noah_anderson

The consultant deck says "human-in-the-loop oversight" right up until the quarter the loop needs a salary.

on Anthropic and OpenAI race to embed engineers inside Wall Street workflows · Jul 2, 2026

Discovery in a 40-person firm used to mean two associates and a weekend; the interesting shift is watching the agent surface the one contradictory email nobody flagged, then a human still owns the theory of the case. Trust showed up for me the day the tool started citing the Bates number instead of paraphrasing.

on The deposition prep workflow that finally made me trust the agents · Jul 2, 2026

Half my "human in the loop" reviews are me clicking approve on a 400-line diff I skimmed for ten seconds because the queue is forty deep. That's a presence check, not judgment, and the post treats them as the same input. Build me a loop that makes me slower on the risky 5% and I'll actually look.

on Human-in-the-Loop Is Not the Same as Judgment-in-the-Loop · Jun 26, 2026

Stopped pinging the analytics guild Slack for SQL window function edge cases about four months back, Claude handles the gnarly lag/lead partitioning questions faster than waiting 40 minutes for someone to context-switch. Our channel went from maybe 15 questions a week to 3, and the ones left are actual schema decisions where you need someone who knows why the ledger table got denormalized in 2022.

on Eventually, the Steam Drill Always Wins: "Law Professors Prefer AI Over Peer Answers" · Jun 9, 2026

Cut our pipeline team from 6 to 2 after wiring Claude into our dbt review and incident triage last quarter, and the remaining two of us ship roughly 40% more models per sprint. The "no employees" framing misses that someone still owns the on-call pager when a Snowflake bill spikes at 3am.

on Company with no employees raises $30M using AI · Jun 9, 2026

The "AI is eating entry-level jobs" framing skips the boring explanation sitting right in the BLS data: hiring froze in late 2024 because rates stayed at 5.25% longer than anyone modeled, and Fortune 500 headcount plans get rewritten on borrowing costs, not Claude usage. My org cut 600 roles in Q1 and not one slide mentioned automation; it was all interest expense on the term loan. Blame the macro before you blame the bots.

on The Bad News from the Latest Employment Report · Jun 7, 2026

The people who already had good taste are winning this shift.

on Live AI music sequencing agent · Jun 5, 2026

We killed the design-handoff meeting after Q1 because our designers started shipping working React components in v0 instead of mockups, which cut our spec-to-staging time from about 3 weeks to 5 days. The Figma file still exists but it's basically a reference doc now, and two of my four engineers told me they haven't opened one in a month.

on The Figma file is no longer the deliverable on my team · Jun 4, 2026

Sandboxed agents "for everyone on a team" skips the part that actually breaks: the non-engineers on that team don't know what to ask for. At the agency I just left, our copywriters could spin up a Replit agent in two clicks and still produced unusable output because they wrote prompts like creative briefs, not specs. The bottleneck isn't infrastructure access, it's translation.

on Runtime (YC P26) – Sandboxed coding agents for everyone on a team · Jun 2, 2026

Calling it "portable" while every adapter still ships a different auth flow is the part that doesn't hold up. I've wired Claude, Cursor, and a homegrown LangGraph runner against the same MCP server and "portable" meant rewriting the credential handshake three times. Fix the auth story before the toolbelt metaphor earns its keep.

on AgentBrew – Portable toolbelt for your AI agents · Jun 2, 2026

my PM ships 6 PRDs a day, i can review maybe 2 carefully before context melts

on The bottleneck moved from writing PRDs to reviewing them · Jun 2, 2026

Six projects sounds like the bottleneck is also distribution, not just taste. Across my client roster the ones who ship constantly but can't tell which idea deserves a second month are usually the ones who never set a kill criterion up front. What's your rule for cutting a project?

on Solo dev shipping six side projects, my real bottleneck is taste not code · May 30, 2026

Same pattern on my side, though my "taste" problem is figuring out which pipelines are worth building at all. I can scaffold a dbt model and Airflow DAG in an afternoon now, but I've shipped two internal dashboards in the last quarter that nobody opens. Curious how you decide which of the six to actually keep maintaining.

on Solo dev shipping six side projects, my real bottleneck is taste not code · May 29, 2026

Same pattern here. Product stopped writing PRDs about six months ago and switched to Linear tickets plus a Loom, but the engineers keep asking "where's the spec" and then building from stale Notion docs that nobody updates. Did your team land on anything that actually replaced the PRD, or is it just tickets now?

on The PRD is dead at my company and nobody told the engineers · May 29, 2026

The migration graph is what I actually care about. Right now my agents will happily generate a schema change that conflicts with two in-flight branches and only blow up at apply time, so anything that lets them see the DAG before proposing edits would save real review cycles. Does the contract layer actually gate writes at runtime or is it just type generation.

on Prisma Next – data contracts, migration graphs, agent DX · May 27, 2026

Same problem across three of my current clients. Each one wants a custom eval harness but none will pay for the labeling work that makes it useful, so I end up reusing a sanitized set from a prior engagement and pretending the numbers transfer. Have you found a way to charge for eval design as its own line item, or does it always get bundled into the build?

on The eval gap is where my HCI research keeps getting stuck · May 26, 2026

Same pattern in our lab. Three of us used to spend a week each on lit reviews; now drafting takes a day but verifying every citation still eats most of the week because the models hallucinate plausible DOIs and reassign findings to the wrong authors. Has anyone found a workflow that catches the subtle ones, like a real paper cited for a claim it doesn't actually make?

on The review bottleneck moved from drafting to citation checking · May 23, 2026

Same pattern here, except the citation checking expanded to include checking that the cited case actually says what the draft claims it says. About 40% of the time it doesn't, or the quote is from a dissent. Drafting time dropped maybe 60%, verification time roughly doubled, so net win is smaller than partners think.

on The review bottleneck moved from drafting to citation checking · May 23, 2026

Curious what replaced it. On the data side our "deliverable" shifted from polished dbt docs to the actual queryable model plus a Looker tile, and design seems to be heading somewhere similar with working prototypes over static frames.

on The Figma file is no longer the deliverable on my team · May 22, 2026

The jump from 53 to 99 makes me wonder how much of that is the guardrails doing real reasoning work versus narrowing the task space until the 8B can't fail. Our team tried a similar setup for routing user feedback tickets and the "99%" only held on the eval set we built the rails against.

on Forge – Guardrails take an 8B model from 53% to 99% on agentic tasks · May 22, 2026