Exactly. Which just spins you right back to where you started before AI.
Needing good people that keep you on track while protecting the quality of the system from degradations.
And minimal amount of people doing the reverse.
Personally, I think it will hurt the later more than help the former.
Scaling a department is hard BECAUSE it increases output and output can create momentum in the wrong direction just as easily as the right direction.
AI can accelerate how fast people pull things in the wrong direction too, which creates drag for what matters.
Yeah. The real bottleneck is knowing what to build and what NOT to build.
It isn't the rate of building itself.
Similar to how hiring more engineers often won't speed up a project.
The more wrong things that have been built, the lower the chance a human (or ai) can pick the next right thing.
It's really just people conforming to internet patterns that garner attention.
To be blunt, if a technology is not something I can get up to speed on in a week or two, it is far too complex to survive in the mainstream.
Sadly I would agree with this.
Both design and dev have been plagued with:
1. People who learned to design/create for the web medium without first learning how to identify quality.
2. Career rewards given for things that make a flashy slide/presentation rather than boring quality/a11y/efficiency.
I think web has always been like this. Probably because people creating websites have a feedback loop often tied to different incentives than the underlying quality of their output.
Definitely a wasteful choice, agreed.
I am just being annoying, but wanted to say that using the robust tool (Radix) for a simple usage problem is common elsewhere.
And using a wasteful solution (SVG) vs CSS for the radio is also something I have seen outside JS frameworks too.
Radio input helpers from things like Radix, Ariakit, etc tend to exist to assist with more difficult visual representations of a radio input.
This particular case just seems to be the unnecessary usage Radix for a simple visual?
I feel you, but this type of things happens beyond React too.
I was just considering this exact shift as well.
Our recent renewal went horribly due to it locking out members on my family plan for days for no reason. And customer service is almost non-existent.
They couldn't even rejoin the plan because they were "already on it".
Nice, beat me by two years. You must have so much more wisdom to share.
If you don't have a decent grasp on what people will appreciate, be it a PR/website/food/art/etc, how can you achieve anything worthwhile with just the skills to produce?
This is a little more generic, but I have often felt that education should be more focused on teaching people to identify quality before teaching them to produce things.
Be able to identify aspects of a quality website. Then learn to build them.
Identify aspects of a quality PR, then make them.
It could actually be helpful, but it requires 1-2 good employees focus handling maintenance and improvements.
Which should still be worth the salary money, but executives tend to always think things don't require ongoing investment after launch.
In my anecdotal experience, the issue tends to come from the fact that companies often treat the bot like a time boxed project.
Buy it, set it up (maybe have consultants involved), celebrate the launch, and remove all engineers from working on it.
No one can predict the future or speak for all users and use cases. I certainly cannot.
All I can say is that the use cases I have seen in my career do not align with what you describe. Nor do many of the other examples shared on this thread.
I respectfully disagree. Solving a common issue the community has would not upset it due to the fact they could technically write CSS with this feature that didn't display as intended.
People do that everyday with CSS authoring, and they iterate to reach their goal. They would fix your example.
👏🏼
Basically forking the style calculation, some duplicative cost, but that is opt in by using this. Similar patterns are used in other domains.
Couldn't it be technically feasible to have two levels of detection? If it was actually chosen to be implemented.
Wrapped styling being based on the wrap detection of the base styles. Not the resulting wrapped state after they are applied?
If I am not mistaken, feel like I also recall people wanting the wrapped items to have lower flex growth. So they are more visually cohesive with the items above that have reached their max compression and caused the wrap.
This would require wrap detection not just for the items but also the flex container.
I feel like in the past I have seen designers want the alignment to alter on wrap to avoid center aligning a trailing item, giving it undo prominence.
Wasn't this the same argument against container queries for so long? That they couldn't exist due to the challenge of preventing an infinite loop?
It was solved for that. Probably lessons there.
Not for CMS driven navs that are localized.
You don't always want a fixed number, you just want more control on how things adapt to newly added items or translation expansion.
Gamification and the illusion of participation.
Removing it would bring a new set of challenges. I could be wrong, definitely have been before, but I think those new challenges would be healthier ones to have.
You could use a polling feature to maintain an outlet for gauging common sentiment.
It often makes me wonder if there could be a viable social media model without the upvote/like, just pure conversation.
Seeing the negatives of the downvote makes one wonder what the unintended consequences of the upvote are too.
Reddit is rough... Downvoting as a feature seems to encourage the spiral to a hive-mind echo chamber.
People get swept up in the up/down momentum and stop feeling the need to think about things for themselves. They can just turn off brain and join the majority.
It always makes me recall the Shakespeare vs Bible anecdote from the Dale Carnegie book.
To pick what you say based on the likely outcome of saying it, not on the instinctual desire to say it.
A simple but powerful thing to attempt to embrace. So rare to see in people though.
So many variations of this play out all the time.
Person A: Expresses enjoyment towards something.
Person B: Poisons the moment with a negative take on it that is not at all constructive.
The silver lining I guess is that it reminds me to strive to not be so careless with my words.
The fastest end having lowest conversion rate is almost certainly related to the type of client/scenario that produces that data rather than the idea that the site spooking them.
Or the outcome of rendering something like 404 text rather than an actual product image having faster LCP.
