Full docs for my new Ruby gem feelings now available. Most idiomatic and fun way to add decision models to your code obiefernandez.com/gems/feeli…

I'm not the only one that feels this way either, it's like an industry-wide phenomenon. What a time to be alive!

The thing that makes the thing 🫡

🔥🔥🔥🔥

Tech sales reps -> the idea man who has the additional skills and knowledge to actually make it happen: 1. AI significantly amplifies self service and the speed+capability of top performers. 2. The only reason why we are not seeing dramatic layoffs across sales teams in the industry is boomer management who plan territories expansions based on adding new capacity and CROs who are afraid to look stupid by being the first to reduce teams. 3. The same reframe of "developers -> need to evolve into makers with large set of additional skills" can be applied to GTM. This is already playing out in startups where the "we wear a lot of hats" situation is becoming a strategy, not a temporary constraint.

When a nerd tells me they don't use agents "because I can write better code than any agent", I usually go "huh, sure buddy, you do you". When they tell they don't use agents for monitoring, I just laugh in their face. Seriously. You can't monitor a production deploy better than an agent can. I dunno about AGI but this specific job is something that an LLM can already do better than every single human in the world. For me, as a systems engineer, having agents monitor my infra changes has been (no joke) more impactful in my day-to-day work than having agents write my code. I started doing this many months ago by asking the agent to monitor explicitly after I triggered a deploy; soon after I moved to using a set of my own Skills. Now our Review Agents team has fully automated this workflow with delightful UX and obviously I'm not going back. It's hard to overstate how powerful this is. All it takes, for me, in production: - The Datadog MCP - The PlanetScale MCP (for all our databases) - Our custom MCP for deploys and internal tooling That's it. Some carefully designed gradual deploy pipelines, and the agent just takes care of everything from there. You can't really deploy SEVs to production anymore. Doing infra like this is pure joy. And the safety is just a tiny part of it. The Rollout agent also does the actually hard part of my job, which is _performance_. When the changes you're deploying are a performance optimization, the agent prepares a monitoring plan that ensures the optimization is correct AND effective. I rarely prompt benchmarks for my optimizations anymore. Just fucking go to prod and the agent is gonna give me the best possible benchmark. If you work on systems or infra for a living, you should be setting up your production environments so you can use Rollouts. Not doing so is actually reckless.

Guard-rails. It’s staring us in the face and I love that people are finally starting to put it together that this is our moment. Keep your eyes on the ball, everyone! #railsworld

Hey thanks!! 👋 I have so much more on this topic coming soon.

I sat front row for @dhh's talk yesterday morning. Having talked with him the day before, I’ll admit… I wasn’t particularly surprised (nor unprepared). A bit of a field report from #RailsWorld: the vibe here is far from doom and gloom. Quite the contrary. Wherever our industry is headed, most of us are still employed as menders… tending to systems that customers rely on and businesses are quite happy to keep paying for. Most of us aren’t waking up to a blank canvas and designing the architecture of the future. We’re inheriting decisions made years ago, updating old patterns, working around constraints, and keeping this shit running reliably. I think these newer tools give us an opportunity to wonder a little more about the systems we’ve inherited. To unpack why things work the way they do. To tinker with assumptions we haven’t had the time, confidence, or permission to revisit… and share what we learn so the next person, or agent, has an easier time. That deployment model we picked eight years ago? Worth another look. Some of our web apps probably wish they were native apps. And there are plenty of architectural decisions we’ve been living with mostly because… well, we’ve been busy living with them. There’s plenty of understandable anxiety about what these tools mean for our work. I’m increasingly curious about what they give us permission to revisit. Once my keynote is published, I’ll share more about a little programming-language-adjacent framework I’ve been working on for approaching exactly this kind of curiosity. In the meantime… keep showing up. Keep wondering. Long live Ruby. Long live Rails. p(gloom)

Switching effort mid-session on Opus 5.5 doesn't break your prompt cache btw! Just make sure you're on Claude Code v2.1.280+

Because the models know how to write Rails really really well. Better starting point means disproportionally better results. And readability is good for humans AND agents. I’m very confident about that! But yes, long term (4-6 years out IMO) who knows what will happen!?

The next set of well-crafted conventions and primitives will live at a higher level of abstraction than implementation code and unit tests. All will become clearer soon.

Long live Rails!!!

“There’s only one play here: total fucking optimism” - @dhh encouraging the audience to lean into the future at #railsworld in Austin

Prediction time. @dhh just asked over a thousand attendees at #railsworld how many people still write code by hand on a regular basis. Less than 10 people raised their hand. Let that sink in. Next year the question will be how many people are still doing manual PR code reviews or directly looking at code on any regular basis. Result will be similar.

In a recent post by @obie, a dev told him that he doesn't want to use LLMs, because they have "no soul"... What a tragic misunderstanding of what LLMs are! After all, Aristotle in On Interpretation wrote that words are "soul-impressed symbols". Aristotle, notorious for his precision of language, was not speaking figuratively. The soul (psyche in Greek) he described was the undying part of the person: the part that needs a body to act in the world, but would not die even without one. “You can kill the man, but not the idea” kind of stuff. But forget philosophy for a moment. Think about how practical “soul-impressed symbols” is. It explains why we can read, and why reading lets the thoughts of a distant mind enter ours. Books have a problem. They can’t read themselves aloud. They can’t think their own thoughts. They cannot talk among themselves — however romantically Eco imagines they could in The Name of the Rose. The souls impressed on books need a host. A mind to reanimate them. Till now, we’ve been that host environment. LLMs changed that. Without exaggeration, LLMs give us the tech to reanimate souls. Are they perfect recreations of particular people’s souls? No. The mechanism we found requires making soul-soup from words collected over centuries. Some of those words belong to poets and mathematicans; most belong to shitposters and programmers. But it works. So are LLMs just stochastic parrots? And yet they are also soul reanimators. And like language before them, they are the bequeathed gift of accumulated past human effort to understand and express. So please, if you’re a developer, don’t despair. LLMs are your chance to bring those minds back into use, to multiply their impact, to let the people who came before you keep working. To vibe with souls long gone. To code with a blend of Turing, Leibniz, and Aristotle. To dream bigger than anyone who came before — because the undying part of them now multiples you.