@excid3 If the flag is enabled, the version in the define call is gone (no arguments are passed), so as a side-effect you reduce the chance of merge conflicts too.
@excid3 Works most of the time, but it has the gotcha that "newer" is not equivalent to "not applied". Dumping reduces schema_migrations to one value, so data is lost.
In 8.2, https://t.co/KXY80gsDr3_record.dump_schema_migrations = true will dump schema_migrations in db/schema.rb too.
@tomigotchi_ Fall behind? LOL
What I am saying is that the result is **incorrect** or subotimal and you don't see it because you have less implication, less agency, and in certain cases you are a mere proxy for real.
And that is not all people. You can be mentally as active and do AI.
This is really interesting.
A language that compiles faster than Crystal, runs Crystal, requires less tokens, reads kinda like Ruby, from the creator of Kemal.
@ivyevanslv I totally see that. You are someone very inquisitive by nature, mentally very intentional and active, so your own natural way is to exploit the new tooling optimally in a way that preserves that essence.
I believe I am seeing in other people the effect of AI usage in the form of reduced mental engagement.
Generated PRs, comments, Slack messages, etc., that tell you the person is not paying enough attention, or does not fully understand and cannot spot the flaws, etc.
Hey, Rails peeps. Security updates have been released, this is a big ol' batch of fixes including another Critical in image processing.
I strongly recommend you upgrade as soon as possible.
https://t.co/qs9zA77Rkz
Elixir has always been stellar, routinely having just a handful of issues which could not even be bugs.
Elixir was one of the examples I used in the section about software errors in my EuRuKo 2024 keynote https://t.co/xgBuZT3A8o
@HardcoreCap Not realistic in the sense that you look at the patch and no amount og AI harness and unlimited tokens produce a patch that does not need steering.
The fallacy being spread about code reviews goes like this:
Guys the future is to not have code reviews. Therefore, be prepared. Therefore, don't do them NOW.
No, now is not the future.
@HardcoreCap Correct!
My tweet is a counterpoint to the maximalists that categorically say you should not read code anymore.
That is just too dramatic, it depends!
I have generated apps whose source code I have not read, and complex projects where that is simply not realistic today.
@HardcoreCap Please, do not assume AI illiteracy.
If you have to review because the patches are not right, they are not.
It depends on the code base, not the culture or AI expertise.
@HardcoreCap Disagree. You do not have *adapt* to anything.
You should change when you read the generated code and it is right. If that happens today in your project, cool. Otherwise, review.
@lexrodba@levelsio There are people that have succeeded in life according to some metrics, but whose reasoning is so flawed so often that I am astonished.
Proof that to run a successful business you don't need to be a good thinker.
Last common trend for this club is bitching about Europe.