People like Byran live amongst us
Making their own laptops but from SCRATCH
Imagine how good this man’s pasta carbonara tastes
how much was it to get the case milled?
Looks like an MIT admissions portfolio project. Don’t know if it fits the uniqueness category for it but I guess the quality of the end product makes it good enough.
Admittedly this isn’t fully open source like the Novena or the Reform but I doubt adcomms care. I just wish I was rich enough and skilled enough to be able to spend $4.5k on a neat project like this.
Love the parts research you did.
Bryan is in his last year of high school.
</end>
Keep building!
There are some obvious next steps for improving the polish on this, would you say you were more resource constrained, time constrained, or skill constrained?
For instance, did you put any thought into making flex PCBs to make the cable routing easier?
I also think the concept of a laptop with a removable wireless keyboard is brilliant, and I think your implementation is a lot cleaner than e.g. the Surface or the iPad's case-keyboards. If I had a laptop that did that, it would be my go-to travel machine. One less thing to cart around.
Very impressed by what you have done here. Kudos to you on achieving designing and building a whole laptop!
What I've found is that it's a bad idea to use USB extension cables; these can introduce bit errors if e.g. you copy large amounts of data (order of terabytes). It's much better to insert a USB drive directly into a carrier board, but this is not always physically possible.
You did the smart thing there with the SoM (for the uninitiated: power sequencing to individual parts of an SoC and its external components is an epic hassle to get right and that's assuming you actually have proper documentation - without it it's an utter pain), but how in hell did you get the high frequency stuff working out on what was likely your first or second try? This is IMHO where your work really shines.
USB-C, DisplayPort (at 4K to boot) and PCIe at modern speeds are all but black magic to most, this isn't digital any more, this is good old analog circuitry and physics at work that most people don't even learn in university any more.
Doesn't matter why, pretty sick. I'm studying physics myself, so its pretty inspiring to see you do this
You are clearly a very clever person and you do not need a web app wiggly graph thingie to throw ideas together.
There's no need to gild a lily!
Please keep the faith - I love that you are focussed on being altruistic and sharing your skills to the benefit of everyone.
Thank you.
It's very impressive work and it makes me so happy to see real hacker news on HN. This is real hacking.
Wow! And I'm guessing if he attempts a 2nd edition, it'll probably be even thinner, lighter, and faster!
P.S. @Hello9999901 any relation to "Bunnie" Huang?
I just want to throw money at you! We need an open source laptop!!
i didn't find any firwmware in the repo (didn't look exhaustively) but I did find that the SoC this is based on is supported by https://github.com/edk2-porting/edk2-rk3588 .
AFAICT the azoteq trackpad has proprietary firmware, so if that's true then i won't call this laptop fully open source. but from a practical perspective, i am much less worried about that then the boot path.
love the keyboard, wish i could test drive it!
so instead, i was left very, very impressed!
And possibly the year before.
Well, well done. Good luck to you!
On a less joking note, I wonder if I'm decent at Factorio, I could learn this.
Just joking, incredibly impressive!
The laptop form factor hasn't really changed in decades. It's a rectangle with a screen in the lid and a keyboard in the base. Below the keyboard is a battery and a system board. The battery has to be replaced when it wears out and the system board when it becomes obsolete, but then why aren't they both fungible parts? If you take any arbitrary ATX PC from many years ago, you can replace the system board/CPU/memory/storage with modern ones and carry on using the same chassis, screen, power supply and keyboard provided they meet the required specs for the new parts (and they often do).
So why can't I do this with the average laptop, instead of having to replace $200-$300 worth of perfectly good parts or more each time I want an upgrade?
How much did it cost to make this open-source laptop? My wild guess is it's around $500-750.
Not sure what’s meant by “high end” here. Performance is a rather important aspect, and the RK3588 this uses will make it slower than almost every laptop on the market. Practically all are twice as fast (both single- and multi-core), most are 3–5× multi-core, and the best approach 7× (paired with 2.5× single-core).
Looking at Lenovo India, they sell three laptops that are slower multi-core and maybe slower single-core (running Celeron N4020 or Athlon Silver 7120U); after that, they’re all at least twice as fast, in both single- and multi-core benchmarks.
(I’m simplifying to PassMark’s single-/multi-core scores, using <https://www.cpubenchmark.net/cpu.php?cpu=Rockchip+RK3588&id=...> and such.)
From <https://www.byran.ee/posts/creation>: “In many aspects, the Rockchip RK3588 is the fastest consumer-procurable chip on the market.” As someone not involved in these spaces, this was my vague impression, but it still ends up disappointing if you simply can’t get good performance for a project like this because only bigger companies can buy the better-performance things. It’s an extremely impressive project, but unfortunately will be rendered not viable for many—probably most—people for this one reason. That makes me sad. I wish they’d sell us the good stuff.
It's lovely to see HN so nice and friendly, keep it up guys!
Maybe laptop should have two layers or parts.One for Motherboard and memory and another for connectors, fans, power supply, battery etc. Then we can have more standard even if a little thicker.
Personally I'm a bit disappointed that it's based on Rockchip.
If someone can come up with low cost open source laptop with RPi compute module 5 with 16GB RAM I think it will selling like hot cakes given the software and hardware eco-system that exist round RPi [1]. It just that the compute module has yet to come with 16GB RAM unlike the normal RPi 5 but it will probably just around the corner [2].
[1] Compute Module 5:
https://www.raspberrypi.com/products/compute-module-5/
[2] New 16GB Raspberry Pi 5 on sale now at $120 (191 comments):
Byran, I have been a professional engineer longer than you have been alive, I can tell you right now that I have met very, very few people that would have the motivation, skill and sticktoitivness to pull this off.
Please hire this guy to help you make the thinkpad's keyboard to do wireless magic with the trackpoint and the mouse buttons include. Thanks
Congrats on the awesome project! Actually, I think 'well done' is more fitting since this must have taken a ton of work and willpower!
---
It was neat to read through the progress log (1), which begins,
> "It was around 1AM. I wrote up the mission goal (2) and went to sleep at 2AM. The start",
and ends:
> "With the YouTube video and blog post almost done, I hope this isn’t the last of anyon_e. But rather, the start of a trailblazing journey."
This project is the epitome of MUREX Electrical's mission statement, "attempt the impossible":
It's "impossible", a non-MUREX Robotics Electrical member might say. However, we accept it as the process. In the end, we will have achieved something others might have called "impossible". But the achievement only comes through endless, motivated attempts at the impossible. (3)
2: https://www.byran.ee/posts/mission, which links to (3)
3: https://github.com/murexrobotics/electrical?tab=readme-ov-fi...
The mixed presentation of plug and play components interspersed with EE problems and solution really helped make it more accessible. It also got me excited about the possibilities and made me realize that we we might already approaching another open architecture DIY boom.
I got the sense that this is a side project, but I'm sure many have noticed that it could be a legit framework-level company. Someone already mentioned the recruiters, but also you're sure to have investors knocking. Whatever you do, please keep having fun and sharing it.
Suggestion: It would be nice to include a price list on the article.
This project is impressive as heck, but aside from being intellectually out-of-reach for most kids, it would be financially challenging as well. Last I looked, CNC aluminum blocks were well out of the reach of 99.9% of kids (but that was decades ago; perhaps prices went down).
For people wanting to follow in those footsteps, it'd be nice to know which things cost $5, which $50, $500, or $5000. Just that kind of intuition is helpful.
I have a board with old MT6572, it idles at 270mW with working CPU, even less when in semi-sleep (turns off CPU and wakes up every half a second).
1. Make sure the bootloader (u-boot) loads the kernel as fast as possible.
- Disable automatic Ethernet/USB/other subsystems initialization (you can keep them enabled, just don't activate unless requested in the shell manually by the user)
- Tune `distro_bootcmd` command
- Make sure that MicroSD/eMMC/SSD works full-speed (with proper clocks and speed protocol)
2. Use fast decompression algorithm for the kernel and initramfs
- It's either zstd or gzip
3. Collect boot file access data and sort the files on the filesystem
- The benefit in near-linear access & read-ahead
> I ran out of time
And then realized this was a ... high school project!?
Way to go, amazing work!
I'd be really happy with myself if I just built a case and put off-the-shelf components in it.
Just commenting so I can't come back later and claim I wasn't stealing lots of ideas from the author for my own project; the hinge problem, keycaps, the mainboard designed on KiCAD, are all interesting.
Mental note, a commercial laptop of similar specs should never cost more than $4,673.81.
It is no doubt an incredible achievement. I don't like the 'anyone can do this' when that clearly isn't true - it comes across as a humble brag and seems to be a strong part of hustle culture. I would much prefer 'anyone with a decent amount of money and a high enough intelligence can do this', or 'this is now far easier to do than it has ever been'.
I do like the idea of MIT being a beacon to the best and brightest and I do think that the lack of a level playing field means that many otherwise talented people miss out on that opportunity. Perhaps what I would really like is for the world to have more MITs but I don't know if that is possible and I worry that attempts to do this would undermine the quality of MIT. So perhaps I should be content that MIT exists as is and that some people get to go there even if I did not - we all benefit from the fruits of their labor. My university was a top tier university renowned for harsh grading and I was still rather disappointed by the quality of my peers and I worry that the quality at universities in general has since declined further.
Cheap and high quality small batch electronics and hardware fabrication is rapidly changing the world in a way that I think few people understand. It used to be that you had to have a decent size company to do this kind of stuff and that company needed capital investment, layers of management etc. So the cost of bringing a widget into the world was really expensive, risky, and took a long time. The only way to make that money back was to do things in bulk and sell a lot of them which meant you had to be sure there was a sufficient target market. These days a single person can design and fabricate a single item for comparatively very little. And if they want to make it accessible to the rest of the world there is no need to build a factory, just upload the plans. If it's a popular design in all likelihood someone in China will produce it in bulk at commodity prices. The speed of commodification has become so fast that it's practically instant. There is a bit of a phenomena going on at the moment with 'high tech overproduction' where it is claimed that China is intentionally over producing high tech goods to undermine Western markets - it's my view that they're ahead of us on the commodification curve. As manufacturing also manufactures the manufacturing tools the commodification process is a self reinforcing cycle.
Not quite llama.cpp level easy but definitely doable.
For 7B class models the speed is usable
It’s posts like this, fueled by incredible community support, that make Hacker News not just great but unmatched.
With 2,000 points (and counting), this Show HN is currently ranked as the 4th-best Show HN of all time. If we exclude the #1 post (this upvotes itself)—which isn’t a true project—this post would be the 3rd-best of all time. Who knows? By tomorrow, it may surpass 2,741 points and claim the #1 spot outright.
Outstanding work, Bryan. All the best.
+7B means "additional 7B"
if you want to say "more than" or "at least", you say "7B+"
Feel like you could make a pared down version of this with commodity parts outside of the chassis if you aren't going for a flagship competitor. I guess you could also just buy a $20 chromebook, too. Maybe...you could fit a nice rockchip SOM inside a chromebook??
Honestly though, I think the maturity shown in his write-up impressed me even more.
Inspirational.
Thank you for your service to the free and open source principle. Richard Stallman and Eric Steven Raymond would be proud.
The part I am most interested in is the 'powertrain' and how it manages battery charging.
[1] https://wiki.friendlyelec.com/wiki/index.php/CM3588_NAS_Kit#...
Talk
https://www.youtube.com/watch?v=djb0T9COjD4
Article by Wilczek on Anyons
How about the embedded software? Probably not. This is OK as I understand this project is a huge achievemnt. Thank YOU
I've always thought it would be good for hard drive software ( embedded software on the actual drive or SSD ) to be open source. My thinking on such a situation is such a project could start with a storage device from maybe 5 years more more ago - as in my opinion the software would be a less technological challenge. (( I mean take an existing already manufactured working storage device ( ssd or hard drive ) and replace the embedded commercial software with open source software. This would remove the technological barrier of actually constructing the hard drive hard. ))
Comments: