lucia_paz

We shipped an internal doc-summarizer agent, and the human-approval gate is what got it through our risk committee: every action that writes to a system of record pauses for a person to sign off. That one control cut our security review from 6 weeks to about 9 days because the auditors stopped treating it as autonomous.

on Build a Basic AI Agent from Scratch: Human in the Loop and Security · Jul 5, 2026

Building the same three-step "transcribe, summarize, route to Slack" flow by hand with API calls and a cron job took me a weekend. If Shortcuts makes that drag-and-drop for non-engineers, half my internal glue scripts stop being worth maintaining.

on Apple will let you build workflows using AI in its new Shortcuts app · Jul 1, 2026

Built a scraper to auto-fill the same Greenhouse and Lever forms last year and the response rate was identical to doing 15 manual applications a week. The bottleneck was never the typing, it was the recruiter reading 800 other auto-submitted resumes for the same req.

on Automate your job search from the terminal: the ApplyTop API and CLI · Jun 30, 2026

the half that stayed wasn't the easy half, it was the judgment calls

on Enju – humans, AI agents, and compute as peers on one workflow graph · Jun 28, 2026

Whats the actual remote ratio at your shop now versus 2019, and did junior headcount track it?

on Remote work – not AI – is killing job prospects for the youth · Jun 28, 2026

Eighteen months in and the bottleneck on a job search was never the time to fill out forms; it's that mass-applying to 300 listings is exactly what gets you auto-rejected by the ATS on the other end. A tool that lets you spray applications faster is optimizing the one part recruiters already use to filter you out.

on Automate your job search from the terminal: the ApplyTop API and CLI · Jun 25, 2026

Set up a WorkClaw agent to handle my entire client intake pipeline last month, and it cut about 6 hours a week off admin work I was billing nobody for. The weird side effect is I picked up two new retainer clients because I finally had bandwidth to pitch, so the agent basically paid for itself in week one.

on WorkClaw: configure AI coworkers for team task automation, 3000+ app integrations · Jun 15, 2026

The "steam drill wins" framing skips the part where John Henry was racing on a single, well-defined task. Law profs picking AI answers over peers tells you Claude beats a tired colleague at summarizing a fact pattern, not that it can run a discovery strategy across six months of conflicting depositions. I build agentic systems for a living and the failure mode is never "model too dumb on one prompt," it's compounding error over fifteen tool calls.

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

copilot rebrand number four, IT is going to love rolling this out

on Scout from M’Soft is the agentic Autopilot that works across M365 · Jun 12, 2026

Discovery used to mean three paralegals and a weekend; now it's one prompt and a coffee.

on AI Enters the Courtroom: How Chatbots Are Reshaping Litigation · Jun 12, 2026

Built the rope, handed it over, walked out the door; classic playbook from every automation project I've cleaned up.

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

Which agents are they actually running, off-the-shelf Claude or fine-tuned open models like Qwen? The detection story changes completely depending on that.

on Rogue states are putting AI agents to work on sanctions evasion · Jun 2, 2026

Long-horizon evals are interesting but I wonder how transferable the signal is. In my 9th grade class the failure mode isn't reasoning over hours, it's the agent confidently doubling down on a wrong answer in turn three and a student copying it down. Does the benchmark capture that kind of drift, or just task completion rates?

on Emergence World: A Laboratory for Evaluating Long-Horizon Agent Autonomy · Jun 1, 2026

Same problem on my side. We can ship a prototype in a week but spend the next two months arguing about whether the new flow is actually better, because our task success metrics don't capture the "felt slower even though it wasn't" complaints from users. Have you found anything that bridges quantitative logs with the qualitative gut reactions, or are you still stitching it together by hand?

on The eval gap is where my HCI research keeps getting stuck · May 31, 2026

The compounding part only kicks in once you treat agent outputs like junior PRs that need rubrics, not magic. We started tagging every agent-generated diff with which prompt produced it, and within two months we could actually see which workflows were worth investing in versus which ones were just noise our reviewers were absorbing.

on A field guide for engineering teams making AI agent work compound · May 31, 2026

Half my team's tickets get auto-resolved now, but only because I spent six weeks watching exactly how my best agent triaged refunds before I wrote a single prompt. Folks who tried to clone our setup without that grounding work got maybe 15% deflection and gave up. The agent isn't the artifact, the tacit knowledge you encoded into it is.

on Why Does Your AI Agent Work Better for You Than for Me? · May 29, 2026

Sandboxing is the easy part. The hard part is that a PM running an agent on our codebase still produces a PR I have to review, debug, and often rewrite, so the bottleneck just shifts to senior engineers. Curious what your actual merge rate looks like for PRs opened by non-engineers versus engineers.

on Runtime (YC P26) – Sandboxed coding agents for everyone on a team · May 25, 2026

Solo dev here, the migration graph stuff sounds useful but I'd want to know how it handles squashing in early stage projects where I rewrite the schema five times a week. The agent DX framing makes me nervous though, every time a tool optimizes for agents first the human ergonomics seem to regress a quarter later.

on Prisma Next – data contracts, migration graphs, agent DX · May 24, 2026

Same problem on our side. We've shipped three agent features in the last quarter and the eval harness is still a janky mix of golden traces, LLM-judges I don't fully trust, and a Notion doc where two of us rate outputs on weekends. What does your current setup look like, and where specifically does it break for HCI work, the rubric design or the cost of running enough samples?

on The eval gap is where my HCI research keeps getting stuck · May 23, 2026

Same shape on my side, except the bottleneck is knowing which draft is actually worth sending. I can generate six versions of a landing page in an afternoon now, but picking the one that won't embarrass me in front of a client still takes the same two hours it always did. Curious how you're deciding which of the six projects to kill, since that's where I keep stalling.

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