Asking dumb questions has become one of my moves in interviews. This is to expose LLM users, as LLMs are extremely unlikely to answer a "how do I do X" question with "you shouldn't do X", or "how would you replace Y" with "you should keep using Y".
I guess now I have an additional reason to ask, so that I can screen out the assholes that will say it's a dumb question...
This sort of question works well for assessing senior / lead skills.
For example, when asked how to go about replacing PostgreSQL with MongoDB, it'd be okay for someone fairly junior to either give a straightforward answer or to turn their nose up and say "eww, no" - but you'd expect someone with more seniority to instead flag that it was an unusual scenario and ask for more detail.
Prepare a vaguely-plausible reason (eg. "we're storing JSON documents and the team are more comfortable with Mongo's tooling"), and this can form the basis for a productive discussion which you can drive in whichever social or technical direction you feel is most appropriate.
I was not aware of this. (We haven’t been interviewing recently.) Honest question, is it not possible to see the candidate’s eyes scanning the text? In video meetings I notice this when someone is reading.
Fun fact, there are multiple pieces of software that will real-time correct your eye tracking. Useful for virtual all hands or talks to fix teleprompter eyes.
Last time I said "you shouldn't do X" in an in person interview, I failed and the feedback was "candidate was not cooperative with the interviewer".
So even if the candidate is not using an LLM, the interview process is so messed up now that people willingly go along otherwise they may be rejected for, you know, actually stating what their experience was like.
Funnily enough, that was a YC funded company that I was interviewing for, and I expected a lot better.
Not sure how your delivery was, but maybe the best way is, “If I were doing it that way, here’s how I’d go about it. However, I might consider this other way and here’s why.”
In all honesty, my words were something along the lines of "I don't think that approach is a good idea. It's because of X reason."
Another viewpoint I've seen when sharing this story is that the interviewer wants me to start with the "bad" implementation first, then we move on to the "proper" one, but what's the point if I actually have pointed out the flaws beforehand?
That's very disappointing. As an interviewer the ideal situation I want in an interview is one where we are discussing tradeoffs. It allows the interviewee to show a deep understanding of the solutions we are discussing and, depending on the problem, even share insights around business impact.
I wonder if the pushback is more of a common reaction from more junior interviewers or once that aren't very empowered and just need to go down a script and worry about getting in trouble if they deviate.
Calling someone a "disagreeable asshole" just for saying that something is a dumb question is a bit strong. So he went out for challenging management on an internal document and he was 100% correct according to that tweet.
Seems to me that the actual assholes are still employed in that company. Perhaps he was just hired to fix the mess. Once the job was done, he was no longer needed.
Telling the interviewer that his first question is stupid is a sign of a disagreeable asshole.
Now granted, you can push back about questions irrelevant to the job but asking candidate for a data engineering role a MySQL question is fair game.
And even if it were irrelevant you’d want to frame it better. I’ve pushed back twice. Once the interviewer apologized (you’re right that’s not really relevant to the role) the other time the interviewer was playing tech bingo and we simultaneously invalidated each other.
However inappropriate the reaction to the SQL question may be, tweeting in a thinly anonymized manner about the incident is worse or perhaps even illegal.
As for management, perhaps we should coin the term "agreeable assholes" as a neologism for Machiavellian.
"It was very benign and was probably the most practical question I asked him! Let's say you want to randomly sample 1,000 users out of a table with a user_id column, e.g. so you can get a list of IDs to send a customer feedback form to. How would you do that?"
To me, that's not a dumb question. I mean, there is a simple answer but there are also complicated answers (especially if your table size is large, and performance matters):
1. SELECT user_id FROM table LIMIT 1000 -- this gives unordered results, but is not truly random (it's dependent on database retrieval patterns)
2. SELECT user_id FROM table ORDER BY RANDOM() LIMIT 1000 -- this is uniform random, but it's expensive for big tables.
3. SELECT user_id USING SAMPLE reservoir(1000 ROWS) -- this is a fast and memory-efficient streaming method, i.e. Reservoir Sampling [1] in which n is treated as large and unknown (DuckDB supports this sampling method and others: https://duckdb.org/docs/stable/sql/samples.html) Reservoir sampling (https://en.wikipedia.org/wiki/Reservoir_sampling) was a hot topic among data scientists on Quora circa 2011.
Sometimes simple questions have deep implications.
Could it be construed as a dumb question because the details depend on the specific engine you're using? If you have TABLESAMPLE, then the question is trivial. The question doesn't really test anything general other than "are you aware that your engine probably already has a feature for this"
It's not my <insert company name>. Has been my mantra lately. If I mention that a quarterly update cycle for mission critical software is too long in the context of security and the desktop enginners don't agree...I just repeat my mantra. I did recently develop a fix to a stupid bug a company said was working as expected. So I don't always stick to the mantra. But I'll disagree once. But then I commit to whatever management wants no matter how boneheaded. They're smart enough to see their dumb idea failed without me rubbing their noses in it.
> But then I commit to whatever management wants no matter how boneheaded.
This is always the right thing to do when you decided to sell yourself as a cog of a bigger machine. Some times, management is right and the know-it-all engineer isn't.
I'm surprised they didn't follow up with "why it that a dumb question?" It would likely lead to an interesting answer on its own, in the context of the interview.
Sure, on a case by case basis, it’s easy to make the argument that the smart asshole should’ve been more agreeable and tactful. But in the long term, keeping these managers hurts the company more and it’s how you end up with zombie orgs that scrape by.
I interviewed an individual with similar sounding experience/background. Said my first question was silly and insulting. I apologized but said our interview process is standardized, and please just entertain me. They proceeded to spend the rest of the interview giving the absolute worst answer I’d ever heard in any interview to any question. Did not make it past the initial interview.
One thing that you, and many others commenting on may be missing, is that we're getting an uncorroborated, retelling of a story that may well be simply reflecting the teller's own attitudes and reflections rather than a blow-by-blow account of what actually happened.
For example, say I was the interviewee and actually said of the SQL question: "Well, to be frank, that question isn't so good for understanding my SQL abilities." I don't think that would be rude or evidence of "being an asshole". A retelling of that story may simply be condensing it down to how that statement was interpreted: as a nice way to say, "that's a stupid question".
We shouldn't get too far into parsing what's in the story as a verbatim telling of what happened. We have the author's impressions of those events and their interpretations of the outcomes to drive home a larger point the author wants to make. That's it.
Honestly if I was interviewing, even if you told me the future and this was the outcome of hiring the guy, I still wouldn't hire him.
It's clear he's brilliant, and experienced, and seems like he's just refusing to "play the game". Speaking truth to power works if you know when to also NOT do it, or how to do it in a way that gets people to listen.
It doesn't really matter how brilliant a guy is; if he's not capable of "playing the game", then he's going to be (and WAS) out on his ass within a short timespan. That's a sink on my team's resources having to hire AGAIN and train AGAIN so short after his hiring. I don't want brilliant jerks, I'd rather have someone decent who can stick around by not getting himself fired for saying the wrong thing to the wrong person.
I feel similarly. However, not because I'd be worried about speaking truth to power but how they might treat peers and people more junior/with less power than them. Maybe they in fact treat less powerful people better and mostly punch up but while that exists it seems unlikely. Brilliant jerks can be incredibly effective but they can also destroy your entire team and now when they leave you are screwed.
Being kind and knowing how to deliver disagreement in a constructive way is important. Human beings aren’t perfect logical subjects immune from emotion (thankfully). Hopefully a good manager can take steps to address this or help mediate conversations where such a person might struggle. But places where “playing the game” (like withholding valid concerns with a decision) is a criteria for success are not long for this world
One danger I see with brilliant jerks is that they can also lead to others not sharing their legitimate concerns. If someone who is usually right calls you stupid frequently when you speak up or just make a suggestion you'll eventually stop making suggestions. Psychological safety is core to teams self-correcting. Of course you could also only hire people with very aggressive communication styles. I've seen this work but it obviously Congress with its own downsides.
I don’t disagree with this at all, psychological safety is important and abusive jerks don’t belong on any team. It’s also important to try to understand if someone is being intentionally rude and callous or simply being blunt/potentially being misunderstood.
Honestly, this whole story feels like a massive red flag for the company, not the data scientist.
They loved his attitude when it challenged broken systems, but the second he aimed that same energy at management's BS, he was out. Makes you wonder... is "culture fit" just code for "won't call out your boss"?
If I were on the board / an investor of the company, I’d let the managers go and hire back that engineer.
The goal of a company is to serve its customers and make a profit, not stroke managements egos.
And it sounds more like this guy pointed out what was wrong and maybe embarrassed people, but that doesn’t make him an asshole. It’s one thing to be mean or call people names, that’s an asshole. But this guy was an “asshole” because he was right and it embarrassed people? I’ve seen the type of management that is so insecure they can’t be pointed out what they are doing wrong, sounds more like what happened here.
A good company has management that has a thick skin and acknowledges its missteps and raises up talent. If you can’t admit an employee “lower” than you can point out your flaws and give you criticism, then you’re the problem.
It depends on so much that wasn’t covered here. If you tell me I’m wrong and show me why, I’ll thank you. If you tell me my idea is stupid and make it a point to maximally humiliate me for the mistake, I may have to bury your body in the desert.
It’s never a good idea to make someone feel extra embarrassed for an error. That’s how you get them to loathe you and to make them hide future mistakes from you.
Yes but once the guy is sacked for cause, or trumped up poor performance, bad attitude etc. the remaining narrative is crafted by the layer of management that remains.
Asking dumb questions has become one of my moves in interviews. This is to expose LLM users, as LLMs are extremely unlikely to answer a "how do I do X" question with "you shouldn't do X", or "how would you replace Y" with "you should keep using Y".
I guess now I have an additional reason to ask, so that I can screen out the assholes that will say it's a dumb question...
This sort of question works well for assessing senior / lead skills.
For example, when asked how to go about replacing PostgreSQL with MongoDB, it'd be okay for someone fairly junior to either give a straightforward answer or to turn their nose up and say "eww, no" - but you'd expect someone with more seniority to instead flag that it was an unusual scenario and ask for more detail.
Prepare a vaguely-plausible reason (eg. "we're storing JSON documents and the team are more comfortable with Mongo's tooling"), and this can form the basis for a productive discussion which you can drive in whichever social or technical direction you feel is most appropriate.
What interview format are you using that a candidate could be using an LLM during it?
Anything not in person?
There is software that sends video call transcripts to LLMs and shows you answers in real-time.
I was not aware of this. (We haven’t been interviewing recently.) Honest question, is it not possible to see the candidate’s eyes scanning the text? In video meetings I notice this when someone is reading.
Fun fact, there are multiple pieces of software that will real-time correct your eye tracking. Useful for virtual all hands or talks to fix teleprompter eyes.
Video call?
Last time I said "you shouldn't do X" in an in person interview, I failed and the feedback was "candidate was not cooperative with the interviewer".
So even if the candidate is not using an LLM, the interview process is so messed up now that people willingly go along otherwise they may be rejected for, you know, actually stating what their experience was like.
Funnily enough, that was a YC funded company that I was interviewing for, and I expected a lot better.
Not sure how your delivery was, but maybe the best way is, “If I were doing it that way, here’s how I’d go about it. However, I might consider this other way and here’s why.”
In all honesty, my words were something along the lines of "I don't think that approach is a good idea. It's because of X reason."
Another viewpoint I've seen when sharing this story is that the interviewer wants me to start with the "bad" implementation first, then we move on to the "proper" one, but what's the point if I actually have pointed out the flaws beforehand?
Sounds like you learned something valuable about them and dodged a bullet!
I won't disagree about it being messed up though.
That's very disappointing. As an interviewer the ideal situation I want in an interview is one where we are discussing tradeoffs. It allows the interviewee to show a deep understanding of the solutions we are discussing and, depending on the problem, even share insights around business impact.
I wonder if the pushback is more of a common reaction from more junior interviewers or once that aren't very empowered and just need to go down a script and worry about getting in trouble if they deviate.
Yeah it sounds like maybe OP was too direct. “Uncooperative” is strong language for feedback
I had a systems scenario recently where I was expected to troubleshoot a MySQL master that was trying to backup via rsync to a hung NFS mount.
I successfully resolved the problem and only then detailed how messed up their design was to get into such a situation in 2025.
Calling someone a "disagreeable asshole" just for saying that something is a dumb question is a bit strong. So he went out for challenging management on an internal document and he was 100% correct according to that tweet.
Seems to me that the actual assholes are still employed in that company. Perhaps he was just hired to fix the mess. Once the job was done, he was no longer needed.
You can be a disagreeable asshole even while being completely correct.
Being an “asshole” relates to inter-personal relationships, it has nothing to do with facts or correctness.
In the immortal words of The Dude: "You're not wrong, Walter, you're just an asshole."
found the disagreeable asshole
(I'm a disagreeable asshole)
I knew it, I'm surrounded by disagreeable assholes!
Telling the interviewer that his first question is stupid is a sign of a disagreeable asshole. Now granted, you can push back about questions irrelevant to the job but asking candidate for a data engineering role a MySQL question is fair game. And even if it were irrelevant you’d want to frame it better. I’ve pushed back twice. Once the interviewer apologized (you’re right that’s not really relevant to the role) the other time the interviewer was playing tech bingo and we simultaneously invalidated each other.
However inappropriate the reaction to the SQL question may be, tweeting in a thinly anonymized manner about the incident is worse or perhaps even illegal.
As for management, perhaps we should coin the term "agreeable assholes" as a neologism for Machiavellian.
For anyone who doesn’t want to have to register:
https://nitter.net/ryxcommar/status/1930985076965142918#m
The SQL question (mentioned in the thread) was:
To me, that's not a dumb question. I mean, there is a simple answer but there are also complicated answers (especially if your table size is large, and performance matters):1. SELECT user_id FROM table LIMIT 1000 -- this gives unordered results, but is not truly random (it's dependent on database retrieval patterns)
2. SELECT user_id FROM table ORDER BY RANDOM() LIMIT 1000 -- this is uniform random, but it's expensive for big tables.
3. SELECT user_id USING SAMPLE reservoir(1000 ROWS) -- this is a fast and memory-efficient streaming method, i.e. Reservoir Sampling [1] in which n is treated as large and unknown (DuckDB supports this sampling method and others: https://duckdb.org/docs/stable/sql/samples.html) Reservoir sampling (https://en.wikipedia.org/wiki/Reservoir_sampling) was a hot topic among data scientists on Quora circa 2011.
Sometimes simple questions have deep implications.
Could it be construed as a dumb question because the details depend on the specific engine you're using? If you have TABLESAMPLE, then the question is trivial. The question doesn't really test anything general other than "are you aware that your engine probably already has a feature for this"
It's not my <insert company name>. Has been my mantra lately. If I mention that a quarterly update cycle for mission critical software is too long in the context of security and the desktop enginners don't agree...I just repeat my mantra. I did recently develop a fix to a stupid bug a company said was working as expected. So I don't always stick to the mantra. But I'll disagree once. But then I commit to whatever management wants no matter how boneheaded. They're smart enough to see their dumb idea failed without me rubbing their noses in it.
> But then I commit to whatever management wants no matter how boneheaded.
This is always the right thing to do when you decided to sell yourself as a cog of a bigger machine. Some times, management is right and the know-it-all engineer isn't.
I'm surprised they didn't follow up with "why it that a dumb question?" It would likely lead to an interesting answer on its own, in the context of the interview.
Absolutely. And then regardless of the answer: "what is a good SQL question and what makes it better than the first one?"
Sounds like they fired the wrong person.
Sure, on a case by case basis, it’s easy to make the argument that the smart asshole should’ve been more agreeable and tactful. But in the long term, keeping these managers hurts the company more and it’s how you end up with zombie orgs that scrape by.
I interviewed an individual with similar sounding experience/background. Said my first question was silly and insulting. I apologized but said our interview process is standardized, and please just entertain me. They proceeded to spend the rest of the interview giving the absolute worst answer I’d ever heard in any interview to any question. Did not make it past the initial interview.
Jerks come in all sizes.
Maybe I’m not reading the post right but the dude sounds like a dick? What am I missing here?
The dicks are the management.
One thing that you, and many others commenting on may be missing, is that we're getting an uncorroborated, retelling of a story that may well be simply reflecting the teller's own attitudes and reflections rather than a blow-by-blow account of what actually happened.
For example, say I was the interviewee and actually said of the SQL question: "Well, to be frank, that question isn't so good for understanding my SQL abilities." I don't think that would be rude or evidence of "being an asshole". A retelling of that story may simply be condensing it down to how that statement was interpreted: as a nice way to say, "that's a stupid question".
We shouldn't get too far into parsing what's in the story as a verbatim telling of what happened. We have the author's impressions of those events and their interpretations of the outcomes to drive home a larger point the author wants to make. That's it.
You can read that, as a bad manager or inexperienced interviewee that asks stupid questions.
Moral of the story, you can be a star but still have to "work with" your managers.
Possible engineer was filtering the places he would want to work at.
Any place that fires me for calling bullshit — I would not want to work at anyway. Sayonara!
Or maybe... don't be a brilliant jerk
Or maybe... Don't be a jerk
But likely, it's hard to find an adequate management.
Some people interpret a "no" as "the guy saying no is a jerk". Don't be this kind.
Brilliance doesn’t balance out being a jerk. The net result is always negative.
Honestly if I was interviewing, even if you told me the future and this was the outcome of hiring the guy, I still wouldn't hire him.
It's clear he's brilliant, and experienced, and seems like he's just refusing to "play the game". Speaking truth to power works if you know when to also NOT do it, or how to do it in a way that gets people to listen.
It doesn't really matter how brilliant a guy is; if he's not capable of "playing the game", then he's going to be (and WAS) out on his ass within a short timespan. That's a sink on my team's resources having to hire AGAIN and train AGAIN so short after his hiring. I don't want brilliant jerks, I'd rather have someone decent who can stick around by not getting himself fired for saying the wrong thing to the wrong person.
I feel similarly. However, not because I'd be worried about speaking truth to power but how they might treat peers and people more junior/with less power than them. Maybe they in fact treat less powerful people better and mostly punch up but while that exists it seems unlikely. Brilliant jerks can be incredibly effective but they can also destroy your entire team and now when they leave you are screwed.
Being kind and knowing how to deliver disagreement in a constructive way is important. Human beings aren’t perfect logical subjects immune from emotion (thankfully). Hopefully a good manager can take steps to address this or help mediate conversations where such a person might struggle. But places where “playing the game” (like withholding valid concerns with a decision) is a criteria for success are not long for this world
One danger I see with brilliant jerks is that they can also lead to others not sharing their legitimate concerns. If someone who is usually right calls you stupid frequently when you speak up or just make a suggestion you'll eventually stop making suggestions. Psychological safety is core to teams self-correcting. Of course you could also only hire people with very aggressive communication styles. I've seen this work but it obviously Congress with its own downsides.
I don’t disagree with this at all, psychological safety is important and abusive jerks don’t belong on any team. It’s also important to try to understand if someone is being intentionally rude and callous or simply being blunt/potentially being misunderstood.
Share is name so half of HN can hire him!
I'm assuming his name is "T. Scottsman".
Honestly, this whole story feels like a massive red flag for the company, not the data scientist.
They loved his attitude when it challenged broken systems, but the second he aimed that same energy at management's BS, he was out. Makes you wonder... is "culture fit" just code for "won't call out your boss"?
If I were on the board / an investor of the company, I’d let the managers go and hire back that engineer.
The goal of a company is to serve its customers and make a profit, not stroke managements egos.
And it sounds more like this guy pointed out what was wrong and maybe embarrassed people, but that doesn’t make him an asshole. It’s one thing to be mean or call people names, that’s an asshole. But this guy was an “asshole” because he was right and it embarrassed people? I’ve seen the type of management that is so insecure they can’t be pointed out what they are doing wrong, sounds more like what happened here.
A good company has management that has a thick skin and acknowledges its missteps and raises up talent. If you can’t admit an employee “lower” than you can point out your flaws and give you criticism, then you’re the problem.
It depends on so much that wasn’t covered here. If you tell me I’m wrong and show me why, I’ll thank you. If you tell me my idea is stupid and make it a point to maximally humiliate me for the mistake, I may have to bury your body in the desert.
It’s never a good idea to make someone feel extra embarrassed for an error. That’s how you get them to loathe you and to make them hide future mistakes from you.
Yes but once the guy is sacked for cause, or trumped up poor performance, bad attitude etc. the remaining narrative is crafted by the layer of management that remains.