carlosf

10 ms·

31.
by carlosf 5y ago
Yeah, modern CPUs already have hundreds of millions of logic gates and can do calculations much faster than the human brain, yet we need an ridiculous amount of code hours to display some stuff in a screen in a way that barely works. My 100
32.
by carlosf 5y ago
Exactly, periferia.
33.
by carlosf 5y ago
> without arguing (which doesn't work) My anedote: I went from 290 lbs to 190 lbs over the last year. I totally knew that I was eating garbage, but I thought I was young, so I thought I could just switch to being healthy a few years
34.
by carlosf 5y ago
> Most of what you get from elixir in terms of redundancy, high availability, etc you can have that anyway from kubernetes, heroku or any PaaS.... you will need more than 1 server anyway, so... Yeah, last time I tried to get a hang in Er
35.
by carlosf 5y ago
We can't predict shit like whether is gonna be sunny tomorrow, truth is the consequences of climate change are unknown, as the earth is a terribly complex system. Just like Covid, economics, politics... my only prediction is that most
36.
by carlosf 5y ago
1 - Nope 2 - Pretty much yeah
37.
by carlosf 5y ago
Nice read. As a sysadmin / DevOps / SRE / whatever, I also realized at some point that being constantly busy is actually a state of extreme fragility. Nowadays I try spending a significant part of my day just trying new stuff
38.
by carlosf 5y ago
Search for voice coding. That's how many voice coding algorithms work, you try to find a digital filter that generates a sound that is as close as the original according to a perception based metric, then transmit the filter coefficien
39.
by carlosf 5y ago
AWS's own offerings such as RDS and internal control plane stuff are very likely using ARM behind the sheets. I have evaluated going ARM, but I ended up deciding the savings were not worth it. Not only you need to mantain 2 archs simul
40.
by carlosf 5y ago
I migrated from Terraform to AWS CDK and I'm very happy. The good: - Tooling for any major language (Typescript, Python)… is lightyears ahead of anything you will find for Terraform / HCL. - You can write your code as declarativel
41.
by carlosf 5y ago
> In a modern telecommunications service provider, new equipment is deployed, configured, maintained and often financed by the vendor. Just to let that sink in, Huawei (and their close partners) already run and directly operate the mobil
42.
by carlosf 5y ago
I gave up on using mypy as I was spending more time telling it to shut up than actually using it catch mistakes ahead of the time. That said using type annotations made me realize I want to be fluent in a language with a strong static type
43.
by carlosf 6y ago
When I was like 14 years old one one my math teachers explained to us what calculus was using zero equations. He made everything look really easy. At college, it was the complete opposite. My professor started with limits and convergence, d
44.
by carlosf 6y ago
Food for thought: if every method for predicting event x is terrible, then you might as well not try to predict x and build your life in such way that you never expose yourself to the risk of x happening.
45.
by carlosf 6y ago
I agree, but that's equivalent to my original claim. Being bad at extrapolation is a consequence of assuming all training data can describe your phenomena distribution and being wrong.
46.
by carlosf 6y ago
I disagree, every ML model has some implicit statistical assumption, which is often not well understood by practitioners. At minimum you must assume your underlying process is not fat tailed. If it is, then your training/validation&#x2
47.
by carlosf 6y ago
I am increasingly worried with people applying ML in everything without any rigour. Statical inference generally only works well in very specific conditions: 1 - You know the distribution of the phenomenon under study (or make an explicit a
48.
by carlosf 6y ago
Lambda has a ton of caveats and limitations... but it seems their platform is full of limits as well: https://deno.com/deploy/docs/pricing-and-limits AWS sucks hairy balls at providing things that are simple for d
49.
by carlosf 6y ago
Just a thought: that is more related to security models than to GUI vs CLI. In Windows you can delete and break things if you're a local admin. Warning messages got better, but using a PC also used to be intimidating. Smartphones on th
50.
by carlosf 6y ago
Kinda agree with the author. I also realized a while ago that pinning Dockerfile dependencies is a terrible practice unless you want to hire people to basically do Dockerfile maintenance. BTW this site is full of gems I had to learn through
51.
by carlosf 6y ago
I'm looking for a managed Knative solution and Google Cloud Run looks like a great product. The only reason I'm not using it is because I also had first hand experience with Google's atrocious customer service in the past.
52.
by carlosf 6y ago
Not sure if it's already GA, but AWS announced a product called EKS / ECS Anywhere last year that covered this use case. Edit: not GA yet
53.
by carlosf 6y ago
Nice writeup. The catch is it's very hard to measure correlation reliably for badly behaved random variables (like the ones we have in finances).
54.
by carlosf 6y ago
GitLab customer here. Sometimes I also feel like they are adding a lot of stuff instead of improving their core services. BTW I already gave up on using GitLab's native caching in favor of doing all my CI/CD with Kaniko + Multi St
55.
by carlosf 6y ago
I've recently found about Protocol / structural subtyping in Python and found this sort of pattern really useful as duck typing is so common in py code. Highly recommend reading: https://www.python.org/dev/pep
56.
by carlosf 6y ago
I actually had a cool experience opening it on the mobile. It sucks on the desktop though.
57.
by carlosf 6y ago
> let's do so with eyes open that we will be headed to uncharted territory. That's how I think. In economics, society and very complex systems in general, one can rarely predict the consequences of most actions. The best we can
58.
by carlosf 6y ago
Great read. I believe that's an issue not just with container security scanners, but with automated security in general. One example: I enabled AWS Security Hub to have some sort of security score on my account and although it found a
59.
by carlosf 6y ago
Some Udemy courses are pretty good and are taught by practicioners, while Coursera is much closer to what you get in University. Coursera -> Helped me to graduate Udemy -> Helped me to get useful skills quicker and make money
60.
by carlosf 6y ago
As a GitLab / Docker user, testing pipelines is definitely a major pain point in my devops process. Cool stuff!

More ›