jw887c

5 ms·

1.
by jw887c 1y ago
You are the poster child for depreciation. You might have received more utility than the price you paid but at the end of the day, that dollar value once you gave it away was $0
2.
by jw887c 2y ago
>Avoid using Poetry for new projects. Poetry predates many standards for Python tooling. This means that it uses non-standard implementations of key features, such as the dependency resolver and configuration formats in pyproject.toml fi
3.
by jw887c 3y ago
Multiprocessing is great as a first pass parallelization but I've found that debugging it to be very hard, especially for junior employees. It seems much easier to follow when you can push everything to horizontally scaled single proce
4.
by jw887c 3y ago
Yes
5.
by jw887c 4y ago
>We were on AWS Managed Airflow, but to stay on it and have a solid platform, I would have been writing Github Actions for CI/CD, standing up ECR and IAM roles with Terraform, setting up EKS to run Kubernetes jobs, managing infra mo
6.
by jw887c 4y ago
This is the case for most column oriented data warehouses (including BigQuery, but Snowflake does allow for 1 PK). It's just the nature of the technology.
7.
by jw887c 4y ago
Skydiving is cold. Skydiving while naked sounds freezing
8.
by jw887c 5y ago
My favorite flaw of averages isn't even mentioned in this article. It's the aggregation of averages across covariates. The more covariates (higher dimensions) your problem has, the less likely the population will exist "in th
9.
by jw887c 6y ago
Great tip.
10.
by jw887c 6y ago
A $40B manager once told if I could deliver 6% annual return with 1% volatility, he would give me all his money. Yeah, they want as little volatility as possible.
11.
by jw887c 6y ago
GCP is behind. They're are behind in niche services to be fair but behind nonetheless. Nothing that prevents building similar services, but coming from AWS, I routinely feel frustrated by lack of finesse and comparable offerings. IMO,
12.
by jw887c 6y ago
If you prefer yaml > Python or if you prefer installing one k8s app instead of managing all of the airflow dependencies (scheduler, webserver, workers, etc)
13.
by jw887c 6y ago
Special room for classified stuff. Even once you're inside the classified room (usually windowless and behind a locked keypad) a lot of stuff is also behind locked filing cabinets.
14.
by jw887c 6y ago
Oh definitely. I have TONS of buildup after a few days. There is a (anecdotally) significant difference in hearing ability post cleaning via q-tip. If I don't use the q-tip, Airpods get very gross within a few days.
15.
by jw887c 6y ago
Depends on which part of the ladder you are on. If you're the fund manager of a big fund, you got rich. If you're the LP of growth seed rounds, you probably got rich. If you're the LP of later stage rounds, you probably didn&
16.
by jw887c 6y ago
Ex-aerospace engineer here, now software developer. This is one major reason why a lot of younger folks at Boeing leave (at least the ones with a more broader view of their industry). A lot of "engineers" at Boeing are actually pr
17.
by jw887c 6y ago
Yeah this doesn't work for non-white folk. Just makes us look like white folk
18.
by jw887c 6y ago
Softbank is every VC/PE's analyst and associates punching bag
19.
by jw887c 7y ago
I view it as: Explainable models = statistical models (regression / multilevel models). Black box models = machine learning (tree based models). AI = black box models for computer vision / NLP.
20.
by jw887c 7y ago
Its not surprising when you realize how bad wealth inequality has made investors searching desperately for any place to put money.