Erick ๐Ÿ™ƒ Navarro (@erickgnavar) header
Erick ๐Ÿ™ƒ Navarro (@erickgnavar) profile picture

I write(well AI does it now) bugs and get paid for that โค #emacs #python #elixir #lisp Stream coding stuff in Spanish at https://nitter.gq/t.co/8MKwqi50r6

the cloud โ˜๏ธ
Joined December 2008
Photo by Erick ๐Ÿ™ƒ Navarro (@erickgnavar)Photo by Erick ๐Ÿ™ƒ Navarro (@erickgnavar)Photo by Erick ๐Ÿ™ƒ Navarro (@erickgnavar)Photo by Erick ๐Ÿ™ƒ Navarro (@erickgnavar)Photo by Erick ๐Ÿ™ƒ Navarro (@erickgnavar)Photo by Erick ๐Ÿ™ƒ Navarro (@erickgnavar)Photo by Erick ๐Ÿ™ƒ Navarro (@erickgnavar)Photo by Erick ๐Ÿ™ƒ Navarro (@erickgnavar)Photo by Erick ๐Ÿ™ƒ Navarro (@erickgnavar)Photo by Erick ๐Ÿ™ƒ Navarro (@erickgnavar)
Erick ๐Ÿ™ƒ Navarro retweeted
lol my lawn mower is only used 1x a week. Should I rent one from OpenAI instead? Also, who caused the DRAM shortage? It wasn't consumers...
sam: local inference is a terrible allocation of scarce resources โ€œa lot of people want local inference, and somehow they assume that the problem is solved now because we have open-source models, we're going to be able to run them on our computers, and so we're not going to need all the stuff in the data centre but from a global allocation perspective, actually it's quite terrible to have everybody, like, have a bunch of the world's limited supply of DRAM sitting in their own computer and being used five percent of the time you want maximum utilisation in the cloud if you're trying to deal with this very supply-constrained environment somehow many people's intuition is the opposite of thatโ€
188
434
8
7,070
103,000
Erick ๐Ÿ™ƒ Navarro retweeted
34
239
20
2,525
68,293
Erick ๐Ÿ™ƒ Navarro retweeted
OpenBSD 8.0 is out now! See what's new: openbsd.org/80.html
22
323
29
1,518
107,355
Erick ๐Ÿ™ƒ Navarro retweeted
So I was trying to understand why pull requests speeding up the Elixir version were being denied by the clankers while they were being accepted to Go/Express/Rust. It turns out that, not only each project had a different starting point and a different prompt, but also each project also has a different set of verifications that they are required to pass! For example, Elixir is the only guided to "preserve the pinned Rails reference behaviour", the only with a explicit verification that we are using Redis and draining its queue (so we cannot remove it while others can), and the only one actively enforcing it. So it turns out that the people telling you to not read the code cannot even use AI to set up fair benchmarks (probably because they would have to read the code in order to do that). You don't have to believe me, you can do your own research or ask your agent to do it for you. E.g. here is what Claude Code came back with: gist.github.com/josevalim/aeโ€ฆ
Replying to @dhh
Also extracted a separate verification project that's responsible for the benchmarking and checking of new implementations and improvements to the existing ones. The full report from the latest runs is there too. github.com/basecamp/once-camโ€ฆ
68
106
16
1,033
59,599
Erick ๐Ÿ™ƒ Navarro retweeted
Oh FFS. It turns out the req/s in the initial vibed Rust version is within the same order of magnitude as the initial vibed Elixir/Go versions, and even performs worse in some of them. The history is in the repository! The Rust version only got faster with a bunch of additional work. The initial Elixir/Go versions presented in the tweet below saw none of this additional work. It was comparing a 300 commits Rust version with a 2-3 commits Elixir/Go version. Source: github.com/basecamp/once-camโ€ฆ
I've had agents implement and optimize the Campfire web app in Elixir, Go, and Rust. Yes, Ruby is the slowest. I didn't care about that when the pay-off was huge in terms of productivity and developer joy. But if you're no longer reading the code? github.com/basecamp/once-camโ€ฆ
Readers added context they thought people might want to know
The shown benchmarks use non-equivalent implementations: Elixir routes all DB queries through a single GenServer with Redis for jobs (retained from Rails), while Rust uses in-process queues, omits CSRF tags/fields, and applies other caching changes. github.com/basecamp/once-โ€ฆ raw.githubusercontent.com/basecamp/once-โ€ฆ github.com/basecamp/once-โ€ฆ github.com/basecamp/once-โ€ฆ x.com/zachsdaniel1/sโ€ฆ
77
109
32
1,359
170,737
Material de estudio didรกctico
56
Erick ๐Ÿ™ƒ Navarro retweeted
If you'd used Emacs, you would never have had to switch away to begin with.
One nice thing about the IDE being dead, at least for me... I first used UltraEdit as my "IDE" for 10 years, then I switched to Sublime for the next 10. I had just started to convert to VSCode when the AI revolution took over... allowing me to forgo the transition and revert back to Sublime.
17
6
1
94
8,621
GLM 5.3 doing a security review about my project ๐Ÿคฃ
1
107
Erick ๐Ÿ™ƒ Navarro retweeted
all linux distros are just piles of bash scripts except for nixos, which is configured using an advanced lazy functional programming language... ...which evaluates to a pile of bash scripts
33
35
5
1,214
66,591
Hi @Pebble, why do I need an account to be able to see battery usage? ๐Ÿค”
137
Usa #rust decรญan, serรก divertido decรญan
1
156
Cuando el junior(deepseek-v4-flash) no encuentra el error y llamas al mid(deepseek-v4-pro) y tampoco lo encuentra y toca llamar al senior(glm-5.2) que cobra mรกs caro
1
12
1,039
si estรก mas caro GLM5.2 comparado a DeepSeek4 ๐Ÿ˜…
72
No GitHub, no voy a instalar eso que encima estรก bien caro
1
120
Erick ๐Ÿ™ƒ Navarro retweeted
Elixir v1.20 released! Now officially a gradually typed language: Elixir type checks every single line of code, finding bugs and dead code, without developer overhead (no typing signatures) and extremely low false positives rate. Plus a faster compiler! Links and reports below.
55
307
47
1,486
129,835
Erick ๐Ÿ™ƒ Navarro retweeted
When someone replies to a pull request comment with obvious AI content, it genuinely saddens me. PRs used to be a place to teach/learn/discuss software but now thereโ€™s nobody on the other side. If I wanted an agent response, Iโ€™d ask mine. Social coding is dead.
63
59
12
544
17,954