--- Log opened Sat Jun 19 00:00:42 2021 00:25 -!- cpopell [~cpopell@c-24-22-170-81.hsd1.wa.comcast.net] has joined #hplusroadmap 01:26 -!- Urchin[emacs] [~user@user/urchin] has joined #hplusroadmap 02:02 < fenn> http://www.macmillandictionaryblog.com/new-years-resolution-no-adverbs 04:59 -!- yashgaroth [~ffffffff@2601:5c4:c780:6aa0:e975:8c69:19cd:7a28] has joined #hplusroadmap 08:03 -!- cpopell [~cpopell@c-24-22-170-81.hsd1.wa.comcast.net] has quit [Quit: Connection closed] 08:04 -!- drethelin [~drethelin@096-042-035-082.res.spectrum.com] has joined #hplusroadmap 10:00 -!- ptrcmd [~ptrcmd@user/ptrcmd] has quit [Ping timeout: 272 seconds] 10:00 -!- ptrcmd [~ptrcmd@user/ptrcmd] has joined #hplusroadmap 10:12 -!- L29Ah[XMPP] is now known as L29Ah 11:11 < kanzure> https://en.wikipedia.org/wiki/Heavy_Press_Program 11:25 < srk> https://ygoliya.medium.com/manufacturing-next-generation-electronics-like-they-are-t-shirts-526e8551748b 11:41 < L29Ah> photoresist looks better wrt pcb-making 11:45 < L29Ah> https://www.copprint.com/ resistance comparison 12:33 < srk> I prefer milling for prototyping since it's quite fast and requires no chemicals, "just" a cnc :P 12:34 < srk> copprint inks look cool but it's 550Eur for 1kg, would prefer smaller sample 12:36 < L29Ah> yeah, with photoresist you still need milling to get beyond simple SMD circuits 12:37 < L29Ah> making precise stencils is probably involved as well 12:37 < L29Ah> for printing 12:38 < L29Ah> and it's fun how they tell subtractive tech isn't sustainable while their stuff requires a few times more material to obtain the same conductivity as copper 12:40 < srk> indeed :) classic marketing bs 12:42 < srk> had some fun with stencil precision before, also dots that need dissolving 14:50 -!- drethelin [~drethelin@096-042-035-082.res.spectrum.com] has quit [Ping timeout: 252 seconds] 15:06 -!- drethelin [drethelin@096-042-035-082.res.spectrum.com] has joined #hplusroadmap 15:33 -!- drethelin [drethelin@096-042-035-082.res.spectrum.com] has quit [Ping timeout: 265 seconds] 15:43 < fenn> https://www.motherjones.com/kevin-drum/2018/02/an-updated-lead-crime-roundup-for-2018/ 15:51 < jrayhawk> https://www.gla.ac.uk/media/Media_774797_smxx.pdf 2021 meta analysis 16:03 < fenn> too many fancy statistics words for me to understand 16:03 < fenn> also i find it odd that they only searched for papers about "crime" "jail" "arrests" "convictions" and so on, rather than violence, homicide, assault, etc. 16:07 -!- cpopell [~cpopell@c-24-22-170-81.hsd1.wa.comcast.net] has joined #hplusroadmap 16:07 < cpopell> o/ 16:12 < fenn> heteroskedasticity 16:20 < fenn> what is with all these people and not starting at 0 on the y axis 16:39 < cpopell> either to zoom in on differences, or manipulate perceptions 17:22 < fenn> jrayhawk: well i read the meta analysis front to back and still have no idea what i just read 17:23 < fenn> is "ideal specification" correcting for publication bias? 17:27 < fenn> the abstract for the largest study by gromqvist et al says "Following 800,000 children differentially exposed 17:27 < fenn> to the phaseout of leaded gasoline in Sweden, we find that even a low exposure affects long-run outcomes, that boys are more affected, and that changes in noncognitive skills 17:27 < fenn> explain a sizeable share of the impact on crime and human capital" 17:27 < fenn> but in the meta analysis it's shown as a big fat zero 17:28 < fenn> (pg 42 3rd row) 17:46 < fenn> i'm reading that study now, they say that sweden had much lower lead exposure levels than the US at the same time. also it's listed as an "individual" study (as compared to an "area" study) even though they are measuring the lead in the moss around where the individuals were born, not the lead in the blood of the individual 17:51 < fenn> the correlation values in this one study are tiny, but the N is very large, so i wonder if perhaps the meta-analysis is giving too much weight to it, even though it only explores a narrow portion of the range from no-lead-exposure to a-little-lead-exposure and does not study lots-of-lead-exposure at all 17:59 < fenn> if the effect is non-linear (and it is, according to gromqvist et al, saying the threshold is 5ug/dL blood in early childhood) 18:00 < fenn> then you would expect a single large study with a low average exposure to skew a linear regression relative to a complementary set of small studies with high levels of exposure 18:15 < fenn> some other studies: (in ug/dL) reyes 2009: 10-20, reyes 2014: 7.6, billings & schnepel 2015: 10-20, gazze 2016: 10-20, rau 2016: ~10, aizer 2016: <10 18:16 < fenn> average childhood blood levels of US adults in 2002: 10, and in 2018: 3 18:19 < fenn> my bad, 5ug/dL was the threshold for school problems 18:19 < fenn> 7ug/dL was the threshold for crime 18:20 < fenn> in urban sweden during the relevant period the average was 6ug/dL and fell to 4 after banning leaded gasoline 18:21 < fenn> or in other words, enough to show effects on cognition, but not enough to cause crime 18:23 < fenn> of course you would expect there to be more variation in individual exposures, but their sampling method just considers the concentration of lead in moss in the area, which seems to be a smooth gradient from urban to rural 18:50 -!- yashgaroth [~ffffffff@2601:5c4:c780:6aa0:e975:8c69:19cd:7a28] has quit [Quit: Leaving] 19:04 < fenn> ok i think gromqvist's thresholds should probably be the same value 19:14 -!- cpopell [~cpopell@c-24-22-170-81.hsd1.wa.comcast.net] has quit [Quit: Connection closed] 19:15 < fenn> in 1979 average blood levels in US were 14 ug/dL 22:38 -!- justanotheruser [~justanoth@gateway/tor-sasl/justanotheruser] has quit [Ping timeout: 252 seconds] 23:10 -!- drethelin [drethelin@096-042-035-082.res.spectrum.com] has joined #hplusroadmap 23:34 -!- Urchin[emacs] [~user@user/urchin] has quit [Ping timeout: 268 seconds] 23:49 -!- drethelin [drethelin@096-042-035-082.res.spectrum.com] has quit [Ping timeout: 265 seconds] 23:56 -!- Urchin[emacs] [~user@user/urchin] has joined #hplusroadmap --- Log closed Sun Jun 20 00:00:43 2021