This is a great insight from TestDouble and it is even more true with Gen AI about how speed comes from flow and flow is a systems concept.
Source: https://t.co/XuKZhfPefe
“But the AI changes that equation. The parts of Hotwire that frustrated me, writing all those Stimulus controllers and wiring up data attributes by hand, are exactly the parts the agent handles well.” Agents like Hotwire as much as I do!
Rails 8.2 will use Sec-Fetch-Site instead of cookie tokens for CSRF protection by default. Love seeing browsers compressing complexity like this. Stellar work by Rosa to extract this from Fizzy 👌https://t.co/eXgjulSASn
This is just a drop in the AI ocean, but as a long-time fan of IDEs, it’s amazing how LLMs routed around decades of progress in language processors and tooling to arrive at an interface that is 100× better.
https://t.co/F27eWi0m2D
It's great to see this feature (although limited) finally supported out of the box.
P.S. We could have had it for 10 years already: https://t.co/nZeB2mbiWu.
(Well, I had it https://t.co/UptBjFr3CM.)
P.P.S. There is also store_model for turning bags of values into proper objects:
https://t.co/Ri51qff58m.
My view on all the LLM-powered code review processes I've seen: 90% of it is just creating text slop absolutely nobody reads.
Doing it yourself, before you push to GH, is fine. But making your co-workers interpret this noisy slop is wasting their time.
That moment when your Ruby Kafka producer goes from 160 to 800 msgs/sec with ONE line change 🤯
Just wrap it in Async do...end. That's it.
Wrote about how WaterDrop + fibers finally makes Ruby feel fast for I/O work:
https://t.co/hVC33hZ7KM
#ruby#rubyonrails#apachekafka