What are your thoughts on this 2023 comparison?

So to confirm, you don’t trust blogs where the company is selling a product or service, even if they don’t mention it in the article? If so, that would cover a lot of articles shared on this instance.

For what? I don’t see any products or services being promoted in this article.

People who care about performance are using loops

Well that depends, generators are faster than loops when you’re using Bun or Node.

The conclusion aligns with my own belief, which is that it’s better to create a minimal context by hand than get agents to create it:

We find that all context files consistently increase the number of steps required to complete tasks. LLM-generated context files have a marginal negative effect on task success rates, while developer-written ones provide a marginal performance gain.

When I have got Claude to create a context, it’s been overly verbose, and that also costs tokens.

However in this case the opposite is true, as Chromium currently doesn’t support this feature.

Do you mean features only currently available in Chrome?

Well spotted, the article states:

The :heading pseudo-class is currently available in nightly builds only. You can test it now in:

• Firefox Nightly (behind a flag)
• Safari Technology Preview

There are some really good tips on delivery and best practice, in summary:

Speed comes from making the safe thing easy, not from being brave about doing dangerous things.

Fast teams have:

• Feature flags so they can turn things off instantly
• Monitoring that actually tells them when something’s wrong
• Rollback procedures they’ve practiced
• Small changes that are easy to understand when they break

Slow teams are stuck because every deploy feels risky. And it is risky, because they don’t have the safety nets.

It"s working now

I think there’s many solutions to this, including setting a minimum account age to accept pull requests from, or using Vouch.

There’s a great rebuttal to Shumer’s post, Why I’m not worried about AI job loss.

Guys, can we add a rule that all posts that deal with using LLM bots to code must be marked? I am sick of this topic.

How would you like them to be marked? AFAIK Lemmy doesn’t support post tags

I try to stay well read on AI, and I regularly use Claude, but I’m not so convinced by this article. It makes no mention of the bubble that could burst. As for the models improving aren’t the improvements slowing down?

More importantly, the long term effects of using AI are still unknown, so that for reason the adoption trajectory could be subject to change.

The other factor to consider is that the author of this article is a big investor in AI. It’s in his interest to generate more hyperbole around it. I have no doubt that generative AI will forever change coding, but but I have my skepticism about other areas, especially considering the expensive controversy of Deloitte using AI to write reports for the Australian government.

I find articles like this about agentic engineering both compelling and anxiety inducing. It’s staggering what people are achieving with agents toiling away in the background. However, it also leaves me feeling insecure, as I have many ideas that I would love to build, and previously my excuse was a lack of time, and now I worry I have no excuse when an agent could potentially build it whilst I sleep.

What I’m saying is the post is broadly about programming, and how that has changed over the decades, so I posted it in the community I thought was most appropriate.

If you’re arguing that articles posted in this community can’t discuss AI and its impact on programming, then that’s something you’ll need to take up with the moderators.

In fact, this garbage blogspam should go on the AI coding community that was made specifically because the subscribers of the programming community didn’t want it here.

This article may mention AI coding but I made a very considered decision to post it in here because the primary focus is the author’s relationship to programming, and hence worth sharing with the wider programming community.

Considering how many people have voted this up, I would take that as a sign I posted it in the appropriate community. If you don’t feel this post is appropriate in this community, I’m happy to discuss that.

My nuanced reply was in response to the nuances of the parent comment. I thought we shared articles to discuss their content, not the grammar.

Regardless of what the author says about AI, they are bang on with this point:

You have the truth (your code), and then you have a human-written description of that truth (your docs). Every time you update the code, someone has to remember to update the description. They won’t. Not because they’re lazy, but because they’re shipping features, fixing bugs, responding to incidents. Documentation updates don’t page anyone at 3am.

A previous project I worked on we had a manually maintained Swagger document, which was the source of truth for the API, and kept in sync with the code. However no one kept it in sync, except for when I reminded them to do so.

Based on that and other past experiences, I think it’s easier for the code to be the source of truth, and use that to generate your API documentation.

There are plenty of humans using em dash, how do you think large language models learnt to use them in the first place? NPR even did an episode on it called Inside the unofficial movement to save the em dash — from A.I.