Pinned Tweet
I was paying >$300/month for accounting software that couldn't send invoices reliably. So I built Zippato.
Carlos Souza retweeted
This article exposes the real challenge in AI SDLC: our industry is still trying to figure out how to let go of old patterns. It is like fighting nature instead of aligning with it.
Here is a quick story that exemplifies this perfectly.
Once a sailor named Amir Klink asked an engineering professor: how do I build a boat that never capsizes?
He planned to row from Africa to Brazil, but waves get crazy in the Atlantic.
The professor laughed and said it was impossible, but suggested he build a boat that always gets back up after capsizing.
That is what Amir did and it worked. He came rowing and tumbling from Africa to Brazil.
Most people are trying to build the fastest, safest way to build software with AI. We will never build a system that never crashes. The nature of systems is to crash, even if the logic is perfect, hardware isn’t.
The challenge is to design systems and processes that recover from crashes virtually instantly.
That is the biggest takeaway from this article.
Latte Macchiato update: used a little bit of honey syrup and the espresso layer settled in the middle a lot easier.
A verdadeira rinha de backend! Ansioso pra acompanhar a contribuição da galera e ver esses números🍿
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-cam…
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-…
raw.githubusercontent.com/basecamp/once-…
github.com/basecamp/once-…
github.com/basecamp/once-…
x.com/zachsdaniel1/s…
Carlos Souza retweeted
I mixed Jev + an espresso (state) machine.
Turns out, Jev's a pretty good barista ☕
Introducing Jevspresso: XState determines what's possible, Jev decides the next move. Order in plain language, or break stuff and watch Jev cope.
Try it: jevspresso.dev
Carlos Souza retweeted
There's been a ton of talk about the role of humans in code review, and when and how humans should be signing off on changes.
I believe that, long-term, humans have no role in routinely reviewing code.
I don’t care how you ship, I care that you are accountable for it
The "whiteboard defense:" I should be able to pull you aside at any moment and ask you to explain any customer-facing system you've shipped. You should be able to clearly explain how it works and defend the decisions you made. This is my benchmark for responsible AI usage.
I don't expect line-level familiarity with the code. I don't care if you remember the exact function name or implementation detail. You may not even know it. I don't care.
But if I ask "why did you do X instead of Y?", "what happens if this actor behaves maliciously?", "what data structure did you use here and why?", or "where does this fail?" you should be able to answer confidently.
For PoCs, demos, experiments, whatever: I don't care. Generate 100% of it and understand none of it. Speed over quality every time in those specific scenarios.
But if you're shipping customer-facing work, you can't be shipping things you don't understand at a high level.
Carlos Souza retweeted
Does anyone here look for seasoned #elixirlang developer located in EU for remote work?
I am looking for work.
#myelixirstatus
Sounds like a great way to build an awful company culture.
We're using a new method to hire engineers at @HelloUntangle.
I think this playbook will become the standard for hiring technical talent in the age of agents.
Here's how it works:
1. We ask candidates to submit a screen recording of their whole screen, shipping a new feature on a current app they work on.
Note: This is usually 45 minutes of video - no editing or speeding up - we want to see them thinking, prompting and problem solving with the agent - we also want to see their devops skills.
2. We watch these videos and pick a subset to go through Round 2.
3. Round 2 is giving them a seat on @DevinAI so they have secure access to our repo.
4. We ask them to ship a new real-world feature and get the PR merge-ready.
5. We pay them a fixed fee to do this, and they have 16 hours to complete the work.
6. We watch the videos, review at all the Devin threads, have an agent do the same, and then choose which one to hire.
Note: We don't have any meetings with the folks before any of these things. It's all about the work and the results.
Carlos Souza retweeted
For the first time in a while, I have an opening for new clients (in September).
Do you need help with your product, engineering team, or an Elixir/Phoenix project?
How I help teams:
- ✨ Product: building the best possible product for your customers. I care about the solution end-to-end.
- 🧑💻 People: strengthening your team's technical and product skills as we improve the product (remember, your team is the engine behind your product). Mentoring, pair programming, workshops, whatever is needed to make your team the best version of itself.
- ♻️ Process: continuously improving how your team builds, coordinates, and delivers your product
Through it all, I practice and teach a culture of continuous improvement so that, long after I'm gone, your team can continue the virtuous cycle of self-improvement.
DMs are open, or email me at german(at)germanvelasco.com
For more on how I work, you can see germanvelasco.com
#MyElixirStatus #ElixirLang
Carlos Souza retweeted
I published a little post about @elixirlang GenStage demand: andrealeopardi.com/posts/gen…
It has interactive charts which I totally build line-by-line by hand without any AI. Nah but I’m still really happy to finally be able to put a "visual" explainer like this out there! 🎆
