> Even if LLMs make mistakes, the ability of LLMs to deliver useful code and hints improved to the point most skeptics started to use LLMs anyway

Here we go again. Statements with the single source in the head of the speaker. And it’s also not true. The llms still produce bad/irrelevant code at such rate that you can spend more time prompting than doing things yourself.

I’m tired of this overestimation of llms.

Even where they are not directly using LLMs to write the most critical or core code, nearly every skeptic I know has started using LLMs at very least to do things like write tests, build tools, write glue code, help to debug or refactor, etc.

Your statement suffers not only from also coming only from your brain, with no evidence that you've actually tried to learn to use these tools, but it also goes against the weight of evidence that I see both in my professional network and online.

I just want people making statements like the author to be more specific how exactly the llms are being used. Otherwise they contribute to this belief that llms are a magical tool that can do anything.

I am aware of simple routine tasks that LLMs can do. This doesn’t change anything about what I said.

All you had to do is scroll down further and read the next couple of posts where the author is being more specific on how they used LLMs.

I swear, the so called critics need everything spoon fed.

You need to meet more skeptics. (Or maybe I do.) In my world, it's much more rare than you say.

My person experience: if I can find a solution on stackoverflow etc. the LLM will produce working and fundamentally correct code. If I can‘t find a already fullfilled solution on these sites, the LLM is hallucinating like crazy (newer existing functions/modules/plugins, protocol features which aren’t specified and even github-repos which never existed). So, as stated my many people online before: for low-hanging fruits LLM are totally viable solution.

I don't remember the last time Claude Code hallucinated some library, as it will check the packages, verify with the linter, run a test import and so on.

Are you talking about punching something into some LLM web chat that's disconnected from your actual codebase and has tooling like web search disabled? If so, that's not really the state of the art of AI assisted coding, just so you know.

But you have just repeated what you are complaining about.

Do you want me to spend time to come with a quality response to a lazy statement? It’s like fighting with windmills. I’m fine with having my say the way I did.

> Here we go again. Statements with the single source in the head of the speaker. And it’s also not true.

You're making the same sort of baseless claim you are criticising the blogger for making. Spewing baseless claims hardly moves any discussion forward.

> The llms still produce bad/irrelevant code at such rate that you can spend more time promoting than doing things yourself.

If that is your personal experience then I regret to tell you that it is only the reflection of your own inability to work with LLMs and coding agents. Meanwhile, I personally manage to effectively use LLMs anywhere between small refactoring needs and large software architecture designs, including generating fully working MVPs in one-shot agent prompts. From this alone it's rather obvious who is making baseless statements that are more aligned with reality.

> Here we go again.

Indeed, he said the same as a reflection on 2024 models:

https://news.ycombinator.com/item?id=42561151

It is always the fault of the "luser" who is not using and paying for the latest model.

This whole thing worldwide will lead to division in the society and to people having secret social media accounts. Only the most lazy and stupid ones will be handing over their private data to any gov security forces.

With so much browser fingerprinting and the biggest social networks being US companies, I wouldn't be surprised if already a company like Palanti has a dataset with all your history

They don't need fingerprinting, all the American social media has probably already been given a sealed FISA order to hand over account lists with ip access logs.

I feel like it's the same as the "I am not a terrorist" declaration check-box. You know your socials, they know your socials. They want to see if you lie.

What do you mean, secret? Your every action online is stored in massive data centres either operated by the US government or American mega corporations.

I can’t believe people STILL believe there is anonymity left online!

Unless you take some “enemy of the state”(great movie btw) level actions, they know everything worth knowing about you.

That's what I'm thinking every time I hear or have to use the term "AI". It is not intelligent, but everyone is so used to call it so. LLM is much better.

It would be great if all these companies contributed to a some kind of a unified modular platform like Project ARA. I see a lot of new devices, but they all do their own stuff. They produce hardware for their software, the end result is the same as with big brands. Most of these devices are usable while they are supported by these companies. Some of them allow installing custom Android roms, but not many.

Looks like the market just gets more fragmented without any improvements towards better sustainability/reusability. The only thing that really caught my attention recently was Pilet, a handheld Raspberry Pi. That's a really cool thing, that gives mobility while maintaining functionality.

I hope not. Projects like that to have any chance at surviving have to be good phones first. Adding modularity will make it worse in terms of specs, more expensive and in the result dead on arrival. Once they launch a few successful (or at least sustainable) products, they can maybe try doing some modularity

Am I the only one who just feels burnt out on these type of projects? We have a plethora of raspberry pi and other arm mobile developer kits that all just fail to deliver. They make great pet projects but fail at what most mobile phones do great which is provide a computer I can reliably and safely take with me in life. This pilet thing has 7 hours of battery life, is huge and will probably explode if I put it in my bikes bag.

While it's not perfect I've been investing more time into learning to live with grapheneOS. I can run Emacs and clang on the go. It's a better start that won't turn into a paperweight.

I'm not sure what you think Jolla is, but they have a track record of releasing phones that are good enough to be used as daily drivers. They are also targeting enthusiasts, but I've been using exclusively phones that run Sailfish OS (their main product) since 2014.

Sorry if my post is confusing I'm referring to the poster I replied to mentioning the Pilet which is a raspberry pi based project. Jolla phone I really can't speak too. It sounds closer to graphene where they understand the benefit of reasonable hardware quality and battery life.

Graphene requires a Pixel device. I can hardly call it accessible.

7 hours is not bad, considering my iPhone 13 mini can only last for day with occasional usage.

> and other arm mobile developer kits that all just fail to deliver

Librem 5 and Pinephone didn't fail to deliver.

The next sentence, which I read as the rest of the thought, was

> They make great pet projects but fail at what most mobile phones do great which is provide a computer I can reliably and safely take with me in life.

And I can't speak to the Librem 5, but the I'd say the pinephone did in fact fail to deliver a reliable daily driver, remaining merely a pet project full of rough edges.

Wow. I turned on the randomized tracks under the "Deconstructing 1/2" and it's beautiful.

Wishing these guys all the best. It's not just about following the market. It's about the ability to just be yourself. When everyone around you is telling you that you just have to start doing something and it's not even about the moral side of that thing. You simply just don't want to do it. Yeah, yeah, it's a cruel world. But this doesn't mean that we all need to victim blame everyone who doesn't feel comfortable in this trendy stream.

I hope things with the AI will settle soon and there will be applications that actually make sense and some sort of new balance will be established. Right now it's a nightmare. Everyone wants everything with the AI.

> Everyone wants everything with the AI.

All the _investors_ want everything with AI. Lots of people - non-tech workers even - just want a product that works and often doesn't work differently than it did last year. That goal is often at odds with the ai-everywhere approach du jour.

>When everyone around you is telling you that you just have to start doing something and it's not even about the moral side of that thing.

No, that's the most important situation to consider the moral thing. My slightly younger peers years back were telling everyone to eat tide pods. That's a pretty important time to say "no, that's a really stupid idea", even if you don't get internet clout.

I'd hope the tech community of all people would know what it's like to resist peer pressure. But alas.

>But this doesn't mean that we all need to victim blame everyone who doesn't feel comfortable in this trendy stream.

I don't see that at all in the article. Quite the opposite here actually. I just see a person being transparent about their business and morals and commentors here using it to try and say "yea but I like AI". Nothing here attacked y'all for liking it. The author simply has his own lines.

By victim blaming I meant some comments here. I can relate to the author, and the narrative that it's my fault for trying to be myself and keep to my ways triggers me.

I never found it comfortable to work with grids. The syntax and layout just feel off. Flexbox is a much more flexible and easy thing to work with.

There's plenty of overlap, but they solve different problems: flexbox when the content should control element sizing/fit, grid when the container should control element sizing/fit.

Another way to think about it: flexbox is for alignment of boxes in one dimension: horizontally or vertically.

CSS Grid is for two dimensional layout of rows and columns.

Back in the day, developers wanted page layout instead of the hacks on top of hacks with table-based layouts, floats and positioning to create layouts.

We’ve had CSS Grid designed for page layout on the web, in all browsers since 2017; as of 2022, only 12% of the top 1 million websites used CSS Grid, which to me is ridiculously low.

I use flexbox for grid purposes, simply because the syntax is straightforward and easy to read. Yeah, it’s one dimension, but if you nest it, it becomes two with no issues.

With nested flexbox the nested dimensions are not aligned to each other. With grid the items are aligned across both each row AND each column. With subgrid even nested grids can be aligned across nesting levels.

True. But so far I haven't faced layouts that I could not implement using flexbox.

I think the grid syntax is really off-putting - it’s one of the few places of “typical” CSS I tend to avoid unless something really calls for it. In my experience, it feels like the people most familiar with grid display, the syntax, and using it, are more on the design side than the programming side - most of my frontend peers who use it tend to misuse it when flex would work fine and be less rigid for their goals

But flex grow and align stretch exist, which moves control back to the parent...

A grid really feels like a list flexes to me too, functionally.

I agree. I occasionally turn to them to see if they work in a new setting, but find they never expose the features of a grid I would find useful. Everything must be manually placed, rather than allowing content to intelligently snap to multiple axes. Possibly I never have grasped some fundamental concept, possibly they are not suited to the sorts of layouts I usually work on. But more and more I feel they are designed to fulfil some purpose orthogonal to what I would need them to do.

I’m going to use it on iOS and macOS. Feels like the previous Safari.

On Linux I’ll keep to Firefox.

For me this feels irrelevant. These tools are marketed for developers for their day-to-day jobs that involve building products. Devs don't look up information on people or build some complex mathematical things daily. They build things that consist of different parts, which in turn can consist of different contexts and can be a combination of other things as well. It can be a straightforward approach or it can be a legacy codebase that also need to incorporate new features with new stacks. The real test is in the real world scenarios. But every time it's about a narrowly scoped thing, the tests, the marketing. And they try to build an image that the combination of these scoped tasks can somehow bring you the ability to build at large scale. They don't say it, but they implicitly mean it with the way they present all this. Computers can compute, they can detect patterns and do analytics part, they can build assumptions based on the data they have. But they need the data, they need parameters, they need not only an operator, they need the source for the material they base their computations and output on. And somehow all the marketing completely ignores this fact. And this is damaging.

This is amazing. Once a company that built its reputation on quality and performance wants to focus on that. iOS became a parody of itself. 26 feels like a cheap knockoff of an iOS. Hundreds of people get paid to enshitify experience for millions of users. That’s incredible. And then they simply say that yeah you know it’s time to work on quality and performance. Isn’t it what always had to be there in the first place? Ridiculous.