Richard Hamming famously mused that “He who works with the door open gets all kinds of interruptions, but he also occasionally gets clues as to what the world is and what might be important.”
I struggle with this advice.
Arguably, since you’re reading my public newsletter (presumably discovered via a project on my public website), my door is at least somewhat open.
It certainly felt open last week when Hackaday and Hacker News found my notes on PCB stepper motors and several folks wrote to discuss using such motors for applications in, e.g., laser optics and instrumentation indicators.
But while I love hearing from fellow nerds, these sorts of interactions aren’t exactly the same as a casual drop-by-the-office chat.
One key difference is that anything you see on my newsletter or website has been filtered, processed, and made legible: I’ve defined a problem and situated it the context of my motivations and goals. This usually takes quite a bit of closed-door work; despite the “research log” title, my PCB stepper page is a single entry summarizing a few months of work.
During that work, there was no way for you, dear reader, to glance into my metaphorical office to ask about the weird shapes I was drawing in CAD, to help me understand the gist of inductance, or to experience the distinctive smell of motor drivers in thermal overload.
This feels like a missed opportunity, for two reasons:
The mundane: Constructive technical work requires a ton of tacit knowledge. I don’t mean subtle, hard-to-communicate wisdom, but rather all the trivial stuff that makes the difference a day of fruitful work and a day of banging one’s head against the wall for the lack of a command-line flag.
While YouTube is incredible for disseminating tacit knowledge in general, it can’t hold a candle to someone strolling by while you are working and remarking in two seconds, “oh, you’re holding that backwards”.
The change of perspective: Sharing a vague idea might firm it up, and hearing it reflected back might illuminate new dimensions. I can’t recall all the times folks have pointed me toward prior art and potential applications that I would have never discovered on my own.
So why am I not working more in public? Why do I put hundreds of hours of work in various projects that I share only sporadically, off-hand with friends during one-on-one calls? Why do I publish work in finished “chunks” of articles rather than as my daily commits to public Github repos or livestreams from my desktops (digital and literal)?
I’m not entirely sure, but will guess at a few reasons:
I prefer goofing around with new ideas in a code repository / schematic / sheet of paper rather than trying to articulate it on Twitter or some chat app. It’s easier (or more fun) to just show someone a built artifact than trying to describe a hypothetical thing.
I’m afraid that if I do spend time workshopping vague notions on Twitter, my agenda will devolve into checking Twitter constantly to discuss apps for organizing the note-taking apps needed to organize the tweets.
I don’t have an appropriate mechanism for requesting help on the mundane. E.g., I spent a full day this week isolating a C++ / WASM fault that I’m sure one of y'all would have recognized immediately. But if I’d live-emailed/tweeted my debugging, most of y'all would’ve (understandably) unsubscribed after the first few dozen updates. (Popular coding livestreams look amazing for this, but I’m not sure if becoming popular whilst maintaining one’s research direction is less effort than just developing the expertise oneself.)
Working in public now may constrain me in the future (i.e., raising funding or capturing a new market may be difficult after publishing a hundred hour masterclass for interested competitors.)
Livestreaming imposes a self-censorship tax. E.g., I often cold-email people for elaboration on their blog posts and PhD theses. I’d need to either inform them that the correspondence would be made public or remember to not accidentally stream it; in both cases one of us takes on “can this be public?” cognitive overhead.
The public is full of ding-dongs, who sometimes form a mob. I’m not worried about this now, but it’s a source of anxiety for my Internet-famous friends that I’m not keen to adopt. (And I’m not sure how to avoid.)
Even putting aside the mob, I worry that larger audiences might incentivize pursuing more popular and broadly understandable topics over than the ones I find interesting.
But perhaps my framing here is all wrong. After all, Hamming’s door didn’t open to the public — he worked at Bell Labs! This, I suspect, greatly reduced his exposure to time-wasting ding-dongs, the risk of getting attacked by a mob, and the cognitive overhead of self-censorship and donning a public persona.
Perhaps the right question isn’t “how can I work more in public?” but rather “how can I eat in the Bell Labs cafeteria?”. How can I keep up with my peers in a casual, high-trust, private setting?
While a smaller peer group might not have the breadth of technical trivia like the public masses of Stack Overflow, Twitter, etc., the feedback is arguably better in two other ways:
Feedback from peers is likely to be relatable and actionable; if they’re doing similar work, they’ll know where to find the cheap turpentine.
Feedback from peers is easier to qualify. I’ll take technical advice from the people whose technical accomplishments I respect. I’ll discount the advice of peers who tell me to use X when I know they just discovered X last week and are on real kick about it.
It may be much easier to ignore advice from people you know; than to ignore dozens of emails from strangers (or a hundred thousand likes/views) that may plant seeds of doubt; “maybe all those people are right?”
(This is a common dilemma for popular artists; I particularly enjoy the discussions on The Modern Maker Podcast, where YouTubers opine about pursuing their personal furniture design interests versus selling out with DIY epoxy river table and hot tub videos for the masses.)
While I’m a member of several private Slacks, Discords, Zulips, Patreons, and mailing lists of various themes, none of them feel quite like scenius.
It’s interesting to think about why not, if only to frame the question in the positive: What social norms and supporting technologies would enable scenius?
I have a few ideas, but, alas, it may take me several months to filter, process, and make them legible. In the mean time, it’d be easier to have a private conversation — feel free to email me if you have any thoughts on this topic =D
Craig Mod reflecting on the third year of the Special Projects membership program.
Andy Matuschak’s Lessons from 2021, including reflections on his own membership program.
Database researcher Jamie Brandon on his membership program and making past updates public. Check it out, especially if you like databases.
Andy Matuschak got many replies on Twitter for asking “given the prevalence of $400k++ comp packages for ~30-year-olds in Big Tech, why aren’t there lots more "gentle[wo]man scholars”? i.e. people quitting to do non-remunerative creative projects" My $0.02:
I’ve been enjoying these HandmadeCon game developer interviews on networking Starcraft and Diablo, developing control schemes and the Halo and Destiny asset pipelines. Unlike conference talks which can skew abstract and normative, these interviews stay grounded in the context of successful games; the texture of the actual challenges and expertise of the practitioners shines through.
If you want your fusion chamber in under 200 years, you gotta pattern high-powered lasers so you 3D print metal faster.
“As I write, Quorn is pretty much the only commercially viable [single cell protein] product that can be a primary source of protein for humans.” Also interesting: “Things that grow fast and produce proteins quickly contain a lot of RNA for translation. This is a problem because humans metabolize RNA into uric acid and orotic acid, and the overaccumulation these can lead to gout and liver damage, respectively.” The WHO recommends you limit your daily consumption of RNA to 2g.
“This update provides the Generation 12 vaccine design, which includes mutations found in the omicron variant of concern.”
“Be careful of purely verbal, syllogistic reasoning. We make these arguments in conversation all the time. They seem plain, convincing, and commonsensical, but in reality they’re pretty weak.”
Deviations from Chromium (features we disable or remove). A dry reminder of how much of your web browsing goes through Google.
“One of the surprising things to me about web3, despite being built on ‘crypto,’ is how little cryptography seems to be involved!”
Inside the machine that saved Moore’s Law. Sweet photos of ASML technology.
This Amazon review is a masterclass in short (presumably?) fiction.
“It might make sense for San Franciscans to retreat even further into a digital phantasm, given how grim it is to go outside there. But Xi will want Chinese to live in the physical world to make babies, make steel, and make semiconductors.”
“Gifts will not, for heaven’s sake, prevent developers from implementing bad ideas occasionally that turn into security holes. Nothing will. Have you met developers?”
“Villeneuve has pulled them together in an effort to create this post-colonial reading of Dune, not by preaching to us about the evils of imperialism but simply by turning the camera, as it were, and getting us to view the question from the Fremen point of view.”
“Basically any quibbles or hesitations are erased by images like the above, of the Bene Gesserit departing that spaceship, and about which I can only say: That’s fucking sick as hell. This shit rules.”
“Most people are extremely unwilling to grant that faith in textbook knowledge should ever be crowned with success. We have a very strong narrative bias against such stories.”
US net electricity generation hasn’t changed since about 2005.
“Money is a claim on energy yet its creation is not tethered to energy availability or cost.”
“Like with extraverts and introverts (who are about equally numerous), wambs by virtue of being wambs — and therefore more focused on other people — are going to have a greater influence on social consensus, including what personalities are considered normal. Social normality doesn’t reflect statistical normality because the construction of normality is a social process and not a statistical one.”