"His workers must agree to install spyware on their computers so Crossover’s productivity team can track the number of times they click their mouse or stroke their keyboard. The tracking software takes screenshots every ten minutes and, in some cases, snaps photos from PC webcams."
If they really believe they can measure programming ability by the number of times their workers are clicking things, I'd give a wide berth to their software products.
They're hiring the software equivalent of Subway sandwich artists, not chefs at two-Michelin-star restaurants. Sure, both jobs involve placing cheese on top of bread. But in one of them the goal is to do it repeatedly and reliably, not creatively or innovatively.
As a former cook in one and two Michelin star restaurants, I can assure you I move 10X faster than a Subway sandwich artist. I make 10X as many moves with precision. A single mistake can make the whole flow go to crap. Moreover, I can continuously track the temperature of upwards of 100 pieces of different types of meat and fowl, 5 or 6 venison loins, 15 veal chops, 25 pieces of filet, 25 pieces of rack of lamb, ect. I can do that with the optimal rest period before the meat touches the plate. Another place I was leading the fish station. We would plate 40 scallops dishes and 30 of the sea bass dishes at a time with a team of 3. (https://www.foodnut.com/i/Picasso-Las-Vegas/Picasso-Las-Vega...) The cook who did this didn't get the correct angles on the cauliflower quenelles -- so far from perfection.
I write software now. I helped open that kitchen. The first couple months is engineering, how to get organized to plate 90 quenelles in 4 or 5 minutes under a heat lamp. Writing software is just like engineering that kitchen everyday all day, month after month. It's not about creativity or innovation with food. It's about creativity and innovation with solving engineering problems.
I believe it's the parallelism. The subway setups I've seen only allow for about 5-6 concurrent sandwiches, an assembly line of sandwich artists working serially, and their ingredients are basically preprepared. Sounds like this Michelin cook was handling the preparations of each ingredient, including making sure each ingredient is ready at the same time, and then plating them. The quantities of each ingredient they were handling at a time was also much greater.
Awesome life story, with that Michelin star kitchen skillset you have the grit and interpersonal leadership skills necessary to lead a team of devs into the unknown and produce results that matter, hopefully at a director or above level.
Can you see yourself doing this? We need more diversity at the senior leadership level, please consider it if you haven't!
Also if you don't mind me asking, can you share what made you leave your former profession, was it the stress/physical exhaustion? It seems like grueling work to say the least.
Thanks for sharing your background and perspective on HN!
In case anyone is curious, the reason why he said cook rather than chef is chef is a position in the kitchen. It's French for chief and there's only one in each kitchen. The chef doesn't actually cook much himself.
Are software engineers for hedge funds sandwich artists, or chefs at two-Michelin-star restaurants?
Is the answer to every question, "Who is actually doing something creative and innovative?" always, "Me, and everyone else is not?"
Is it bad for software engineering that its reality is starting to resemble subway sandwich artists, and its fiction inside people's heads resembles being a Michelin star chef?
I don't really think that there's such a huge difference between developers in remote countries and developers in first-world countries. Just like many programmers would have imposter syndromes, many unheard-of programmers in poorer places are equally great. The only (or at least foremost) reason they're paid less is simply that they can afford to live a good life with such pay anyways.
The biggest disadvantage to those outsourced developers might be that they didn't have the chance to learn English as well as the first-world devs, and thus might have issues with communication, which is a huge part of software engineering for sure. But in terms of abilities to learn, they are not worse than anybody.
What really worries me is that most non-tech companies (and even some tech companies) don’t understand the difference and frequently allocate resources inversely - paying the equivalent of a Michelin chef to make subs and trying to get top dollar dinner cooked with teams barely skilled in the subject.
The “cloud wage” story could be very compelling in such situations, but it will likely lead to a lot of issues with the more complex systems. Someone has to be the true “architect” and own the successful evolution of the software/business.
If you only have automation or low-end jobs, you are going to lose the ability to make transformative/revolutionary moves. You will still have the ability to refine existing processes, but there are definitely diminishing returns to consider.
The other item that’s also glossed over is sheer productivity - when dealing with more complex issues, experienced developers are a lot more productive, so your $15/hr can easily turn into 10-20x more hours with the wrong experience choice, in the end costing you more.
Hate to use a cliche, but “the right tool for the right job” applies to software too.
And then they evaluate people on their Michelin-star-worthy work, which leads to the phenomenon of "promo-driven development" - people building new complex systems (that immediately turn into legacy code) because they don't get rewarded for doing something simple and maintainable. And then you need more Michelin-star-worthy chefs to keep things running because no sandwich artist will be able to figure it out.
> The other item that’s also glossed over is sheer productivity - when dealing with more complex issues, experienced developers are a lot more productive, so your $15/hr can easily turn into 10-20x more hours with the wrong experience choice, in the end costing you more.
When you are the "consulting company", you don't care. You just get to bill more hours.
If you are the client company, you don't care, either. Anyone with half a brain cell in this day and age knows that outsourcing software development is wholesale idiocy. So, if you're doing that, you are either 1) idiots or 2) outsourcing for political reasons or 3) both.
Interesting article. At least three super-cars, or rather, hyper-cars, are built by hand : McLaren, Koenigsegg and Pagani all use plenty of high tech but have workers who are lifers.
It’s probably not aimed so much at actual productivity, which you can measure in other ways, as it’s a measure for behavioral control- in other words, making sure everyone is always a little bit anxious, feeling a little bit guilty, so that they don’t try to unionize or complain too much.
previously worked on-site at a web hosting company which had similar sorts of spy systems in place - these weren't really checked or used as a primary measure, but more that spot checks were done to see if people were 'generally busy', and also when actual productivity (support request qty/quality) was down or discipline problems occurred the systems would be used to inspect/audit other performance.
programming is for sure less 'mechanical' than routine system support support requests, but I can see the same sorts of measurement as useful in a remote work / no trust environment.
I measure profundity of novels in ink expended, and the success of my parties in the amount of confetti cast. According to the spec this is the perfect procedure. My pet hamster has his doubts.
But writers actually do get paid by the word. Charles Dickens was famously paid by the page, and his novels ended up profound enough. I'd wager most professional writers are in the business to make a living, not to be profound.
That resource says he was paid for every 32 pages. In any case, it seems like for at least some of his early books like The Pickwick Papers, he was paid proportional to volume of output and not to sales/quality. (For some of his later works he earned a percentage of sales.) It's a perfectly fine model for people who aren't already as famous as late-career-Dickens and it can nonetheless produce profound work.
so either they don't know that you can use an issue tracker (or VCS) to get insight into productivity, or they don't use an issue tracker (or VCS), or both.
Someone needs to create a tool that exercises your IDE and gives the impression of activity. You could run it in a VM while you get on with the real work.
If they really believe they can measure programming ability by the number of times their workers are clicking things, I'd give a wide berth to their software products.