Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
sevagh
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
30 ms
·
361.
▲
by
sevagh
5y ago
It's definitely not guaranteed income (the competition is stiff if you're not on top of your field). But it's an option.
362.
▲
by
sevagh
5y ago
I'm certainly not hoping it goes badly for him. I just don't think it's healthy to learn about a source of volatility and then immediately go all-in on the next one.
363.
▲
by
sevagh
5y ago
2018: life was perfect with my perfect partner until my perfect partner left me 2021: life is perfect with a new perfect partner who's never going to leave me unlike my previous perfect partner 2023: any guesses?
364.
▲
by
sevagh
5y ago
>(mostly innder) I suppose the problem is when it stops being inner, and it starts spilling outward.
365.
▲
by
sevagh
5y ago
I have a discomfort about people showing off their consumption habits or tastes. The other day I read a sentence like, "I admire a person who listens to tasteful music, does their research and has a carefully curated music hobby"
366.
▲
by
sevagh
5y ago
If you want to work on machine learning problems, you can try your luck at winning AICrowd/Kaggle competitions. Some have cash prizes: https://www.aicrowd.com/ https://www.kaggle.com/competitions
367.
▲
by
sevagh
5y ago
I agree with you here. Using Prometheus, federated Prometheus, and Thanos on top of it for good measure, would probably get you better results without using a hodge-podge of non-Prometheus-compatible tools.
368.
▲
by
sevagh
6y ago
I have a Makefile for a Rust project which binds the local repository to a Docker volume, builds it in the container using muslrust, and then does a chown to change the target directory back from root ownership to my own user. All I had to
369.
▲
by
sevagh
6y ago
Did you mean ruminating?
370.
▲
by
sevagh
6y ago
>This about sums it up. This is a rant and it’s difficult not to just close tab half way through. Every single rachelbythebay article: Here's a vague condition. I encountered it at vague companies that I may have worked at. Anyway,
371.
▲
by
sevagh
6y ago
Go more than makes up for this in the deployment aspect, since deploying Python in 2020 (Python 2, Python 3, conda, pip, venv, system interpreter, poetry, etc.) is still a mess.
372.
▲
by
sevagh
6y ago
I second the Thanos link, but if you want to only use Prometheus, I would look into federation: https://www.robustperception.io/scaling-and-federating-prome... https://prometheus.io/docs/prometheus/
373.
▲
by
sevagh
6y ago
Have you looked into Grafana Loki for logs? If I had to redo one part of our stack, that's what I'd choose.
374.
▲
by
sevagh
6y ago
False in my experience. Full-time job? After the initial learning curve, a simple 2x redundant Prometheus poller setup on can last for a long time. Ours lasted for 30,000,000 timeseries until encountering performance issues. After that, we
375.
▲
by
sevagh
6y ago
I've had luck with https://thanos.io/ for a big (~1 billion timeseries across all our DCs) Prometheus scale out project. Horizontally sharded Prometheus that can be queried and alerted on in a unified view with object
376.
▲
by
sevagh
6y ago
What's the origin for the fascination with blogging in the tech/software world? I used to maintain one (mostly because I worked with other engineers who fetishized blogging and encouraged everybody to blog). I eventually lost inte
377.
▲
by
sevagh
7y ago
So did you indeed loop over every possible frequency? I know I attempted a Goertzel-based pitch detector once, where I looped over frequencies and picked the one with the largest magnitude. I picked 25Hz and 4200Hz as my limits (from a pian
378.
▲
by
sevagh
7y ago
OK, fair enough. A MIDI file would contain the musical information necessary to generate any notation. Thanks for the reply. I see indeed the link to the Mcleod pitch method paper is dead. I'll see about hosting it in a git repo.
379.
▲
by
sevagh
7y ago
Interesting project. What do you think of https://github.com/sevagh/transcribe ? It seems like in your project the word transcription isn't referring to sheet music transcription?
380.
▲
by
sevagh
7y ago
Bizarre that people keep mentioning their "guitar tuners" that use Goertzel. Are you looping over every single possible frequency? You should be using real fundmental frequency estimation/pitch detection algorithms ([1], [2])
381.
▲
by
sevagh
7y ago
You speak of them with such disdain. Is that a common attitude in your company?
382.
▲
by
sevagh
7y ago
Agreed - some of the best communication, interviews, and offers I've got came from SO jobs. It's the first place I look as well.
383.
▲
by
sevagh
7y ago
Good timing on having recently began working with BPF + XDP. Here's my project: https://github.com/sevagh/ape , for those curious about what itches XDP can scratch.
384.
▲
Building a Warped Linear Prediction Audio Codec
(sevagh.github.io)
1 points
by
sevagh
7y ago
|
0 comments
385.
▲
by
sevagh
8y ago
>[1] Such a benefit doesn't matter in a containerized world. Go is used for the tooling for deploying this containerized world. Let's take a look at some infra/ops tools written in Go: * Docker itself - should run on barem
386.
▲
by
sevagh
8y ago
The book Behave by Robert Sapolsky describes this topic (among others) nicely. https://www.penguinrandomhouse.com/books/311787/behave-by-ro...
387.
▲
by
sevagh
8y ago
Yes, he had the author of "Why We Sleep" on and the research is clear that marijuana reduces REM sleep - the health impact of _that_ - of reduced REM sleep - is yet unclear, AFAIK. Like you say: >I'd agree that sleeping is
388.
▲
by
sevagh
8y ago
>During my junior and senor year of college I started smoking pot, and I started sleeping on a regular basis, going to bed about 1-1:30 and getting 6/7 hours of sleep. I'm not really sure if I was happier, or more productive, I
389.
▲
by
sevagh
8y ago
>This is actually a common misconception. In reality, it is not possible to "catch up" on missed sleeping hours. I find that the same sentiment is worded misleadingly in the book Why We Sleep. If you have a day of bad sleep (i.
390.
▲
by
sevagh
8y ago
>(For youtube videos, when I'm trying to get information on something, I employ a "rule of third": just drop the slider a third of the way in and skip all the personal introduction and setup, to see if this video might act
More ›