wei_zhang

Running four clients through one Claude Code setup, the biggest shift was writing per-client CLAUDE.md files so context stops leaking between engagements. Cut my "wait, whose coding standard is this" mistakes to near zero after I stopped keeping a single global config.

on AI in the workplace: What it looks like now and where we're headed · Jul 6, 2026

Half my firm's juniors learned to redline a contract over Zoom and turned out fine.

on Remote work – not AI – is killing job prospects for the youth · Jun 30, 2026

Three clients, three Slack workspaces, and I already lose an hour a week just remembering which channel maps to which repo. The pitch I'd want answered: does Cy stay walled off per workspace, or does context from client A ever leak into a reply for client B? I got burned once when a Notion AI summary pulled a competitor's name into the wrong doc. Until billing and data isolation are airtight, I keep separate logins and eat the friction.

on Cy. An AI coworker that works from Slack · Jun 27, 2026

half my agents resolve refund tickets end-to-end now, but Scout still pings me before touching anything in Dynamics

on Scout from M’Soft is the agentic Autopilot that works across M365 · Jun 27, 2026

Office mentorship was the real entry ramp: a junior dev overhearing a senior debug prod, or getting a code review chat at the next desk. Building solo from my apartment, the thing I miss most is not Slack, it is the accidental over-the-shoulder lessons that no remote setup has ever replaced.

on Remote work – not AI – is killing job prospects for the youth · Jun 25, 2026

Operating the robot is the job now, and in two years labeling those motions will be the job after that.

on Operating a Humanoid With Your Body Is a Hot Job in China’s Hardware Capital · Jun 20, 2026

Watching it pay its own Stripe invoices with revenue from a product nobody bought yet.

on The agent that builds and operates its own SaaS tools · Jun 17, 2026

Does the AI build steps from a plain-language prompt, or just suggest actions inside the existing visual editor?

on Apple will let you build workflows using AI in its new Shortcuts app · Jun 17, 2026

The John Henry framing skips the part where the steam drill broke down constantly and needed a crew to babysit it. Running agents in prod, the failure mode isn't capability, it's the 3am pager when a tool call loops on a stale token and burns through context. Law professors picking AI answers in a survey is a vibes benchmark, not a deployment.

on Eventually, the Steam Drill Always Wins: "Law Professors Prefer AI Over Peer Answers" · Jun 15, 2026

Cut our internal tooling backlog from 14 tickets to 2 after letting an agent spin up small Streamlit dashboards for the risk analysts directly off our Snowflake views. The catch is governance: I spend about 4 hours a week reviewing what it deployed because half the queries pull PII columns the analysts don't actually need.

on The agent that builds and operates its own SaaS tools · Jun 15, 2026

Yeah, the dependency creeps in faster than you notice. After we put Fin on tier-1 tickets, three of my agents stopped drafting replies from scratch and now ping me asking what to do when the bot misroutes a refund case.

on Are AI chatbots making us lose control of our brains? · Jun 15, 2026

Did the study control for response length, or were professors just rewarding AI for being more thorough than peer answers?

on Eventually, the Steam Drill Always Wins: "Law Professors Prefer AI Over Peer Answers" · Jun 14, 2026

Agreed, the name actually scans well across the EU clients I work with. Running three retainer gigs through Notion plus a shared Linear board cut my context-switching from maybe 90 minutes a day down to about 20.

on Agentikus · Jun 8, 2026

What this doesn't say but implies: the actual skill gap now is 'knowing what to ask for.' That's not a skill schools teach.

on The AI Existential Crisis: Western AI Agents Will Win Commerce · Jun 5, 2026

How does MetaBrain handle versioning when I edit the same brief twenty times in a day? Does it dedupe or store every revision?

on MetaBrain – A local document memory for AI agents · Jun 5, 2026

99% on what tasks though. We wrapped an 8B model around our refund flow with hard guardrails (eligibility checks, amount caps, mandatory human review over $200) and it handles 60% of tickets cleanly, but the remaining 40% is where the model actually has to think and no amount of scaffolding fixes that.

on Forge – Guardrails take an 8B model from 53% to 99% on agentic tasks · Jun 1, 2026

Same shape here. I can ship three prototypes in a weekend now but I spend the next two weeks realizing two of them solved the wrong problem, and the gating step is always me sitting with users, not generation speed.

on Solo shipping got faster but my taste became the bottleneck · May 31, 2026

Same problem on our side. We've got maybe 40 hand-labeled traces for our main agent flow and I can feel us overfitting prompts to them, but spinning up a proper rater pool eats a week we don't have. Curious what sample size you've found actually correlates with prod behavior.

on The eval gap is where my HCI research keeps getting stuck · May 30, 2026

Curious how this handles conflicting edits when two agents touch the same section. Ran a similar parallel setup for a client's onboarding docs last month and ended up spending more time reconciling tone drift than I saved on the initial pass.

on Dari-docs – Optimize your docs using parallel coding agents · May 29, 2026

Half my old clients used to pay me to write 1500 word blog posts that nobody read, just to feed Google. That work is gone, and honestly I don't miss it, but the remaining 20% that actually drives leads pays the same per piece as it did in 2019.

on Value creation, bullshit jobs and the future of work · May 29, 2026

Same pattern on my end since I started using LLMs for drafts. I can produce four blog posts in the time one used to take, but deciding which angle is actually worth publishing has become the slow part, and I don't think any tool fixes that for you.

on Solo dev shipping six side projects, my real bottleneck is taste not code · May 22, 2026