minhmeoke
WiFi USB Client: Fingernail, eg: Asus USB-AX56 adapter (25.5 x 16 x 9mm) [8]
WiFi Card Client: Postage stamp, eg: Intel Dual-Band Wireless Adapter AC-7260NGW M.2 2230 (22mm x 30mm x 2.4mm) [10]
802.11n Wireless N range: theoretically 230 ft (70 m) range indoors and 820 ft (250m) range outdoors, which is approximately 1 FIFA soccer field's (105 x 68 m) width for indoor range, and 2 soccer fields end-to-end by 3 side-to-side for outdoor range. Newer protocols do not seem to have extended the range of individual radios but rather rely on repeaters or range extender devices to provide additional coverage.
https://www.hummingbirdnetworks.com/articles/what-is-the-dis...
https://en.wikipedia.org/wiki/IEEE_802.11#Protocol
https://publications.fifa.com/de/football-stadiums-guideline...
Some of these are slightly older devices dating from around 2020, but the sizes should generally still be approximately the same in 2025.
Router/Access point: Mobile operating systems like WebOS, Android, and iOS have had hotspot capabilites built-in since around around 2010 (Mobile Hotspot on Palm Pre and Pixi Plus in 2010 [1], Wifi Tethering in Android 2.2 Froyo in 2010 [2], and Personal Hotspot in iOS 4.3 for iPhone 4 and 3GS in 2011 [3]). Today you can configure your phone to become a personal Wifi router with one button press [4], [5].
[1] https://www.cnn.com/2010/TECH/01/07/ces.palm.pre.plus.pixi/i...
https://web.archive.org/web/20101125023042/http://articles.c...
[2] https://www.wired.com/2010/05/android-22-froyo-features-usb-...
https://web.archive.org/web/20140707210122/https://www.wired...
[3] https://www.engadget.com/2011-03-09-ios-4-3-spotlight-person...
https://web.archive.org/web/20251018061108/https://www.engad...
[4] https://support.google.com/android/answer/9059108?hl=en
[5] https://support.apple.com/guide/iphone/share-your-internet-c...
For a discrete router device, TP-Link TL-WR802N (57 × 57 ×18 mm) [6] and GL-Inet GL-MT300N (58 x 58 x 25mm) [7] are matchbox-sized travel router devices that fit into the palm of your hand, cost less than $30 USD, and consume less than 2.75W (they can be run from a basic 5V, 1A USB connection).
[6] https://www.tp-link.com/us/home-networking/wifi-router/tl-wr...
https://web.archive.org/web/20250724184720/https://www.tp-li...
[7] https://www.gl-inet.com/products/gl-mt300n-v2/
https://web.archive.org/web/20250822045920/https://www.gl-in...
Wifi Client: Wifi capabilities are now built into mobile phone, TV, and many camera chipsets, there is no separate card required. If you want a separate WiFi adapter, you can buy a USB adapter that is about the size of a fingernail or 5 small coins stacked together (eg: 5 US pennies, or 5 Euro 1-cent coins), basically take the metal part of a USB-A male connector and extend or extrude it a bit. This Asus USB-AX56 adapter [8] is 25.5 x 16 x 9mm and supports WiFi 6 (802.11ax) with theoretical speeds up to 9.6 Gbps (actual real-world speeds appear closer to 800 Mbps). Streams in Wifi 6 can be up to 160MHz wide via channel bonding of up to 8 adjacent 20 MHz channels, but in practice this is only feasible in locations with low or no interference. Note that wider channels have a reduced effective range and perform poorly at a distance or through obstructions.
[8] https://www.asus.com/networking-iot-servers/adapters/all-ser...
https://web.archive.org/web/20251018061524/https://www.asus....
https://www.wi-fi.org/wi-fi-macphy
Discrete Wifi cards can still be found in some laptop computers (although many Wifi cards are now soldered down), one of the more recent (introduced around 2013 to 2015) and smaller sizes is M.2 2230 (22mm x 30mm x 2.4mm) which is also the form factor for some M.2 PCI-Express SSDs (solid state storage drives, typically flash devices). These are about the size of a postage stamp, weigh around 2.3 grams, and cost around $20 USD or less.
One example is Intel AX210 [9] which supports Wi-Fi 6E, on 3 bands: 2.4 / 5 / 6GHz in a 2x2 configuration (2 TX transmit and 2 RX receive antennas) at speeds of 2.4 Gbps, and also features Bluetooth 5.3.
[9] https://www.intel.com/content/www/us/en/products/sku/204836/...
A popular previous-generation device is Intel Dual-Band Wireless Adapter AC-7260NGW [10] with good support for GNU/Linux, again with a M.2 2230 (22mm x 30mm x 2.4mm) form-factor approximately the size of a postage stamp, supporting Wifi 5 802.11ac with up to 867 Mbps theoretical bandwidth on dual bands in a 2x2 configuration (2 TX transmit and 2 RX receive radios), and 433 Mbps per stream. Channels can be up to 80MHz wide via channel bonding of 4 adjacent 20 MHz channels.
An alternative, earlier form factor was Half Mini-PCI-Express, for example Intel Dual-Band Wirelss Adapter AC-7260HMW [10] at 26.80 x 30 x 2.4 mm.
[10] https://www.mouser.com/datasheet/2/612/dual-band-wireless-ac...
https://web.archive.org/web/20251018061520/https://www.mouse...
# Initial Setup
1. Install Android Platform Tools from: https://developer.android.com/tools/releases/platform-tools#... and verify that you can execute the adb tool.
2. Connect the Android device and the computer to the same Wi-Fi network
3. Also connect the Android device to the computer via USB cable (just for this initial setup)
4. Pair the phone with the computer and enable Wireless debugging as described here: https://developer.android.com/tools/adb#connect-to-a-device-...
# Subsequent usage (does not require USB cable):
5. Toggle the Wireless Debugging quick setting toggle to enable adb connections from your computer to your phone
6. Retrieve the Android device's IP address in Settings → About → Status.
7. On the computer command line type: adb connect ip-address-of-device:5555
8. Push the folder to the Android device: adb push /path/to/folder/to/copy/on/computer /sdcard/path/to/destination/on/android/device
python -m http.server 8000
You can download prebuilt binaries for x86-64 Linux, Windows, or Mac OS (sorry, no prebuilt binaries for Apple Silicon, but they could be added if there is sufficient demand) from https://github.com/akovacs/uploadserver/releases/ or compile from source using a nightly rust toolchain if you prefer.
Compared to cloud services or `python -m http.server 8000`, this is extremely fast since the server is written in rust, it is fairly simple (compiled and stripped binary is typically less than 3MB), it sends everything over local LAN, it seems to handle large files (over 4GB) fairly well, and you only need to install the software on one machine.
For additional details, please see: https://news.ycombinator.com/item?id=39665095
python -m http.server 8000
You can download prebuilt binaries for x86-64 Linux, Windows 7 or higher, or Mac OS 10.7 Lion or higher (sorry, no prebuilt binaries for Apple Silicon, but they could be added if there is sufficient demand) from https://github.com/akovacs/uploadserver/releases/ or compile from source using a nightly rust toolchain if you prefer.
Start the file server on a desktop machine:
chmod +x upload_server
./upload_server
# Linux
hostname -I
# Mac
ifconfig
# Windows
ipconfig
curl -X POST --data-binary @file_to_upload.txt http://192.168.1.X:8000/uploaded_file.txt
curl http://192.168.1.X:8000/uploads/uploaded_file.txt --output downloaded_file.txt
Compared to cloud services or `python -m http.server 8000`, this is extremely fast since the server is written in rust, it is fairly simple (compiled and stripped binary is typically less than 3MB), it sends everything over local LAN, it seems to handle large files (over 4GB) fairly well, and you only need to install the software on one machine.
https://www.unravelled.dev/how-architecture-diagrams-enable-...
- Can check the textual description of the diagram into version control and track how it changes over time
- Can easily diff 2 versions of the same diagram
- Adding elements to a large existing diagram using WYSIWYG tools can be tricky (eg: you may have to move or reroute existing elements to make enough space)
Another cool idea is to create multiperspective diagrams with different views for different purposes (eg: code structure, deployment, network architecture, security):
- Clarity. By breaking up diagrams into perspectives, diagram authors can bring much more clarity to a system. Individual concerns can be spread out over many perspectives, allowing each the space it needs without interfering with others.
- Extensibility. Multiperspective diagrams share a model, so creating additional perspectives from an existing diagram is incredibly easy, far easier than starting from scratch each time.
- Maintainability. A corollary to the above, individual perspectives can be modified without affecting the others. Diagrams that are easy to maintain get maintained, while diagrams that aren’t fall out of date quickly. Multiperspective diagrams greatly help in the fight to keep documentation up-to-date.
This is a cool demo: https://app.ilograph.com/demo.ilograph.Stack%2520Overflow%25...
For other textual diagramming tools and ideas, please see this thread: https://news.ycombinator.com/item?id=37222855
Disclaimer: I'm not affiliated with Ilograph, just stumbled onto their blog and I think they have a lot of cool ideas.
A useful filesharing utility that a friend wrote is https://github.com/akovacs/uploadserver - it's basically a nicer version of:
python -m http.server 8000
Determine your machine's IP address using:
# Linux
hostname -I
# Mac
ifconfig
# Windows
ipconfig
chmod +x upload_server
./upload_server
curl -X POST --data-binary @file_to_upload.txt http://192.168.1.X:8000/uploaded_file.txt
curl http://192.168.1.X:8000/uploads/uploaded_file.txt --output downloaded_file.txtA friend wrote a faster and more sophisticated version of `python -m http.server 8000` in rust which also supports uploads, you can read about here: https://news.ycombinator.com/item?id=37628347
File transfers are performed over the local network, and therefore should be very fast. If you don't have a network connection, you can just enable a Wifi hotspot on your mobile device and transfer the files over that. There is no need to install any software on the mobile device, you can just use a web browser to browse to a directory that you share over the network.
A friend wrote a faster and more sophisticated version of `python -m http.server 8000` in rust which also supports uploads, you can read about here: https://news.ycombinator.com/item?id=37628347 File transfers are over the local network, and therefore should be very fast.
python -m http.server 8000
A friend wrote a faster and more sophisticated version of `python -m http.server 8000` in rust which also supports uploads, you can read about here: https://news.ycombinator.com/item?id=37628347
A useful utility that a friend wrote is https://github.com/akovacs/uploadserver - it's basically a nicer version of:
python -m http.server 8000
Start the file server, and then navigate to it using the web browser of your choice on any device (no need for a client application).
chmod +x upload_server
./upload_server
curl -X POST --data-binary @file_to_upload.txt http://192.168.1.X:8000/uploaded_file.txt
curl http://192.168.1.X:8000/uploads/uploaded_file.txt --output downloaded_file.txt
# install rust toolchain
sudo apt install rustc git
curl https://sh.rustup.rs -sSf | sh
rustup install nightly
rustup default nightly
git clone https://github.com/akovacs/uploadserver
cd uploadserver
cargo run --release> Currently it only displays live data - no recording and replay is possible. NOTE: The program only displays the engine data, not airbags, ABS and body control systems. Even if your adapter supports that, you will need a more specialized program for that.
If you have more sophisticated requirements, it seems like Wireshark has the ability to stream and record data in real-time: https://www.csselectronics.com/pages/can-bus-to-usb-wireshar... is an ~$70 adapter which shows up as a serial port
Alternatively, https://canable.io/ with USB port seems to support several open source software packages (there are clones of Canable 2.0 Pro available for ~$17 on AliExpress, but hopefully the official version will come back in stock soon):
- Cangaroo (Windows, Linux): Send/receive standard and FD frames, decode messages from DBC files
- SocketCAN (Linux) Native Linux support using slcand
- python-can (Windows, Linux, Mac) Easily interface with the CAN bus using your CANable and Python scripts
Disclaimer: I have no affiliation and have not used either of these, just posting them as possible options.
It's basically a nicer version of:
python -m http.server 8000
It works on Windows, Mac, and Linux, and therefore should probably work on FreeBSD as well if you can install a Rust toolchain:
pkg install curl
pkg install git
pkg install rust
git clone https://github.com/akovacs/uploadserver
cd uploadserver
cargo run --release
curl -X POST --data-binary @file_to_upload.txt http://localhost:8000/file_to_upload.txt
Alternatively, if you want to stick to ASCII output, https://arthursonzogni.com/Diagon/#Sequence supports several formats (see the dropdown) which can pipe into something like Typogram (https://google.github.io/typograms/) for more beautiful output.
For example, input:
Renderer -> Browser: BeginNavigation()
Browser -> Network: URLRequest()
Browser <- Network: URLResponse()
Renderer <- Browser: CommitNavigation()
Renderer -> Browser: DidCommitNavigation()
.--------. .-------. .-------.
|Renderer| |Browser| |Network|
'--------' '-------' '-------'
| | |
| BeginNavigation() | |
|-------------------->| |
| | |
| |URLRequest() |
| |------------>|
| | |
| |URLResponse()|
| |<------------|
| | |
| CommitNavigation() | |
|<--------------------| |
| | |
|DidCommitNavigation()| |
|-------------------->| |
.--------. .-------. .-------.
|Renderer| |Browser| |Network|
'--------' '-------' '-------'
One more ASCII-based tool that you could add into your workflow is https://arthursonzogni.com/Diagon/#Sequence which can be used to generate the ASCII that you input into Typogram.
For example, input:
Renderer -> Browser: BeginNavigation()
Browser -> Network: URLRequest()
Browser <- Network: URLResponse()
Renderer <- Browser: CommitNavigation()
Renderer -> Browser: DidCommitNavigation()
.--------. .-------. .-------.
|Renderer| |Browser| |Network|
'--------' '-------' '-------'
| | |
| BeginNavigation() | |
|-------------------->| |
| | |
| |URLRequest() |
| |------------>|
| | |
| |URLResponse()|
| |<------------|
| | |
| CommitNavigation() | |
|<--------------------| |
| | |
|DidCommitNavigation()| |
|-------------------->| |
.--------. .-------. .-------.
|Renderer| |Browser| |Network|
'--------' '-------' '-------'
- Grab to-go salad box at Whole Foods
- Burritos at Chipotle
In SF, I would avoid the Tenderloin, parts of SOMA from 6th Ave and higher, Mission district below 20th Ave, Excelsior, and Hunter's Point as well as the Bayview neighborhoods. Avoid East Palo Alto in south bay (the rest of Palo Alto is quite safe).
For short-term housing, this could be a good option if your budget is ~$1000/month: Berkeley is generally quite safe https://och.berkeley.edu/housing-secondary/price-under+1100/...
This looks like a nice Anki deck for reviewing system design concepts: https://github.com/donnemartin/system-design-primer
You might also be interested in Alex Xu's System Design Interview Book.
Ubiquity UAP AC LR also has long-range access points with up to 183 meters range for ~$109 USD: https://store.ui.com/products/unifi-ac-lr
If you are more ambitious, you can deploy an OpenWRT BATMAN-based mesh network: https://openwrt.org/docs/guide-user/network/wifi/mesh/batman
I recommend using something that supports OpenWRT out of the box, such as GL-iNet Mango Mini Router GL-MT300N-V2 (~$20 USD refurbished) https://store.gl-inet.com/products/certified-refurbished-pro...
Please only use this information for humanitarian/civilian purposes and not for military objectives. Thank you!
1. Get back into the habit of writing and reviewing evergreen notes (https://notes.andymatuschak.org/Evergreen_notes, An Executable Strategy for Writing: https://notes.andymatuschak.org/z3PBVkZ2SvsAgFXkjHsycBeyS6Cw...) and refining my Zettelkasten knowledgebase and perhaps making it publicly available at some point.
2. Continue refining the precision of my vocabulary and my understanding of the etymology of words. I've realized that I only have a fuzzy grasp of many of the words I use on an everyday basis, yet by becoming more aware of the nuances and subtleties of the usage, origin, and connotations of different terms, I will be able to express myself more accurately and also perhaps better be able to read between the lines of what others say and write.
For example, consider the following definitions:
Compound: composed of two or more separate elements; a mixture.
Complex: consisting of many different and connected parts.
Consequently, a set of buildings which share the same property but are unconnected (for example a mobile home community, or standalone military barracks buildings) should be referred to as a compound. In contrast, if the structure has bridges between the different parts, it should be called a complex.
Also consider:
Sophisticated: (of a machine, system, or technique) developed to a high degree of complexity.
Complicated: consisting of many interconnecting parts or elements, often involving many different and confusing aspects
(from Latin complicat- "folded together", from the verb complicare, from com- "together" + plicare "to fold")
An automobile is a very complex machine, since it consists of thousands of parts and systems. To a mechanic, it might be complicated (as reflected by manuals which are hundreds or thousands of pages long and hours of frustration), while to a user it is sophisticated - capable of doing many things, and yet still easy to use, since the complexity is abstracted away.
These 4 words are seemingly very closely related, yet through careful word selection, we can use them to communicate very different ideas and emotions.
The Megatronix SC620 Six Tone Car Alarm is also fun to watch: https://www.youtube.com/watch?v=xXYc0gtOswQ&list=PLn67ccdhCs...
As well as the a cappella Starships (Nicki Minaj cover) by Mike Tompkins which has a mind-blowing breakdown of how he built up the composition from different vocals at the end: https://youtu.be/vDqHpWY1UD4?t=209
This is a great way to refine your ideas before you write and makes it easier to develop the outline as an assembly of different notes that you already have.
An Executable Strategy for Writing: https://notes.andymatuschak.org/z3PBVkZ2SvsAgFXkjHsycBeyS6Cw...
The main insights for me were: - Using notes as a way to organize one's writing (by assembling ideas together into an outline, then filling out the details) to avoid writer's block. - Creating "logs" around concepts that extract useful ideas from ephemeral observations and distill concrete insights over time. - Using an "inbox" of new ideas as a way to focus attention and perform spaced repetition of concepts. - Organizing notes via tags, backlinks, and associations/outlines rather than as a hierarchy. - Incrementally iterating on atomic concept notes to form larger chunks in memory which allows thinking about more complex ideas and recognizing patterns.
I was curious about the composition of music, therefore a friend and I created/curated frequency-domain visualizations of a variety of music in different languages and interesting sounds using the Welch power spectral density estimation algorithm and Fast Fourier Transform. The visualizations are fun to look at and with some practice you can distinguish the different instruments and vocal tracks in each song.
Chinese: https://www.youtube.com/watch?v=7cWixC3PVP8&list=PLn67ccdhCs...
Russian / Ukrainian: https://www.youtube.com/watch?v=ysqIz8-pRSE&list=PLn67ccdhCs...
Spanish: https://www.youtube.com/watch?v=gsJtqF5hCyE&list=PLn67ccdhCs...
Japanese: https://www.youtube.com/watch?v=sRjxDSIx3Zs&list=PLn67ccdhCs...
Hungarian: https://www.youtube.com/watch?v=UNY3HOnJiUc&list=PLn67ccdhCs...
English: https://www.youtube.com/watch?v=jXCdJ1Lh-48&list=PLn67ccdhCs...
A few other examples which might be interesting:
1. Dialtone using dual-tone multi-frequency signaling and 56K dial-up modem connection: https://www.youtube.com/watch?v=FomWraKuDFg&list=PLn67ccdhCs...
2. Deluxe Multitone Car Alarm: https://www.youtube.com/watch?v=A4uKcvZL7HM&list=PLn67ccdhCs...
3. Composition using only sounds from Windows 98 and XP https://www.youtube.com/watch?v=6lT-jr9sS6Y&list=PLn67ccdhCs...
4. Piano Music (Ballade Pour Adeline): https://www.youtube.com/watch?v=RnAfrEk429w&list=PLn67ccdhCs...
5. Electronic Music Demo: https://www.youtube.com/watch?v=MllJLIX1glg&list=PLn67ccdhCs...
6. Chiptunes (Mortimer's Chip-disko from Supertux): https://www.youtube.com/watch?v=FWHAHPbYc0w&list=PLn67ccdhCs...
MiniDexed: 8 DX7 Tone Generators
- https://github.com/probonopd/MiniDexed
- https://www.youtube.com/watch?v=Z3t94ceMHJo
MT32-Pi: Roland MT-32 emulator
- https://github.com/dwhinham/mt32-pi
- https://www.youtube.com/watch?v=vaSD_wbzJRw
Optionally, buy a PCM5102/PCM5122 or ES9018K2M I2S DAC to improve the output sound quality since the DAC built into the Raspberry Pi is not that great:
https://raspberrypi.stackexchange.com/questions/76188/how-to...
PCM5102 (~$7): https://www.aliexpress.com/wholesale?SearchText=PCM5102+DAC
ES9018K2M (~$12): https://www.aliexpress.com/wholesale?SearchText=es9018k2m+i2...
- https://selecthealth.org/blog/2020/07/7-health-benefits-of-s...
Getting a lot of sunlight in the morning is also a good way to reduce jetlag by resetting your circadian rhythm to be in sync with your new time zone.
If you live in a climate where you do not have access to sufficient sunlight, one possible alternative is to buy an LED corn bulb (don't use it too late at night, otherwise you won't be able to fall asleep):
- https://blog.plover.com/tech/corn-bulbs.html
- https://meaningness.com/sad-light-lumens
- https://www.benkuhn.net/lux/
You can buy these fairly inexpensively from retailers like AliExpress if you are ok with waiting a month for shipping from China: https://www.aliexpress.com/wholesale?SearchText=corn+bulb
If you live in North America, Japan, Korea, or Taiwan (120V AC), you want an E26 socket lamp, otherwise for 240V AC grids, you should consider E27 socket lamps.
It's very similar yet much fuller-featured than commonmark, with support for definition lists, footnotes, tables, several new kinds of inline formatting (insert, delete, highlight, superscript, subscript), math, smart punctuation, attributes that can be applied to any element, and generic containers for block-level, inline-level, and raw content. In addition, it resolves ambiguities in the commonmark spec and parses in linear time with no backtracking.
Further discussion lower in this thread: https://news.ycombinator.com/item?id=33553293
Quickstart for Markdown users: https://github.com/jgm/djot/blob/main/doc/quickstart-for-mar...
Some more in-depth examples, showing how Djot would be rendered into HTML: https://htmlpreview.github.io/?https://github.com/jgm/djot/b...
It's very similar to the Markdown syntax we all know and love/hate, but fixes many inconsistencies in the spec, and also makes it possible to parse a document in linear time, with no backtracking. It is also much fuller-featured than commonmark, with support for definition lists, footnotes, tables, several new kinds of inline formatting (insert, delete, highlight, superscript, subscript), math, smart punctuation, attributes that can be applied to any element, and generic containers for block-level, inline-level, and raw content.
Some examples, showing how Djot would be rendered into HTML: https://htmlpreview.github.io/?https://github.com/jgm/djot/b...
https://sites.astro.caltech.edu/palomar/media/slideshows/mir...
Click the slideshow and press the left/right arrows to advance the slides or go back.
It seems like they took 11 months to manufacture the 20 mirror, then they spent 11.5 years polishing it down to 14.5 tons in Pasadena... wow! I can't imagine being the crane operator while loading and unloading this thing.