention has been consumed by manual interaction. With AGI, could this attention be automated, freeing us to focus on what we truly enjoy? Wouldn’t that make us more human, rather than being repetitive workers shaped by the industrial age?
I hope this app is helpful for you all.