Wednesday 9am, bookmarks full, work not started
Wednesday 9am, sitting in a coffee shop, I scrolled through a morning report. Bookmark folder gained 12 items. Nothing actually got done.
I turn the report into three columns: follow, ignore, do today
This Paizaobao issue dropped Honor Robot Phone, Google Pixel 11, Insta360 X6, and SpaceXAI Grok Bot all at once. My move: hand it to a ready-made AI chat tool and split it into three columns — "useful to whom, can it be verified, worth acting on today" — instead of letting it recap the news. My friend Lin Ran already runs this way. He does product photography on the side in Hangzhou; last Wednesday he fed the report to AI while shooting for a client and kept only one item — "shoot a vertical sample clip." I got stuck too. I asked AI to write me a piece that ended up reading like "the industry is exciting." Looked complete. Not a single executable action in it.
Replicate today: $0 start, 15 minutes, no code
Money: I started at $0, using the free tier of an AI chat tool I already had. If there's no free tier, I'd budget 20–50 RMB/month. Time: about 15 minutes. Technical barrier: copy-paste, hit send, save the result — no API keys to configure. First step: I click the paperclip or "upload" button next to the chat box, paste in the full morning report, then type "output only three columns — follow, ignore, do today — and explain why." This tool isn't for everyone. Skipping it for now is also fine.
By stage, keep only what fits
Just starting out: When I had no steady clients, I'd try it for a week — keep one shippable action per day, don't chase every launch. 1–2 clients: If client demand was steady, I'd turn new features into a small sample or a question and test it on them — no gear-buying. Scaling up: If I'm moving from gigs to repeatable delivery, I'd have AI collect recurring questions each week, then decide whether to build templates, tutorials, or automation — don't let the news itself become the job.