sevensor

32 ms·

721.
by sevensor 2y ago
The problem here seems to be inheritance. Why share a base class? Why not write Frobulated = Foo|Bar and go about your day writing functions like def frotz(x: Frobulated) -> str: return f”{x.value} is the value of x”
722.
by sevensor 2y ago
> I think you are underestimating python developers. There are a lot of people out there writing Python, and a lot of them identify as analysts, (non-software) engineers, scientists, and so on. Not developers. Some of them write immacula
723.
by sevensor 2y ago
I think it’s quite useful to separate ser/de, structural validation, and semantic validation. This is where I struggle with a library like ruamel.yaml, running deserialization and structural validation together, or Pydantic, running st
724.
by sevensor 2y ago
I think you have that exactly backwards. LLM-based tools do best on shallow-knowledge tasks. It’s depth where they struggle. Show me the syntax for constructing a lazy sequence in your language? LLM. Reason about its performance characteris
725.
by sevensor 2y ago
A/B testing misses the point of statistical design of experiments, which is that your variables can interact. One factor at a time experiments are pretty much guaranteed to stick you in a local maximum
726.
by sevensor 2y ago
Nonsense. Audio calls with lots of attendees were terrible. Nobody knows when to speak, people talk over each other, you can’t tell if some attendees are even there. You’d be better off writing messages in chat. Maybe just don’t have the me
727.
by sevensor 2y ago
My experience as well. Anaconda users have weird problems nobody else has when it comes to running perfectly reasonable python code, and especially struggle when it comes to integrating outside the anaconda ecosystem. I tell people to avoid
728.
by sevensor 2y ago
I’ve led the maintenance and development of a medium sized (100s of kloc) Python codebase over a period of five years. As it matured, the code became significantly less fragile and more maintainable. Using private members of modules (or a
729.
by sevensor 2y ago
I like how straightforward this one is! It’s fast, it’s obvious, and it’s good enough, if you know something about the data ahead of time. I would have reached for this about four years ago, if I’d known about it.
730.
by sevensor 2y ago
I am so glad I asked. This is a wheel I’ve been reinventing in my head for eighteen years now. I’ve even asked in other venues, why are there no online median algorithms? Nobody knew of even one. Turns out, I was asking the wrong people!
731.
by sevensor 2y ago
Awesome, you did one! I’ll give it a read.
732.
by sevensor 2y ago
Thanks!!
733.
by sevensor 2y ago
I’ve definitely had situations where a streaming quantile algorithm would have been useful, do you have any references?
734.
by sevensor 2y ago
I’ve worked as a non software engineer with non software technicians, and I’ve spent a lot of time thinking about the differences. Most of the work done by engineers was also quite repetitive; in our case the big difference had more to do w
735.
by sevensor 2y ago
Good tip! Search in teams is very disappointing.
736.
by sevensor 2y ago
Although I agree with the recommendations, I cringe at the definition of abstraction. In a sane world, abstraction doesn't mean defining classes so much as it means identifying important unifying concepts. DRYing your code by moving
737.
by sevensor 2y ago
Not especially :) Legible but somewhat awkward.
738.
by sevensor 2y ago
Not enough context about why you’re handwriting the code. Personally I sketch out data structures by hand, and that’s most of what I write down. I try to use approximately correct syntax there because it keeps me honest. Sometimes function
739.
by sevensor 2y ago
I'd counter that the South is an interesting place in its own right. Skipping it because it's too much like Western countries is to miss out on all the fascinating ways it isn't. Furthermore, I'd question the value of f
740.
by sevensor 2y ago
Or, you can visit South Korea today, and if you want to go off the beaten path, leave Seoul. The east coast is a beautiful and fascinating place. Go to Gangwon, ride a cable car up Seoraksan, eat the best seafood of your life.
741.
by sevensor 2y ago
> The massive influx of poorly prepared junior programmers in the last decade It's been more than a decade. Take a time machine back to 2014, and we're complaining about religious zealotry among followers of the GoF book, deba
742.
by sevensor 2y ago
It's well-deserved. I admire the author's go-for-broke attitude and his willingness to speak potentially career-destroying truths while the rest of us gripe about misguided ML projects in the safe pseudonymity of HN comment threa
743.
by sevensor 2y ago
This may be a bitter pill to swallow, but three years' experience rounds down to zero. With no other information, it's enough to make me expect basic ability to use git and execute well-defined tasks within a narrow scope. If yo
744.
by sevensor 2y ago
Fabulous podcast. Fascinating how we can use pronunciation shifts to date the arrival of a word into English, and how some words have entered the language multiple times by different routes. Also, I learned that there really was a castl
745.
by sevensor 2y ago
There's a charming transcription error in the first letter, substituting "fun" for sun, no doubt due to there having been a long s in the original.
746.
by sevensor 2y ago
A whole bunch of problems can be set up either way. MILP always has an objective, and the constraints are always linear combinations of the decisions. Gurobi is so incredibly fast that it might be worth contorting your problem into a MILP
747.
by sevensor 2y ago
The data engineers I know started out as data scientists. They discovered that their ability to make models that did anything useful was hampered by low-quality or absent data. Then they discovered that there was a lot more job satisfaction
748.
by sevensor 2y ago
This was a particularly lame argument: > If NASA dumps ISS on the Earth, it will be the worst public relations disaster in its history. At the very moment, new generations of eco-conscious young people are taking the reins of control, a
749.
by sevensor 2y ago
If you find that this is the case in your organization, watch out. It's a sign that you're trying to build performative models that aren't grounded in reality. Depending on your organization's goals, this may take a lo
750.
by sevensor 2y ago
AI problems turn into data problems. The happiest and best compensated people I know in that area have gone into data engineering, because data engineers are the ones selling shovels in this gold rush.

More ›