A Veteran Told Her To Ask Claude. She'd Tried Twice Already.
Blogger Yael Grauer called a senior engineer for guidance on a genuinely unsettled question. His answer: "Honestly? Ask Claude." She had, twice, and what she wanted wasn't a summary but the specific judgment thirty years builds.
Grauer's essay names the moment exactly: a mentor she'd built up over years waved her toward a chatbot instead of sharing the judgment he'd use himself. She says she'd already asked twice.
The 105-comment thread split between engineers who called the redirect disrespectful and others who defended it as an earned shortcut, until one open-source maintainer admitted Claude would likely nail 98% of his own advice anyway.
Yael Grauer wanted thirty years of judgment. She got pointed at a chatbot.
In the essay that hit 183 points on HN, writer Yael Grauer describes scheduling a call with a senior engineer to ask about a hard problem, one without industry consensus, hoping for the answer only three decades of hands-on experience could give. What she got instead: "Honestly? Ask Claude." It kept happening. People she trusted, when handed genuinely difficult questions she'd already wrestled with, redirected her to a chatbot rather than share their own read on the problem.
The essay's subtitle does the framing work: "I already did." Grauer is explicit that she isn't anti-LLM; she uses one routinely and had already run these exact questions through Claude before making the call. What "ask Claude" doesn't give her, she writes, is "the person's specific, lived experience." She compares it to asking a friend for a personal recommendation instead of a search ranking. You don't want consensus. You want their judgment, earned the hard way. She concedes there's a real cost to being "the person other people call," but argues that once a question has already survived a round of LLM consultation and still needs a human, waving someone off "just withholds the thoughtful answer decades of experience could have given."
The thread: an earned shortcut, or a professional cold shoulder
The 105-comment reply split along a predictable line: readers who'd felt exactly the sting Grauer describes, and engineers who defended the redirect as a fair filter against low-effort asks. One of the earliest, most-upvoted replies reframed the whole piece before the fight even started, noting Grauer isn't actually arguing against LLMs. "This seems to be a communication problem," an HN commenter with 55,000-plus karma wrote, pointing out that people redirecting her may not even register she'd already done the legwork.
Not everyone read it that charitably. A self-described 40-year veteran of programming and sysadmin work told a different version of the same story, from the other side of the redirect:
"When I run a coworkers question through the AI tooling and talk through the answer with them, it's because my 40 years of experience tells me that's the next step." — linsomniac, a 40-year programming and sysadmin veteran
The more striking concession came from further down the thread. The creator of the Python database toolkit SQLAlchemy described the current mood among engineers as exhausted, not dismissive, laying out a run of 2026 anxieties, layoffs, a squeezed economy, tools replacing colleagues, before landing on an admission that cuts against his own decades of hard-won expertise:
"Whatever advice I have, Claude is going to have 98% of it plus another 10% that I didn't even have." — zzzeek, creator of SQLAlchemy and Alembic
Then the thread produced the counter-evidence Grauer's essay lacks: a case where the redirect actively backfired. A commenter recounted asking a question on a work Slack, only to have a colleague forward it to Claude and paste back the response as the answer.
"15 minutes later a different coworker responded pointing out that everything Claude said was wrong." — queenkjuul, describing a Slack thread where the AI answer was checked and failed
At the far end of the split sat the bluntest read of all, from a well-known figure in software testing circles, author of Lessons Learned in Software Testing:
"If I am asked a question, and I refer it to an LLM, that's because I have no respect for you." — James Bach
The last laugh landed on Grauer herself. Within hours, a reader posting from a barely-used, newly created account turned the essay's own sentence rhythm into ammunition against it: "You clearly used AI pretty heavily to write this - so whats the point?"
Nobody in the thread produced evidence either way. But the accusation itself is the tell worth sitting with: a piece arguing that human judgment is irreplaceable got read, by multiple strangers within hours, as possibly not written by the human whose judgment it was defending.
Get breakdowns like this every day
The top Hacker News stories, unpacked — by email, Telegram, or RSS.
22 Billion UPI Payments a Month Move Through One Central Switch
Every UPI transfer in India, from a roadside tea stall to an online checkout, funnels through exactly one switch, run by a non-profit called NPCI. The money always leaves the payer's account before the payee's bank is even told to credit it, never the other way round.