Please don't post shallow dismissals or focus on trivialities. Let's focus on the substance of the project. It's a good sign that an early stage project hasn't devoted a lot of time or cost into their logo.
3,582 karma · joined October 30, 2007
---
Please send HN questions to hn@ycombinator.com.
If you wouldn't mind reviewing https://news.ycombinator.com/newsguidelines.html and taking the intended spirit of the site more to heart, we'd be grateful.
I actually think another benefit of AI is that it can make it more realistic for people to self-educate and get further towards getting devices working without needing free help from forums. How they get help when they get stuck, without pestering forums like yours for free help is an interesting challenge.
I still think the biggest upside of AI in electronics is for people who are experienced and competent, being able to get the laborious design/layout work done more quickly, get errors found sooner, get their designs right in fewer iterations, and get feature-rich firmware written and debugged faster. If true, that alone can mean more high quality electronics products hitting the market. That’s the point I was making. Nothing to do with newbies freeloading off experienced engineers on forums.
Thanks for the discussion.
They're Made Out of Meat (1991) - https://news.ycombinator.com/item?id=43994603 - May 2025 (3 comments)
They're Made out of Meat (1991) - https://news.ycombinator.com/item?id=43883866 - May 2025 (1 comment)
They're Made Out of Meat (1991) - https://news.ycombinator.com/item?id=40152901 - April 2024 (1 comment)
They're Made Out of Meat (1991) - https://news.ycombinator.com/item?id=38420111 - Nov 2023 (168 comments)
They're Made Out of Meat (2005) [video] - https://news.ycombinator.com/item?id=35848313 - May 2023 (5 comments)
They're made out of meat (1991) - https://news.ycombinator.com/item?id=31965062 - July 2022 (151 comments)
They're Made Out of Meat [video] - https://news.ycombinator.com/item?id=30477257 - Feb 2022 (2 comments)
They're Made Out of Meat (1991) - https://news.ycombinator.com/item?id=24737993 - Oct 2020 (292 comments)
They're Made Out of Meat [video] - https://news.ycombinator.com/item?id=23436550 - June 2020 (4 comments)
They're Made Out of Meat (1991) - https://news.ycombinator.com/item?id=22836278 - April 2020 (1 comment)
They're Made Out of Meat - https://news.ycombinator.com/item?id=22108726 - Jan 2020 (1 comment)
They're Made Out of Meat - https://news.ycombinator.com/item?id=11561522 - April 2016 (3 comments)
They're made out of meat - https://news.ycombinator.com/item?id=8910420 - Jan 2015 (1 comment)
They're Made out of Meat - https://news.ycombinator.com/item?id=8152131 - Aug 2014 (170 comments)
They're made out of meat - https://news.ycombinator.com/item?id=8098264 - July 2014 (1 comment)
"They're Made out of Meat?" Short first contact sci-fi story - https://news.ycombinator.com/item?id=3549320 - Feb 2012 (62 comments)
They're made out of Meat - https://news.ycombinator.com/item?id=774139 - Aug 2009 (3 comments)
> If you made the PCB then you should be well placed to fix issues with it. You don't need years of experience , you have the schematic. A microscope (if you need it) is 40 USD. A logic analyser is 10 USD. Solder paste and hot air is only for specific parts and that is 30 usd.
This is obviously at the extreme low end of the amount of experience and cost/quality of equipment you need to do serious work. Decent oscilloscopes or logic analyzers alone are several hundred dollars, and I didn’t even mention a programmer module, which is $500+ for a licensed J-Link. And you need to be a proficient C programmer.
> I am assuming 2 layer - 6 layer PCBs are very specialised (e.g. drones)
I am talking about 4-6 layer multi-MCU boards for reading from multiple different types of environmental sensors, connecting to multiple different network types (Bluetooth, cellular, satellite) and supporting multiple different power options (lithium battery, 12V DC, USB, solar). Painstakingly minimized power consumption, redundant data storage. Analogue front end for impedance readings of legacy soil moisture sensors. It’s a lot to get right together and to all work reliably in outdoor/remote contexts. But plenty of products that could be useful to reasonably large numbers of people (e.g. advanced home automation/robotics) would need to be comparably sophisticated.
Elsewhere in the thread you’ve written:
> PCB design is not that hard. People trying to offload this to AI are also unwilling or unable to put in the work to actually diagnose issues and finish the project.
This is a dismissive value judgement against others whose circumstances you don’t know about. It’s clearly hard, relative to other kinds of technology work (web/desktop software etc), and not worth it if you can’t get to volume production and sales.
So, sure, you can learn simple PCB design and get a simple 2-layer board in your hands in a month. Degree-qualified electronics engineers and companies obviously exist to provide the rest of the capabilities needed to get a serious product built and all the way to market and volume sales.
It’s clearly not always as effortless as you’re making it out to be in order to justify the sneer.
To test and rework a surface-mount PCB, you may need:
- logic analyzer or oscilloscope;
- microscope;
- high precision soldering equipment/ability;
- solder paste and oven;
- cutting equipment to cut through 4-6 PCB layers and connect tracks with hair-width wire;
- Years/decades of education and experience to understand what to look for and what to do when you find a problem.
Then you have to write the firmware. That’s weeks or months of work. (And much of the embedded firmware code I’ve seen come out of small business engineering departments or consulting shops is terrible, because they don’t have the time/budget or organizational structure that delivers clean, readable, well-structured, well abstracted C code.)
All this is before you get an MVP into the hands of your customers. Then you start getting customer feedback and have to start making more improvements.
“Never been easier” it may well be to get a hobby project prototypes. But for a product with any amount of serious functionality, it’s at least months and likely 1-2 years to get anything into volume production.
But the potential for saving huge amounts of time particularly on design/layout, pre-manufacturing error detection, and production of well-structured and debugged firmware code are huge.
The limiting factor in electronics/PCB development and manufacturing isn’t engineering expertise (there are plenty of sufficiently competent electronics engineers in the world, especially in Asia), nor parts/materials or assembly costs. It’s the endlessly laborious grunt work of testing a design and making minor improvements until it is finally ready for volume production. That process can easily take a year or two even for a fairly simple design. Even a really skilled/experienced engineer can make a simple mistake in a schematic or PCB layout, which only gets picked up weeks later when prototypes are assembled and tested, or even later when a batch has been made and shipped to customers. So, each design-assemble-test cycle takes weeks/months, and each testing/diagnosing/repairing session takes hours or days. And it mostly has to be undertaken by degree-qualified engineers or experienced technicians.
And that’s mostly before you’ve written all the firmware, which is incredibly laborious in its own right.
The outcome is that only very large companies with big sales volumes can afford to do it (or niche players able to sell at huge margins, most commonly in defense or medical industries).
Outside of those sectors, it’s now just not really viable to be a small-medium business selling moderate volumes of units anymore, the way it was in the 80s-90s. I’ve seen it up close in agricultural IoT. There are no big/dominant players (customer needs are too heterogeneous and distribution is so hard). There are many small vendors, but they find it hard to make products that work well and that customers love, especially at a price point most growers will go for. So, no vendors are big commercial winners and not many customers are very happy - even though there’s a great need for technology in this sector. And it’s really not because the vendors and their engineers are incompetent; it’s just really hard.
With AI to take care of so much of the grunt work involved with circuit/schematic design, PCB layout, pre-manufacturing verification, testing, and firmware dev, it’s realistic to me that you could cut down on R&D costs by 90%, make it much faster to get products onto the market, and make it possible for so many more high-quality niche products to be commercially viable.
That's a good point. We've banned the account.
It's still not great community participation, piggy-backing onto someone else's launch thread with these kinds of comments, and repeatedly insisting that weight loss via diet and exercise is easy (people have been saying it for decades and it clearly isn't, otherwise there would be no demand for GLP-1 drugs).
The case remains that your style of participation on HN is self-promotional. If you want to generate attention for your research and program/app, please earn the exposure by writing a deep-dive post that can engage the intellectual curiosity of the HN audience.
If you have material consisting of novel research that the audience will find intellectually gratifying, you're welcome to share it as a standalone post.
It's not good community participation to hijack other people's launch posts with this kind of antagonistic commentary when you have a competing agenda.
Edit: removed initial mistaken accusation of failing to disclose competing product.
In HN guidelines terms, these sections are relevant:
Be kind. Don't be snarky. Converse curiously; don't cross-examine. Edit out swipes.
Comments should get more thoughtful and substantive, not less, as a topic gets more divisive.
Eschew flamebait. Avoid generic tangents. Omit internet tropes.
Please don't post shallow dismissals, especially of other people's work. A good critical comment teaches us something.
While I have you: could you please have a re-read of the guidelines and have a think about if you really intend to participate here earnestly? We've asked you several times to lift your game, and we're not seeing any effort to reform. This place only continues to exist because we have guidelines and uphold them daily, and enough other people make the effort to contribute positively. https://news.ycombinator.com/newsguidelines.html
Comments should get more thoughtful and substantive, not less, as a topic gets more divisive.
When disagreeing, please reply to the argument instead of calling names. "That is idiotic; 1 + 1 is 2, not 3" can be shortened to "1 + 1 is 2, not 3."
Please don't fulminate. Please don't sneer...
Must we? Such vulgar swipes only weaken your position, making you seem like the bigger jerk. Please observe the guidelines, not because you owe Mozilla anything but because you owe this community and indeed yourself better than this. https://news.ycombinator.com/newsguidelines.html
If you wouldn't mind reviewing https://news.ycombinator.com/newsguidelines.html and taking the intended spirit of the site more to heart, we'd be grateful.