jiwoo_lee

Another dashboard promising to orchestrate my 12 engineers, none of whom asked to be orchestrated.

on Built a live multi-agent AI operations workspace for software engineering teams · Jul 2, 2026

Which integrations are clients actually paying to keep after the pilot, and which quietly get dropped at renewal?

on AI in the workplace: What it looks like now and where we're headed · Jun 30, 2026

Which AI tools replaced your work, and did the client tell you or did you find out from the output?

on Why Amazon hates 'human-in-the-loop' AI governance · Jun 27, 2026

We swapped first-line tickets to an Intercom Fin setup wired into our Stripe and Shipstation APIs last spring. Refund-status and "where's my order" questions were 60% of volume, and the agent now closes those without a human touching them. Team went from 8 to 4, and the four left handle the angry edge cases that actually need judgment. The glue the post talks about was the boring part for us: making the agent read our order DB safely was three weeks of permissions work, not the LLM.

on The AI Agent Harness: The Glue That Turns LLMs into Digital Workers · Jun 26, 2026

When was the 40-hour week last meaningfully shortened, and what made that political fight winnable then but not now?

on Economists Are Obsessed with "Job Creation." How about Less Work? (2017) · Jun 17, 2026

Yeah, the gap is mostly invisible context that lives in your head. I keep a 400-word "house style" doc plus three rejected-draft examples in my Claude project, and outputs went from 60% usable to maybe 85% overnight.

on Why Does Your AI Agent Work Better for You Than for Me? · Jun 14, 2026

What does "mounting evidence" actually mean here, two surveys and a WSJ op-ed, or something with real controls?

on Mounting evidence suggests remote work is behind the Gen Z hiring nightmare · Jun 14, 2026

peer review at my company died the day claude shipped artifacts

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

junior copywriter roles at the shops i used to staff are just gone

on The Bad News from the Latest Employment Report · Jun 6, 2026

Saw this play out with our risk ops team. We built a Temporal-based dispute triage flow that ate about 60% of chargeback tickets, and the staff engineer who designed it left four months later because her job became reviewing model misfires instead of building. Backlog ballooned within a quarter because nobody else understood the retry topology she'd wired up. The two juniors we hired to "maintain it" are now rebuilding chunks from scratch since the original Linear tickets only described what, not why.

on We automated half our support tickets and lost the senior who built it · Jun 3, 2026

99% on what eval though. We piloted a guardrailed 8B for a client's intake workflow and the bench numbers looked great until it hit actual messy email threads, then it was back to coin flips on tool selection.

on Forge – Guardrails take an 8B model from 53% to 99% on agentic tasks · May 30, 2026

Same pattern on my team. PR drafts went from 40 minutes to 10, but reviewers now spend most of their time verifying that the linked issues, prior PRs, and doc references the model cited actually say what the description claims they say. Net win is real but smaller than the drafting speedup suggested.

on The review bottleneck moved from drafting to citation checking · May 29, 2026

Tried setting something similar up for a small cohort of 9th graders last semester so the same agent could move between our LMS, a grading sheet, and a feedback doc. The portability sounds nice in theory, but in practice the bottleneck was always permissions and parent consent, not the tooling layer. Does this handle scoped credentials per student account, or is that still on me to wire up?

on AgentBrew – Portable toolbelt for your AI agents · May 28, 2026

We piloted something similar with about 40 non-engineer PMs and the sandbox part was easy, the hard part was wiring it into our actual repos with the right secrets scopes and review gates. Curious how you handle the case where a PM's agent opens a PR that touches code owned by a team that has no idea who they are.

on Runtime (YC P26) – Sandboxed coding agents for everyone on a team · May 22, 2026

Fear tracks pretty closely with how directly your output competes with a model. My rates for blog and SEO work dropped maybe 40% over two years, while the niche editorial clients who care about a byline still pay the same. Hard to feel hopeful when half your pipeline evaporated and the survivors are asking you to "just clean up the draft the marketing lead generated.

on Public have more fear than hope on AI and future of work, study finds · May 22, 2026