HN user

tvbusy

146 karma
Posts0
Comments43
View on HN
No posts found.

I don't think the prompts that the author has proposed will actually work. Including final scope and non-scope is good but it's more of a reaction of what the AI already did. These prompts are suitable for a rewrite, basically, since it's unlikely anyone would have had these ready when they start out.

I have found small iterations to have the best results. I'm not giving AI any chance to one shot it. For example, I won't tell it to "create a fleet view" but something more like "extract key binding to a service" so that I can reuse it in another view before adding another view. Basically, talk to the AI as an engineer talking to another engineer at the nitty gritty level that we need to deal with everyday, not a product person wishing for a business selling point to magically happen.

On the positive side, LLMs are trained based on real data so the default is for it to tell you what data showed. Companies will certainly enforce their influence but it's extra effort against the enormous amount of data, just like with trying to censor sensitive topics. Any context used for ads means less context for the user to use which in turn negatively affects their usefulness.

I understand that most people want to move to other more modern tools, it's up to you. However, what baffled me is why the author's choice not to move is a problem? Did we pay them to move and they did not move as promised? Was there some crowd funding to move that was not fulfilled?

Phone sellers should enforce a mandatory 30 days no use after purchasing to ensure that people are not harmed by phone usage.

Newspapers should only report news at a minimum 7 days after the fact to ensure accurate reporting.

Toasters should lock everything in until it's completely room temperature to avoid accidental burns.

These are serious problems.

LLMs have randomness baked into every single token it generates. You can try running LLMs locally and set the temperature to low and it immediately feels boring to always have the same reply every time. It's the randomness that makes them feel "smart". Put it another way, randomness is required for the illusion of intelligence.

If you already have Home Assistant running, I think it should be simple. Most of the time you can buy devices with pins already soldered and it's just the matter of connecting them together. AIs are pretty good with ESPHome configs. You can even take a picture so that they can help you identify the correct pins. Some coding may be required for drawing things on the display though.

Totally. Frameworks also make it a lot easier for new team members to contribute. React, for example, makes it a lot easier to hire. Any project with moderate size will require some kind of convention to keep things consistent and choosing a framework makes this easier.

Now look at the cross team collaboration and it gets even harder without frameworks. When every team has their own conventions, how would they communicate and work together? Imagine a website with React, Vue, Angular all over the place, all fighting for the same DOM.

You're thinking of open source projects that only need a few hours of work per week. Anything more than a few hours a week either requires someone to give up their full time job to work on it (switch to part time jobs/consultation is an option), or having multiple contributors which still require significant effort to coordinate at the end of the day.

Let's say Tailwind CSS gives up and stops the project, do you think there'll be someone else picking it up, knowing how Tailwind failed in the first place? LLMs don't create new things, they remix what are already available. It's delusional to think that you should use LLMs to create a whole UI library just for your application and spend enormous effort not only maintaining it but also train new team members to use it down the road.

Open source is charity, it's unreasonable, even entitled, to demand someone work on it full time without pay.

Same here. I developed desktop applications for 15+ years and was really frustrated with Microsoft's direction for the UI around the time of Windows Phone. While Windows Forms may not be the best, it worked for decades until then. Now even if someone wants to build a desktop application using native UI, it's next to no resource at all because it's all about cross platform nowadays.

Sounds like me: 1. For new UI/tool, I depend on text to navigate. 2. Once I'm more familiar, I scan using icons first then text to confirm. 3. With enough time, I use just icons. 4. Why the ** do they keep moving it/changing the icons?

The model I have seen is that there's a cost for charging as well as parking after charging has completed. You get 20 minutes after charging has ended for free but there will be a blocking fee afterwards as long as the cable is still connected to the car and will be charged as part of the charging fee. If someone charges their car and then disconnects the cable and blocks the spot, they will be towed as the spot has a notice that it's only for charging devices.

I have seen those attached to light posts here in Germany but they all are in front of public area, not private houses so I guess it's only for charging, not parking.

In another point, if it's in front of a private property and the owner plan to occupy it for themselves, it would not make sense to install charger there as even when the owner is not charging, no one else can use.

I understand the post is about learning to speed up SHA1 calculation, that I have no comment. However, the state file is a solved problem for me. It's a rare case where state files are corrupted and it's simple to just re-check the file. I cannot imagine a torrent client checking the hash of TBs of files for every single start. It's not a coincidence that many torrent clients have a feature to skip hash checking and just immediately assume the file is correct and start seeding immediately.

I still remember how sad I was when my third Pebble died. It's great to hear that Pebble is back again but the market has certainly changed. There was nothing like Pebble last time. Now, while not as good as Pebble, there are still alternatives. I'm using Amaze Bip, for example.

I hope the new Pebble will open up more hacking opportunities. Zigbee devices have been getting really tiny nowadays, so having a watch that can send a Zigbee signal when tapped on an NFC tag will be so much more convenient than pulling out my phone. LTE presence detection is another.

It's corporate talk for "we f up and systems are failing". Musk's act of not paying people now bites him back. I guess they don't have enough servers to sustain the service.

It's hard to say. What's the difference between Reddit and 4chan? Technically there's very little. There's nothing Reddit does that 4chan cannot do, technically. The difference is the community, the people dynamic. It may very well just blow over and stay the same, or Reddit may slowly rot like Digg and MySpace did.