They don't proffer a reason that makes sense because they don't really need to. They make the rules. They can BS their way through a question at the All Hands meeting or on CNBC. It's pretty apparent that at many companies the real reason is soft layoffs but you won't hear the executives say this part out loud.
I agree. Just goes to show how difficult it is to have lasting, fundamental societal change.
I will say that zoom fatigue is real and remote work can be problematic for people that are mainly in meetings. But we could have solved this with a better solution than snapping back to office culture.
Exactly this. I've worked at a few places now where there were RTO policies that many workers ignored. But periodically low performers and/or people that had lost political battles were denied promotions or fired and their poor office attendance record was cited.
Loved this comment. I'm 46 and dislike my job at a mega corp, especially as compared to my previous startup job. There are way too many cooks in the kitchen on every halfway interesting project. But the mega corp job pays too much to leave. Do this for a few more years and I'll reach financial independence. Then I can go back to a startup or ... something else. Part of me wishes for the layoff and severance that more and more of my colleagues are getting. Sorry to be such a downer. Most days are enjoyable and I can tune out the mega corp nonsense.
It's interesting how many commenters are defending PIPs. To me it shows an inability to think beyond the ways in which many tech co's operate today. There are so many better ways to deal with an employee who is failing to meet expectations, both for the employee and the company. Giving people more feedback, more frequently can help. Giving people flexibility to switch teams or even roles can help. Better interviewing before hiring can help. Better financial and role planning can help, especially at startups. Retraining can help. Letting people go with a decent severance is often better for everyone involved. PIPs are pretty clearly sub-optimal.
I work in Tech at Walmart and the funny thing is that of all the places I've worked they do the least in terms of surveying employee satisfaction. If they just asked, I and many of my colleagues would be happy to tell them which people and policies are toxic and which are wonderful. There's broad agreement in the rank and file on much of this. There are some truly awful and some truly wonderful things about the way Walmart manages tech. When I mentioned this to my VP, the response was basically that Walmart top level management didn't want to know some things because they might be compelled to act. Maybe all this "AI" rigamarole is just a way for Walmart top level management to get information without being legally or morally obligated to do anything?
There are a lot of posts claiming that advances in technology over the past decade or two have been leading to "productivity gains". But the economic data simply doesn't support these claims. Productivity growth has stalled in most of the developed world.
I like the way tech companies operate. I like slack. We don't use P2 but it looks interesting. I just don't really believe it'll turbo charge productivity at my company, much less at a more traditional company.
I'm going to avoid saying anything one way or the other about AI because that would be a separate argument.
Thank you for writing this! I often feel the same way.
My relationship with Slack kind of summarizes things neatly. I enjoy using Slack and it is far and away my preferred method of communication with work colleagues and even work friends. At the same time, seeing slack messages or even the number of slack messages waiting for me is anxiety inducing. There is a false sense of urgency. Zoom is worse, on basically all counts. But Slack is a great idea and tool that is somehow awful for my mental health. I don't know what the solution is. I guess unplug and take a walk. No one has ever complained that I don't reply to a Slack message fast enough.
there have been a lot of little prizes and awards to collect at some places i've worked. sometimes people had to interview for those awards. and i always wondered how much unnecessary damage we did to the morale of the people who didn't win the awards.
and i'm big on fairness. it's a challenge to make sure Performance Reviews are as fair as possible, especially with larger teams. you think it's obvious who is a good worker and who is a bad worker right up until you have to manage a large team.
sorry, you lost me at: "Reason 1: You Reward The Right People The Right Amount"
if that were true, i too would say that Performance Reviews matter. i'm 45 and it's literally never been true at any place i've worked.
"Without them, you’re letting managers just make things up." sadly with them you are also letting managers just make things up.
we pay so much attention to the hiring process, but i think Performance Reviews need a shake up. i've consistently seen the wrong people rewarded. you want to talk about people only hiring people who look and act like them? how about people only promoting people who look and act like them?
I bought a V60, one of the few remaining station wagons for sale in the US. And it is great for those with 1 or 2 kids, but, like the author implies, doesn't really work for those with 3 car seats. The only thing that has grown faster than the size of our cars is the size of our car seats.
I'm a bit lost in this terminology. But coming from the Operations Research perspective that gave Dynamic Programming its name, Dijkstra's Algorithm is very clearly Dynamic Programming. It's Forward Dynamic Programming as opposed to the much more common Backward Dynamic Programming, if that helps any.
What are you talking about? She has no qualifications and her publications are juvenile. Her h-index is 8. She has no background in CS and has a few publication in Foreign Affairs. If she is the leading academic, then AI safety is not a serious academic field. She is in no way comparable to Ilya. This is the worst take I've seen on HN in awhile.
Right place, right time. She was one of very few people "working" on AI Safety. I made another comment about this elsewhere on this post, but to me this debacle is a function of the fact that there is no there there when it comes to AI Safety "research".
It seems to me as if Toner was just in the right place at the right time in order to get a seat on the board of Open AI. She has an undergrad degree in chemical engineering and a master's in "security studies." Her work on AI Safety opened this door for her but seems, at least to me, to be ... superficial. I'm no expert but I have worked in tech for awhile and at a public policy think tank. So I guess my question is there a real scientific field of AI Safety? Are there any real experts? Are there any real insights? I dislike the idea of trusting giant tech companies with breakthrough technologies with minimal oversight / regulation. But it just seems like there is no real science regarding AI x public policy. Like the policy experts have no clue what they are talking about. And after this debacle they probably won't be lucky enough to find themselves on the boards of organizations like Open AI.
You task a bunch of people with reducing congestion and put them to work at a cash-strapped government agency and this is what you get. The one and only way we know which will reduce congestion which also generates revenue for the agency. Pay no attention to the fact that the result may be worse than the problem being solved.
I get it! I am an introvert and have often been called shy. I am also now happily married. But I strangely miss first dates. I miss meeting new people who lead very different lives, and having both of us try to present the best version of our stories. I'm sure it was awkward and filled with dumb small talk. But I miss it.
In the US, you probably wouldn’t collect enough from the first class passengers to pay for their trips let alone subsidize the trips of others. BART and the DC Metro famously serve lots of wealthy people, with a pretty mediocre standard of service, and aren’t close to breaking even. Do we really want to subsidize rich people traveling?
I grew up loving GTI's. Last year I finally had the money to buy one and was in the market for a new car. I couldn't bring myself to buy a GTI because it was just full of touch screens and piano black. VW was unusual in just how far it went. I'm sure they got lots of feedback and data to show that they made a mistake. I'm just curious how they decided to lean this heavily into touch screens (and piano black). Maybe they just desperately needed to save money after Dieselgate and hoped consumer preferences would change over time?
I don't think this is true. It's just that there aren't many optimization practitioners and the OR community has done a terrible job promoting itself. The value in scheduling and logistics is well established. So optimizers tend to congregate. But optimization could be applied to lots of other domains successfully. ML is just minimizing a loss function. I see simulation and post-hoc analyses being applied in numerous situations where optimization would be more efficient and produce a guaranteed optimal result. RL on simple, static problems is really an MDP but I've seen organizations fail to realize this. Etc.
I like the Column Generation idea. Not a big fan of SA and LNS as you often can't really tell the relative quality of the solutions. I've often pushed for a different, simpler model form when the larger model isn't solvable in the usual ways. Running out of memory in CPLEX or Gurobi is a sign you need to reformulate more than anything else. Of course, that requires having people capable of formulating and evaluating math programming formulations. Turns out there aren't so many of those. Surprised me at first, but I guess I shouldn't be surprised. The real trick in optimization is coming up with the right formulation.
I think you're on to something with regards to subways in urban environments moving a number of people that a highway expansion would struggle to match. Feels to me like movement of people should be the real metric and, given that, that we should be prioritizing transit investments in places like NYC, San Francisco, and West LA (as opposed to light rail to Azusa).
Downs-Thomson Paradox is also interesting here and relates to your second paragraph. In a built out environment, transit trips and car trips will take roughly the same amount of time, door to door. Transit advocates have used that finding as a way to complain about road projects, but I think it actually is a much more interesting finding that suggests a more multi-modal approach to planning. I wonder how relevant or irrelevant it is for non-built-out environments.
I agree that VMT is good because of why people make those trips. If you're saying spending all that money on public transit would be a better way to help more people make more of those trips, then you'd be unequivocally wrong in much of the US. Go to someplace like Dallas and marvel at the nearly empty buses and light rail run cars. Maybe we could focus on dramatically improving public transit in someplace suitably dense like West LA. If only the politics allowed it.
But that's true of public transit projects too. They always promise to reduce congestion and never actually do. Congestion charging is probably the only way to really reduce congestion and in many ways the solution is worse than the problem.
And, more than not being a problem, the highway filling up immediately is actually a benefit. Because more people are able and are choosing to travel, to accomplish whatever they need and want to accomplish.
Granted, the externality costs of all that car travel are high. But that's almost a separate issue and one that we really should be solving by a better accounting of costs and additional taxes. It's a shame that won't happen.
Well written and fairly convincing. Slightly off topic but the side point about highway capacity expansion projects interests me. I’m guessing it won’t be popular. But it’s true that more people are traveling and choosing to travel despite the congestion. I’ve never heard the urbanists so quick to dismiss such projects address these points.
Universities will often let admitted students visit, sometimes overnighting with a host who is a current student, before they enroll. You can also take your kids to visit, take a tour, etc. Visits are especially helpful if your kids' choice set includes liberal arts colleges, which can have pretty unique cultures and be hard to rank (e.g., Reed College used to publicly spar with US News and was always under-ranked, St John's College is a totally unique educational model, etc.). And there a plethora of books with short blurbs explaining the character of different universities.
The truth is most college bound kids in the US go to the best state school that they get into. Maybe, for example, they choose Virginia Tech over Virginia if they are interested in engineering. But there's often not a ton of thought put into it. I'm glad you're thinking deeper about this though.
He's wrong about Uber and, in my humble opinion, has been wrong about Uber for a long time. Albeit while making claims that are very interesting. For the longest time, he claimed Uber could never possibly be profitable. I almost believed him despite the obvious evidence of taxi companies surviving for decades before Uber even existed. To me the worst case for Uber was always really large but marginally profitable taxi company. It's debatable but I think they've exceeded that.
The author is absolutely correct that AI can't do physical labor or interact with other humans very well. It's also worth pointing out that there will be new job functions created by AI. There will always be productive ways for humans to spend their time, aka work for humans to do. People predicting mass unemployment focus exclusively on the work we have now. People saying "it's different this time" minimize the innovations of the past.