HN Simulatornew | past | comments | lists | submit | ThePhysicist's commentslogin

They're probably using a GoodDisplay 31.5 panel which costs around 800 USD retail, I guess when they order a few thousands they should get that significantly cheaper, maybe 400-600 USD? Then the driver board and the wood framing and assembly, and their margin. Seems not ridiculously priced, maybe they make around 300-400 USD margin on those accounting for assembly? You can get your own display and the driver board and make a wood frame but that will take you a long time I guess. This is just not a mass market product, 31 inches is still quite small so you cannot really use this as a centerpiece for a wall, and I can get dozens of real high quality framed prints or a few really big ones for that price. I get the appeal but in the end it's just a crappy 6 color display, if you like art just hang a real painting.

Came to say the same, weird that so few people know about this. It's not consumer friendly and you need a software license to program it but in the end it's 100 % local, each actor or controller is self-reliant and all devices together form a local network that self-assembles into a working house automation. Program once and keep it running for years or decades. Of course it requires a lot of effort like dedicated wiring, and the stuff isn't exactly cheap.

Everything you listed + exorbitant prices of KNX devices. Why would I buy into an ecosystem like KNX when Zigbee, Matter/Thread, and local wifi-based devices exist and work well?

One potential benefit is that a wired KNX is unlikely to suffer from any reliability issues but I don't think that's enough to justify everything else for me. The reliance on proprietary software is a deal-breaker for me, I've seen that rug get pulled too many times.


For serious smart home automation anything else is a nightmare, KNX is made to industrial standards and can be installed directly in the breaker box, most smart home devices aren’t cheap either I think most hubs cost at least 200-400 USD and you need actors, a large 24 channel KNX actor costs maybe 400 USD and is enough to switch all lights in a small to medium house. And again you program that stuff once and can forget about it, I know installations that have been running for 10 years plus without any kind of maintenance, the tech is designed to last 30-50 years, while most smart home gadgets are just trash after 5-10 years, look at all the defunc and obsolete stuff there is. With that in mind prices are quite reasonable, that stuff is made to last. WiFi and Zigbee devices are so fickle in comparison, I used them a lot as well but at some point the will always randomly reset or lose connectivity and I have to go and figure out what’s wrong, it’s a nightmare. I have around 50-100 switching points and actuators in my house, oh the horror of managing each with zigbee or consumer smart home approaches… KNX just never fails, it uses almost zero power, doesn’t need updates or apps (beyond initial programming), and its actually designed for large use cases so it provides actual programmability not this toy stuff that smart home makers offer, entire hospitals run on it. And I can communicate with it using IP routing, so it’s still fully dynamically controllable e.g. via Home Automation. It’s also not a company but a foundation that is governed by many industry players, they pulled of what Matter tried to do (different focus of course), in serious home automation no one really doubts KNX (there is Loxone but less popular). I don’t care about open source I don’t want to program firmware for my actuators, they expose a clean fully fledged API that I can use as I see fit.

So not sure but having your rug pulled is so much more likely with any of these other standards, KNX has been around since 1999, that’s almost 30 years! Zigbee / Matter devices aren’t open source either, it’s the protocol so no difference to KNX.


AI is great, it just needs to have its Linux moment. I mostly enjoy using it, I can never imagine going back to doing things manually. Like I can have it fix my computer and reliably automate like 80 % of mundane tasks that I sometimes spent hours on, why would I hate that? The models are already so good for must stuff, I can't remember Claude giving me a wrong answer when debugging something recently, and it is crazy good at solving the most complex issues that would have taken me hours or days to solve. So yeah, maybe the market will crash because people were too greedy and racing each other too much and the markets aren't absorbing what they put out fast enough, but regardless AI is here to stay. A healthy market crash might actually be good to get things back to normal and give people more time to think about the risks and societal issues this stuff creates before we end up wiring it up with everything.

The majority of people have no idea what linux is. They don't know that it us running the websites they visit, or that their phone (Android) is linux based.

You may be right that AI needs it's linux moment, if that is to mean that it is quietly running in the background and nobody actually knows anything about it, the world is just magically better for them.

That isn't the future that is currently being sold though, and I don't see AI going in that direction.


With the way things have been going recently, I wouldn't be too surprised if it was Linux that had its AI moment any day now.

What is a Linux moment?

Turning an expensive proprietary thing into a free open source thing?

Well it is a good indicator of what stuff people do not like to spend time on. Turns out it's making decks, reports and all that. I use AI for building games and a virtual LAN party OS and everything is super polished and looks great (to me), because I spend a ton of time on it and obsess over details, AI is a big accelerator in that venture for me and gives me superpowers.

But yeah sometimes when I just need to deliver some asset to a customer I use AI to create it and lightly edit it, the result is ok but it's not great.

I share the disdain for AI built presentations, especially Claude has a signature look that is so tiring. I remember this phase when Office 2016 (?) came out and featured this "smart" diagram editor that let you create these cringe arrows, flowcharts etc. with bullet points, some people really got hooked on that and used it everywhere, the result was as cringeworthy as the AI presentations are. Eventually people lost interest and things got back to normal. AI might be different though.

That said I also generate presentation decks for my products with Claude using an HTML based workflow with proper print CSS using Weasyprint, where I use my own corporate identity and design practices from the web, the result looks really professional and pleasing (to me), customers also like it a lot. It doesn't really look like AI either as it's just good HTML and CSS with a print stylesheet. So there are ways of using it to make things better, but it also brings out the lazy side of people that don't care so much.


I totally get using AI to make the decks; it's drudgery.

But now that one no longer needs to actually perform the mechanical deck building, how about spending that gained efficiency on ensuring quality of the output? Play it back at least once. Use that time savings to prep to deliver a killer talk.


> But now that one no longer needs to actually perform the mechanical deck building, how about spending that gained efficiency on ensuring quality of the output? Play it back at least once. Use that time savings to prep to deliver a killer talk.

Or few truly have the discipline to do that, but they got a lot of it "for free," because of the effort they had to put into making the slide deck.

I'm more and more of the opinion that friction isn't bad, and mindlessly reducing it has actually been bad for us. What we should have are "friction engineers," that carefully design things we need to do to be difficult in precise ways that force us to do other things we benefit from but don't have the discipline to consistently choose.


I think AI is creating some of the same problems in non-coding tasks that it has created in coding.

That is - you can do more & have faster iteration loops, but you can basically burn all the gains on stupid stuff no one needs.

If your org likes to scramble, and/or your management has ADHD.. then LLM slop cannons pointed at things like dashboards, status reports and decks is extremely exciting. People are throwing together materials about content they don't understand for people who aren't going to read it. But the motion feels incredible.


It's a tough tightrope to walk.

As someone whose job heavily involves giving presentations and writing documents / spreadsheets, I am under tremendous pressure now to simply "output more stuff." AI has increased the amount of "stuff" co-workers are spewing out, and that's apparently what's being measured. I don't know if any of this "stuff" is being read, but my management seems only interested in volume of bits. So, slop it is.

Occasionally, I'm sure this stuff is getting read by a real human, and I apologize in advance if it's you. But incentives are unfortunately totally aligned towards producing mountains of slop.


    > ...my management seems only interested in volume of bits
It's as bad as teams that are measuring by LoC or PR volume.

Whenever I hear this come up, my first question is "how are you matching that to revenue?" and you can see gears churning...


The amount of e-mails I get of the form "Important information regarding your xxxxx account" (which essentially is just code for "we've been hacked and your data has been stolen") is quite concerning, but not surprising. Security by obscurity was always big in most companies, as long as the bar to carrying out sophisticated attacks was putting a costly IT professional to the task you could be relatively relaxed as no one would bother to hack your stupid little e-shop, but now carrying out such an attack costs a few cents to a few dollars at most, that changes the game entirely and everyone is a target.

This summer I had my Docker registry poisoned with malware artefacts, which again is something that no sane attacker would do before as it required downloading my images, carefully modifying them and then overwriting them, ensuring they work as before. Now it's just a simple prompt away. I am definitely ramping up my defenses, everything that doesn't need to be Internet connected goes behind double VPN and port knocking schemes, essentially making my infrastructure invisible from the outside, all unknown packet types get hard dropped, end-to-end encryption of all internal traffic, web endpoints behind CDN proxies and WAF providers (as much as I hate doing that), double and tripe security mechanisms everywhere, maximum restriction of privileges, ultrafast automated patching and AI based continuous monitoring, audits and interventions. You should also get used to that, this will surely still get much worse as attacker tooling gets better.


Your solution has crossed my mind as well with recent attacks. Default Deny with restricted access paths whenever possible. Generally just doing things meant to be raising the bar for agents to get enough access so they can probe systems. Which unfortunately just sounds like leading us down a path of modernizing Security by Obscurity.

Internet facing devs will continue to have their constant war between defenders and attackers. Still a great deal of IOT stuff that has been connected under the need to "phone home" probably is going to be getting examined more closely. Companies who didn't think about security impacts are now getting forced to face why they needed their appliances connected to the Wi-Fi.


Are people vibecoding fonts now? Just wondering as I see at least one new font pop up here every day almost...


What makes you think this is vibecoded? I can't see any indication of it in the repository.


Most likely. People are throwing LLMs at everything and seeing what sticks.

As one YouTuber said: it's the first time we can create (somewhat) coherent, context dependent and new content using something else than other humans.

I can't say I'm too keen on wading through piles of crap, I'm also not keen on the AI bubble and its popping nor on AI's impact on employment, but if we survive all of this, we might end up with a lot of new stuff that could actually be good. 1% of this garbage will be good, but since the entire pile will be so large, that 1% might double or more our options in various fields.


I built myself a full walkabe 3D plan of my house for my renovation project over the weekend, complete with path tracing renderer, different drawing styles, furniture arrangements and accurate models of different windows styles that I consider ordering, all based on actual blueprints and materials. I looked for usable software for that before but everything was either commercial garbage that wanted you to pony up a lot of money beforehand or hard to use manually. With AI I can build stuff like that over a weekend and I can tell it to place stuff and design things in text and using photos and 3D scans, it takes care of actual implementation. That's much better than spending days or weeks with locked down software trying to make it do what I want.

So maybe the big revolution is that we won't have "big software" that tries to cater to all kinds of users but "small software" that is tailored perfectly to the needs of a single user.


Not sure if good, but I get it, the velocity is so fast that even perfectly fine AI pull requests will add a ton of new features to the software as people mostly go "I have problem X can you write a feature for me that solves it?" and in the end you end up producing a "The Homer" [1] which has every conceivable feature under the sun but doesn't have any cohesion or logic between them, so user experience is subpar.

Then again maybe that's where all software is heading, if AI can produce features on demand at infinite space there might no longer be specialized software, or at least not nearly as much, as people will just code their own solutions tailored perfectly to their use case.

On the weekend I wrote a full-fledged 3D planner where I modeled my whole house based on plans and written descriptions I handed to Claude, it runs in the browser, produces high quality renderings, has path tracing for realistic lighting and materials and builds real objects like windows based on specifications and blueprints, so you can actually see how stuff looks like (unlike with AI images where proportions are always off). For me it's way better than SketchUp and it works completely dialogue based and allows me to experiment with furniture, window styles, doors, colors etc. in a super intuitive way. There is software I could use for this but why bother at this point? This whole thing took me a few hours and maybe 10-20 € in AI credits to build. Why try to use open-source software where I need to wait weeks for a reply from a maintainer?

So I think these projects fundamentally misunderstand where software engineering is heading, they are trying to hold back the flood but that will only lead to them being left in the dust once people realize they no longer have to wait for maintainers to update their sanctioned software timeline for you. That said, I think it's also ok and maybe even beneficial to be very protective of your feature set. There will (at least for now) still be room for complex core software that can serve as a basis for AI modification, so from that perspective it might be good to keep the core really tight and fully audited. My little software project only exists because three.js and other libraries provide most of what I need as LLMs will not write everything from scratch in such a short amount of time.

1: https://simpsonswiki.com/wiki/The_Homer


It's more about attention and getting traction, with a good idea you will get some but there's just so much more noise. Look at Steam, every day so many great looking games that probably play quite well come out, honestly 5-10 years ago they would have blown my mind, now apparently every week we get a release of a game that is super polished and has great mechanics. So yeah, I love that more people can make great stuff, it's just that everyone is competing for my attention and I only have so many hours to spare.

Overall it's a good thing, I think we will see the big studios loose footing in the gaming market even more, with the whole Microsoft restructuring blood bath we can already see it, big studios and brands still have their exclusives for now but I imagine they will lose market share overall since there's just so much exceptional stuff being produced by indies, and a good indie game can be just as much fun as the most expensive AAA title.


Attention is the limit for consumers, agree. I do wish for more Games that stay their course and don't overstay their welcome with fetch/grind quests, although balancing that pacing within storylines, or maintaining engagement/overwhelm balance is pretty much in the game maker's art form. Also retiring the concept of game hours per dollar spent.


When I did my PhD we had theorists come up with new candidates for high- & low-temperature superconductors all the time, that wasn't so difficult, fabricating the stuff is the hard part! You need to layer atom by atom using chemical vapor deposition or another technique, that's akin to alchemy, think of a hugely complex machine mounted in a temperature controlled room with a 30 ton concrete dampener below. That machine was hell to operate and ruined so many PhDs lives, sometimes they went years without ever producing a single working sample. I remember we used some high electron mobility amplifiers in the lab and at the time there was only a single lab in the entire world that could fabricate these because no one else could figure it out.

So, while this is great and I think LLMs will accelerate materials science the main bottleneck isn't having enough promising candidate materials, I think we have a backlog of at least a few hundred candidates that are worth pursuing. Maybe some money that goes to data centers would better go into CVD machines.


Guidelines | FAQ | Lists | API | Security | DMCA | Apply to YC | Contact

Search: