clairity

japan also has tunnel infrastructure for cars. one of my most vivid memories of visiting there is driving in and out of many-miles-long tunnels, and coming out of the last one smack dab in the middle of kobe. it was magical.

you'll never get change if your goal is to make cars so inconvenient that it pisses off the overwhelming majority of voters. cars are here to stay, so let's accept that reality and work around it. what you can do is demonstrate a better version of the world that people can get behind. but yes, it's not to create more car infrastrastructure, but actually less, because tunnels are expensive, so we'd be more selective about where to put roads. then we can also strategically reduce street area dedicated solely to cars (particularly street parking, but also lane number and width), reclaiming much of our limited surface area for people first.

> "Then, the only vehicles using the streets are mostly taxis (they don't need parking) and trucks, not personal cars. Narrow streets + no parking lots means buildings are much closer together and the city is dense enough for walking and cycling (with the help of subways of course)."

yes, let's build denser/closer together, build trains/dedicated busways, and let's not build (surface) parking lots. that's not incompatible.

off of the underground roads, you can put central, underground parking to serve urban neighborhoods and eliminate the need for permanent street parking (obviously services could still have surface space for intermittent parking). you could have fun little green alleyways aboveground (like in japan!) to access the central parking lot. you could even use a single underground car alley under the middle of a primarily residential block to provide access to underground garages under homes/apartment buildings. imagine how cool it would be to replace the street between you and your neighbor and set up a sports field or a garden or a picnic area, or any number of possibilities (even a whole row of new housing could fit down most "residential streets" in many places).

japan has an extensive network of tunnels and manages them just fine. and it's not a money problem. california, for instance, is the 5th largest economy in the world (~$3.6T), with greater LA ~$1.2T and the bay area ~$1T of that. cars have great utility, and while the economic benefits of roads/streets more than offset their cost in urban areas, a better mix of transportation reduces the amount of roads & streets we need overall (not to mention needing to repair them less often when usage is lower), mitigating the cost issue a bit.

people react viscerally to moving away from a sole reliance on cars, but most people intuitively enjoy human-centered environments more than car-centered ones. i wouldn't advocate taking away cars, but rather providing a more human-centered environment that de-emphasizes cars in our lives so that we can better build relationships and communities with each other.

my dream: urban roads (and parking) underground and multimodal streets above-ground severely road-dieted, with ample trees and micromobility lanes along them all. trees provide shade, reduces the heat island effect, reduces noise pollution (a bit), reduces air pollution (a bit), preserves soil and helps it hold more water, and overall reduces stress for pedestrians. adding mixed-use everywhere, with smaller pedestrian/micromobility alleys for accessibility, means people have somewhere to walk to rather than hopping into their cars for everything.

then you don't need to rely so much on speed limits, signs, and signals, because you've segregated (most of) the cars from people.

yes, another restatement of the distinction between 'plan' and 'workflow' as used in the talk.

no, the bullwhip effect is caused by uncertainty across coupled tasks along the critical path, no matter the level of detail of your plan. it's a statistical phenomenon that you can mitigate with other measures, like critical chain. the empire state building example shows that they reduced coupling by creating 4 separate, independent workflows, mitigating the bullwhip effect by roughly a factor of 4.

yah, plan is an overloaded word, like most words. here it seems to boil down to the old adage of "good, fast, cheap. choose two."[0] if you set scope and budget (the "plan" version), then you have to be flexible on deadline. if you set deadline and budget (the "workflow" version), you have to be flexible on scope. for the empire state building, they designed the building (controlled scope) to fit the budget and deadline.

whether it's waterfall or agile shouldn't really matter. in agile, there is planning, but often it's workflow based (tracking flow via story points, or what not).

[0]: https://wikipedia.org/wiki/Project_management_triangle

it's subtle, but "plan" (material resource planning, MRP) is essentially waterfall in this context, where you design first, then do work breakdown into a waterfall schedule. it suffers from cascading delays because of the bullwhip effect (among other things). it's project oriented (a 1-off, 1-time event).

in contrast, "workflow" is akin to kanban in this context. you start with constraints (in this case time and money) and then design the system to those constraints. mary, the speaker, mentions that they had 4 different, decoupled workflows, which helped them avoid those pesky cascading delays. workflows are process oriented (repeatable events), so steel construction, for example, was thought of as an separate repeatable (if varying) process (swimlanes, in kanban parlance) as they went up in height. kanban also focuses on realtime learning and adjustments as well as just-in-time inventory systems (important to steel being delivered on time, like using 2 different suppliers to make sure there were no delays).

this is the stuff you learn in operations class in business school (or some engineering programs), as did chris (the author of the article/blog), who went to ucla anderson.

BLS says that roughly 5000 people die in construction accidents annually. those deaths are certainly tragic, but not disproportional.

oh cool, this would have been really helpful when i was working on my amateur sports startup. we needed a 2-way, realtime wireless bridge to proprietary (but analog) scoreboards (where accurate timing is important), with future plans of doing realtime tracking of players via both sensors and machine vision (like a dumbed down second spectrum). it seemed like ultrawideband (e.g., apple's U1) was going to be the future in this space, but it's been slow to materialize.

(we didn't get far enough along to seriously look into building a device, moreso trying to cobble together something via a raspberry pi or perhaps arduino)

yah, i ran across this article[0] that had some response time comparisons with the new authenticate_by feature in the upcoming rails 7.1 that mitigates a specific timing attack, which is what got me interested in seeing what the rails auth landscape looked like now.

[0]: https://blog.kiprosh.com/rails-7-1-adds-authenticated_by/

i've admittedly only skimmed the (long) article for now, but it's a great reference and really shows how much thought needs to go into user authentication flows. i've done this kind of custom implementation in the past and it was definitely a great learning experience, but i'm not sure i'd trust my code to withstand a serious attack.

they basically recreated a simpler but not-as-flexible (yet?) devise. the flexibility/adaptability of devise seems to be where all the extra complexity comes from. i'm just waiting to see this released as a gem, and the cycle to start anew... =)

yah, i bought 100 flat black hangers a few years ago (~$15 at ikea back then) so i could hang all of my casual clothing (rather than folding, which is more time-consuming). the flat ones allowed me to hang like 50% more stuff in the same space. my fancier clothes go on wooden hangers (also ikea) so they have more room to breathe.

i've had the belkin version of this for some years, the Conserve Smart AV (unfortunately discontinued): https://www.amazon.com/dp/B0085Q722A . i have my tv as the master switch and my apple tv on an always on outlet so that it's the only other device that can wake my tv via CEC (it's my home hub as well, so need to keep it powered continuously). (edit:) oops, just remembered my A/V receiver is also plugged into an always-on outlet and can turn on the tv via CEC as well.

i also have a few of the Conserve Switch models, which have a wireless radio switch but is otherwise not "smart": https://www.belkin.com/conserve-switch/P-F7C01008.html

both have been handy and going strong. i'm still puzzled why belkin basically discontinued the conserve line of products. seemed like the logical launch off point for their smart home stuff (now under the wemo brand name).

i went down the rabbit hole a bit and there doesn't seem to be a homekit-enabled millivolt (not low voltage, which is 24V apparently) thermostat available. however, i did find this little project: https://zekesnider.com/using-an-ecobee-thermostat-with-a-mil... it seems you could patch any low-voltage thermostat to support millivolt in the same way (with a 24V transformer and a relay switch), but the ecobee3 lite has homekit support built-in.

any recommendations for a cheap, simple, but "smart" relay+thermometer like this that can be added to a apple home/zigbee network in plug-and-play fashion[0]? i also have a dumb heater that would be nice to be automated. the existing dumb thermometer/switch is too close to the heat, making that thermostat unreliable, and i can't move it without tearing up walls.

[0]: without having to run home assistant if possible

yah, i used to do something like the parent poster until i realized that it's a serious reduction in the brute-force search space for a malicious attacker.

i'm staying on 1password 7 to avoid their subscription fees, and using sync'ed, shared vaults to have access on my devices too (and share vaults with others as necessary). there's some duplication in apple keychain and firefox for convenience.

i use 1password's built-in 2FA (TOTP), but only for a couple accounts as i find it unwieldy generally. i'm also keeping an eye on how passkeys develop over time.

there are other considerations as well. a dishwasher has a breakeven point of about 2 people eating ~2 meals a day (e.g., 4 place settings a day), because that's the least amount of dishes you need to run it once a day. anything less, and it becomes an additional cognitive load rather than being a mindless convenience. moreover, stuck-on food becomes extra dry overnight and that generally leads to less clean dishes. if you're a single person looking for an apartment (or even a couple who eat out all the time), you should consider a built-in dishwasher negative value because it's priced in to the cost but you likely won't use it (a washing machine is always nice to have though).

i do this same method and i don't know how you could get through 50(!) dishes in anything less than 15 minutes, unless they're not thoroughly cleaned. 5 minutes is the time it usually takes to do ~1 set of dishes (7-10 items), though that's partially because it lacks the compounding effects you mention.

let's poor one out for kobe here since his obituary was flagged off hn at the time.

sure, but that fits into the extant ~5%. it just takes a friend of a friend/family member to spread through social circles, since they overlap greatly, even for the tightest ones.

as i see it, masks make sense in 2 general cases:

1) when uncertainty about transmission and lethality is high, as was the case in nov/dec 2019, until population risk is more closely bounded

2) when situational risk is measurably elevated over baseline, like having comorbidities, surgery, packed mass transit, etc.

by april 2020, we (in the US at least) had enough information that masking wasn't going to be that panacea, when taking human behavior into account (trust channels, weariness, etc.). yet, it was still weaponized by the media and politicians.

i know this because i was making these very same points back in april 2020. to be effective, a mask has to be worn at the right times (like being face-to-face with another human) and correctly (creating a tight seal around the nose and mouth using an effective filtering material). most people weren't doing either of these most of the time.

yet, masks did nothing other than stoke fears and division, as you're doing here. our masking behavior was, and continues to be, irrational, in that it was done at the wrong times with ill-conceived intent. people worried about masking against strangers, when most transmission happened from someone you knew and trusted (probably 95+%). they were generally ineffective by design (<n95), worn incorrectly (loosely), and worn at the wrong times (outdoors/retail). basically no one wore masks at home around their family and friends, where it might have actually reduced some transmission. understanding this, a little distancing was just as, if not more, effective and less divisive despite the literal interpersonal distance.

the CDC was initially correct in assessing that (the way we wear) masks would not be a panacea for covid, but that was politically unpalatable which is why it was walked backed so haphazardly. masks, and all the CDC outrage around them, was effectively mediopolitical redirection of our anger toward each other rather than toward our corporatocratic institutions to quickly and correctly assess the threat and then appropriately deploy (only) effective measures.

alternatively, labor is the only good, which other abstractions (like money) should subserve. the whole point of an economic system is to distribute the rewards of labor across people. most systems (outside idealistic communism) acknowledge that resources will be unevenly distributed, and then focus on the fairness (aka social acceptance) of the chosen distribution, as most people accept that we contribute differing levels of value to the world's production, and so in turn, accept that we get differing levels of benefit from it.

you're getting confused between economies not being zero-sum games in the long term (pie grows/shrinks over time) but do approximate them in the short term (distribution of today's pie). the cost of labor is not going up relative to overall productivity but rather is going down relatively. so where does the surplus end up? concentrated in the hands of people who aren't constrained enough to deploy capital efficiently in an economy.

so in the hotel example, capital is going to owners who are using it to gamble on capital itself rather than going to workers who typically spend the money and help the economy figure out where to invest for long term growth (and health of the economy and society). cost disease is not even applicable here, though rents are. the economy is increasingly seeking economic rents, which are inefficiencies in the economy because rents by definition don't produce anything.

yah, the documentary is as much about david cope, the human composer, as it is about the algorithmic work. to me, the human is an integral part of AI creativity. that’s because it’s the human who evaluates the value of the work, and AI is far from being able to model the vast heterogeneity of human evaluation (for the foreseeable future).

available to rent on amazon prime video, a neat documentary into the first opera created using AI back in the 80's.

https://www.amazon.com/Opus-Cope-Algorithmic-Opera-David/dp/...

https://www.imdb.com/title/tt18298266/

that's how i learned it (equally stressed syllables) from my history of science professor, who (sadistically) assigned gravity's rainbow for the meticulously researched, and fanciful, historical digressions strewn throughout the book. glad to know he wasn't wrong about the pronunciation.

that was my intro to pynchon, being thrown head-first into the deep end. i did actually read the book, but it was... challenging. i appreciate the experience in hindsight, but rued every minute of it at the time.

well that's good to know! i had an implicit assumption of load-bearing walls in my prior comments, which is why they skew the way they do. of course for non-load-bearing walls, drywall is perfectly fine, but i'd still choose plaster & lath (backed by plywood for seismic reasons) over drywall.

for a dream home, i'd love steel or mass timber for the exterior framing to reduce (maybe eliminate) interior load-bearing needs, so that interior layout is maximally configurable while still allowing lots of windows. expensive, but dreamy!

that sounds to me like a (bad) renovation. what really sucks is damaged plaster & lath being ripped out and patched with drywall. that's the worst of both worlds, since you don't get the seamless barrier that is plaster or the ease of use of drywall.

i probably overstated my point by saying drywall has no shear strength, but we seem to agree that drywall < plaster & lath[0] < plywood in that respect. i'd suggest you've also overstated the strength of drywall as "significant". it's not nothing, but against an earthquake, it's approximately nothing. plywood backing, of course, is the minimum required by code for earthquake resistance.

plaster & lath is harder to work with, but has many superior qualities (re: moisture, sound, heat, malleability) that it makes it worth it in many residential cases (not so much commercial, where reconfiguration is more frequent).

[0]: a lathe is a machine tool

no, drywall is basically chalk sandwiched between paper and provides no structural strength to a wall. studs provide compressive strength while plywood provides shear (and some impact) strength to walls. metal strapping is usually added to provide tensile strength. plaster is backed by lath (wood slats), which provides some shear strength akin to plywood backing.