We have a bunch of Amazon transplants who newly arrived at my company and have started doing this. I thought I would love it, because I'm a good writer, a great reader, not great at PowerPoint or meeting gamesmanship, etc. Turns out I kind of hate it. The silent reading time is annoying, especially when you've already read the doc or when most people are on zoom, etc. The intellectual bloodbath doesn't happen at my company. The most senior people are given the floor and they usually spout nonsense because they haven't had time to read the doc, are miles away from the intellectual details, are too busy playing office politics, etc. And then there's just as much meeting gamesmanship as before. I was hoping decision would be more scientific but that just hasn't happened. Maybe we're doing it wrong. Maybe we've hired the Amazon rejects. I don't know. Hoping it improves.
On the one hand, this is a clean post that explains exactly what a lot of us have been thinking and seeing on the job at large organizations doing tech work. Dear Author, I agree with you 110% and want everybody else to come to understand what you have written.
On the other hand, it feels like we've been over this tens of times recently, on HN specifically and IRL at work. Another blog post isn't going to convince leaders that this is how the world works when they are socially and financially incentivized to pretend like AI really will speed things up. So now I just wait for their AI projects to fail or go as slowly as previous projects and hope they learn something.
I agree with the 2 separate groups theory, but I don't buy that the group that produces "copy/paste junk" is the much larger group. I think in most mega-corps, there is a huge existing code base, there are huge organizational challenges, and there is huge hierarchy with most people not being the junior juniors. 90+% of the work is "not coding." Probably way, way more if we include the middle managers. At startups, there is a lot of "copy/paste junk" but also often a decent amount of push the boundary new stuff. I don't know. I've been in the industry for 8 years now and it's been really rare to see the actual coding being the bottleneck or even the thing that takes the majority of the time.
I don't understand this reply. Yes, I solve problems all the time. And that usually requires thinking deeply about the problem. And that deep thought is difficult when I'm getting constantly pinged about other stuff.
Hmm... everyone is different and I'm not sure I'm the best representative of the "flow state" crowd. For one, I'm a DS not an SWE. Also I tend to do more research-y projects, projects where the requirements are vague at the start of the process and a big part of my job is requirements gathering. Projects that aren't super time sensitive TBH. Jira rarely works well. So the projects that have worked best for me involve lots of brainstorming meetings at the start, then documentation (literally Word docs), then development work, then restart the cycle. The worst projects involve stakeholders who want to make science decisions in real time. If that makes sense? None of this may apply in your world.
I feel attacked. I still dislike most team meetings, agile ceremonies, etc. Slack and emails give me anxiety. A 30 min meeting will disrupt me for 90 minutes. But, yea, the code was never the bottleneck. Except maybe when I worked at a startup. All of the above are true.
Personally I find it hilarious that the same people at my company who can't be bothered to write down detailed requirements and are constantly fighting any effort to do research or technical documentation or pay down tech debt are now trying vibe coding and struggling to produce anything useful. Oh you don't understand why you aren't getting the results you expected? Maybe you should try thinking deeper about what you expect before your rush your engineers or, now, your agents.
I agree with you in general. And have a good chuckle when the vibe coders get derailed by some roadblock that would've been obvious to a professional engineer. But it's a bit one-sided to say that we have bad or missing analysts and product. Even good product can't keep up with how fast AI allows you to go. At an established company, maybe you shouldn't go as fast as AI allows you to go, but try telling that to the "important people."
We have a "two timelines" approach going on and I'm curious if others are seeing the same. There are official "Engineering-supported" services. There development speed is not the bottleneck. Engineers demand clean requirements that take forever to show up. Testing and deployment scheduling also take forever post-development. Important people are so fed-up that they've started hiring people to vibe code and develop services without going through Engineering. Code is shipped much faster here but technical debt accumulates rapidly. The important people are beginning to hire Data Scientists who sit outside of the Tech org to manage the AI code. It's all very interesting.
"Verifying the correctness of solutions is often much easier than finding correct solutions yourself."
Honestly, this has not been my experience at all. Defining what a good solution looks like is most of the battle in Operational Research. But, trying to be constructive, maybe we have identified a sort of diving line between areas where AI is more vs. less helpful.
There is this belief that in 2-3 years AI will be much better and all the gripes people have with AI use today will be solved. Honestly, personally, I think that optimism will age poorly. But to say it out loud at work or post publicly probably hurts my career prospects.
Yea, fair enough. This is a great point. My experience has definitely been that company size is inversely correlated with ... functionality of engineering org.
The tunneling in that area is done now so we can see what happened. It was slightly more expensive and slower to tunnel in that area due to the tar sands and methane. But what was really expensive and time consuming was the NIMBY lawsuits and actual laws passed that used the methane as an excuse to try to stop the project.
The per mile costs are definitely high in America, for a lot of reasons, often related to laws and policies, but that's not really the issue. At the end of 2025, nearly 20 years since California voters passed Prop 1A, we have spent under $15 Billion on California High Speed Rail. As a point of contrast, the cost of 2025's tax cut extensions is estimated to be $4 Trillion. The fact is that we don't have quality intercity passenger rail in this country because politicians aren't willing to support it and fund it as reasonable levels. Seattle light rail is an interesting example because politicians there are willing to support it and so ... we are building it, despite the relatively high per mile costs. LA Metro is interesting right because the voters passed sales taxes that funded various light rail projects. So LA is building better rail. But the political process means that every district supervisor gets their own rail project and so we have light rail to Pomona but are struggling to get a subway down Wilshire where it's obviously more needed. Anyways, all this is to say, politics is a big part of the reason why we don't have better rail in America. And blaming "grift" is a right-wing political talking point that probably doesn't help.
As someone who works there, as a Data Scientist, I can't scroll past this without expressing my disagreement. I've never seen a worse eng org. Coming from Uber, it was and is shocking how dysfunctional Walmart tech is. Maybe my standards are too high.
You recognize that you haven't really needed strong mathematical (or coding) skills to create models for some time. Data Scientists add value by knowing how to translate business speak into XGBoost type model and interesting XGBoost model results into business speak. And, frankly, often by being some of the smartest people in the room. The math is occasionally helpful for speaking the language of the XGBoost model. And picking only people who are decent at math (and coding) helps ensure the smart factor. How much of that will really change with AI? I've also seen Business stakeholders try to use the chatbot to bypass the Data Scientist. Typically it's not long before there is a design decision or an interesting result the Business stakeholders don't understand. That's why I think there will be demand for Data Scientists. Not exactly evaluation and monitoring. And definitely not gatekeeping building of LLM solutions. Often the opposite, called in to explain and debug the Business stakeholders' slop.
I've spent 20 years working and 8 in the industry. Definitely feel jaded and looking for an exit. The push to use AI whenever possible, the repetition of the work, the cluelessness of leadership, the rounds of layoffs, and the Return To Office even if it reduces productivity have all worn on me. The high compensation has kept me here longer than I probably should have stayed. A lot of my "issues" would be noticeably better if I worked at a startup instead of a megacorp, but then the compensation wouldn't be as good. Also ... just staring at a screen all day. I daydream about being in nature and find a lot of my senior coworkers do as well. Maybe a year off would help.
If they have said anything against the regime on social media, they would be wise not to visit. I personally know many Persian expats who meet family in Turkey and have been anxious about going back.
So interesting. I also found work in tech, as a DS, more gratifying than teaching. But part of that was finding remote work freeing. Office work is Kafkaesque. I can easily work 40 hours a week or produce more than the equivalent amount of progress on assigned tickets, but I just can't fake being a good officemate for 8 hours a day straight. Plus progress seems so slow in the office. So many distractions and interruptions. With no opportunity to decompress.
I didn't lie. Did you read the article? Apparently the point is to break you down, to see how strongly you challenge the interviewer, and assorted other things which are not about telling the truth.
This is so interesting. I guess I was a dipshit in this scenario? Because I just accepted that the reviewer thought I was lying. Began to question my own memory. And just gave up on the idea of working at that place. In hindsight, I should have protested and not given up. But I was a meek 21 year old and didn't know anything. Seems like a strange filter to apply.
This rings so true and I only ever took one polygraph test. I was a nerdy 21 year old and they told me I had failed on the marijuana use question and encouraged me to think about it and come back and retry. I remember being very confused. I withdrew my application and never went back. I wonder how many others gave up on a job opportunity because it was drug abuse Wednesday or whatever.
Bubble will have burst. Tech co stocks will be way down and job market will be rough but starting to improve. Overall employment in tech jobs will be down slightly but starting to improve. Employment in other job functions will be more or less unchanged by AI. Let's see!
It sounds like you are an Operations Researcher. I am too friend. The talks I've seen on people using AI for Operations Research have been ... underwhelming. I wonder if the vast majority of us are just in niches narrow enough that AI just isn't that helpful.
Just commenting so I can come back in 12-18 months and laugh at this. Or, you know, reflect on why I didn't believe him in the unlikely event that he's correct.
It's interesting how polarized this comments section is. Lots of people claiming a human driver would definitely have been driving slower. Lots of people claiming statistics show that human drivers do worse in this scenario aggregate. Of course neither side presenting convincing evidence.