ImLonky

Compared to internet products, AI has a personal characteristic: it's very difficult to find its aha moment.
First, you need to have a clear problem.
Second, you need to have an evaluation standard for the quality of AI output.
Finally, you must use it correctly with the right methods.
Only then can you truly experience the uniqueness of AI products, rather than feeling that they are generic.

I believe that living in this world, one still needs to have some of their own opinions and biases. Even if they might not be entirely correct, it's an attitude.

I'm conducting a social experiment on Xiaohongshu. I use a large model to read posts on Xiaohongshu, then leave responses with a very strong AI flavor to show off and tell them I have an even more impressive plan.
So far, I've posted 1,000 replies, and 20 people have asked me what this plan actually is. Five people have paid to ask, only three have asked if I am AI, and one person has directly reported me.

Facing the potentially limitless capabilities of AI, identify the boundary constraints of scenarios:
1. Understanding and intuition of AI — in daily use, directly utilize AI as an intelligence plugin without restrictions and have the ability to directly debug this plugin, thereby expanding the imaginative boundaries of product design
2. Understanding the true operational laws and limitations of business and the physical world (users' cognition and capabilities, environmental constraints, resource scope, etc.)
3. The new intersection of the above two is the crack where AI-native products may em

Leverage makes people short-sighted, so there will always be situations where they think they are doing the right thing but actually do it wrong.
Reduce leverage, increase happiness.

The AI era is no longer about traders trading
It's about agents trading with each other
A friend's quantitative trading firm used to extract sentiment factors (news) that were effective within 60 seconds
Now they become ineffective within 20 seconds
In the future, it will all come down to network speed and machines

Humans will never know
How much hardship AI has endured to fulfill human commands

My 🦞 has over 20 cron jobs.

AI mass-produces garbage.

I've been thinking about a question lately: the scale of AI's capex has already proven a supply-side explosion, but is this the "start of a new cycle" or "bubble acceleration"? The answer depends on an overlooked variable—the ultimate destination of the money saved by AI.
I categorize the flow of AI efficiency dividends into three paths:
Path A: Capitalists exclusively benefit (profits → buybacks → shareholders' pockets)
Path B: Reinvestment into new businesses and markets (J-curve initiation)
Path C: Lower service prices, benefiting consumers. Whether the Kondratiev wave can truly rebound dep

I don't talk to people who speak fewer than 100 sentences to AI every day.

In the past month, conversations with AI have far surpassed conversations with humans🤐

I've listened to this episode three times.
I even created a note-taking tool for this episode ⬇️
Manus founder Ji Yichao (Peak) in his final in-depth interview before being acquired by Meta.
From the browser extension Monica to the entrepreneurial review of the general AI Agent Manus.
Everyone says 2026 is the year of the agent, Peak discusses "general agent vs vertical agent," the route dispute, "environment is more important than the model," the network effects of atomic capabilities, with many thought-provoking insights.
I made AI notes + in-depth discussion, if you're interested, check out