<azonenberg>
rqou: good mcu rngs are actually crypto level
scrts has quit [Ping timeout: 248 seconds]
gruetzko- has joined ##openfpga
zino_ has joined ##openfpga
zino has quit [Ping timeout: 256 seconds]
gruetzkopf has quit [Ping timeout: 256 seconds]
SaLo has quit [Ping timeout: 256 seconds]
SaLoMoN has joined ##openfpga
pointfree1 has quit [Ping timeout: 240 seconds]
sielicki has quit [Ping timeout: 245 seconds]
nrossi has quit [Ping timeout: 240 seconds]
anuejn has quit [Ping timeout: 276 seconds]
AlexDaniel` has quit [Ping timeout: 260 seconds]
jfng has quit [Ping timeout: 260 seconds]
indefini has quit [Ping timeout: 276 seconds]
digshadow has quit [Ping timeout: 240 seconds]
scrts has joined ##openfpga
scrts has quit [Ping timeout: 255 seconds]
digshadow has joined ##openfpga
digshadow has quit [Ping timeout: 256 seconds]
digshadow has joined ##openfpga
<bofh_>
21:51:53 < rqou> hmm random thought: how feasible is a DIY ghetto-NMR?
<bofh_>
21:52:01 < rqou> bofh_?
<bofh_>
I'm halfway thru plans for one. It's extremely possible to build a small benchtop low-field NMR using NdFeB magnets for the permanent field, and this will actually perform reasonably well.
diadatp has quit [Remote host closed the connection]
ondrej2 has quit [Read error: Connection reset by peer]
ondrej2 has joined ##openfpga
bitd has joined ##openfpga
rohitksingh_wor1 has joined ##openfpga
rohitksingh_work has quit [Ping timeout: 255 seconds]
rohitksingh_work has joined ##openfpga
rohitksingh_wor1 has quit [Ping timeout: 260 seconds]
shuffle2_ is now known as shuffle2
shuffle2 has quit [Quit: never coming back]
shuffle2 has joined ##openfpga
azonenberg has quit [Ping timeout: 255 seconds]
openfpga-bb has quit [Ping timeout: 255 seconds]
openfpga-bb has joined ##openfpga
azonenberg has joined ##openfpga
pointfree1 has joined ##openfpga
rohitksingh_work has quit [Ping timeout: 276 seconds]
indefini has joined ##openfpga
anuejn has joined ##openfpga
nrossi has joined ##openfpga
sielicki has joined ##openfpga
AlexDaniel` has joined ##openfpga
jfng has joined ##openfpga
rohitksingh_work has joined ##openfpga
pointfree1 has quit [Ping timeout: 256 seconds]
jfng has quit [Ping timeout: 246 seconds]
AlexDaniel` has quit [Ping timeout: 245 seconds]
nrossi has quit [Ping timeout: 240 seconds]
anuejn has quit [Ping timeout: 240 seconds]
indefini has quit [Ping timeout: 276 seconds]
sielicki has quit [Ping timeout: 240 seconds]
rohitksingh_wor1 has joined ##openfpga
rohitksingh_work has quit [Ping timeout: 240 seconds]
pie_ has quit [Ping timeout: 264 seconds]
X-Scale has joined ##openfpga
pointfree1 has joined ##openfpga
nrossi has joined ##openfpga
jfng has joined ##openfpga
indefini has joined ##openfpga
AlexDaniel` has joined ##openfpga
sielicki has joined ##openfpga
anuejn has joined ##openfpga
zino_ is now known as zino
pie_ has joined ##openfpga
pie_ has quit [Ping timeout: 268 seconds]
gruetzko- is now known as gruetzkopf
rohitksingh_wor1 has quit [Read error: Connection reset by peer]
rohitksingh has quit [Read error: Connection reset by peer]
rohitksingh has joined ##openfpga
Bike has joined ##openfpga
genii has joined ##openfpga
rohitksingh has quit [Quit: Leaving.]
azonenberg_work has joined ##openfpga
azonenberg_work has quit [Ping timeout: 248 seconds]
m_w has joined ##openfpga
<rqou>
whitequark: why are you using xen?
azonenberg_work has joined ##openfpga
<whitequark>
rqou: because I need virtualized Windows?
<rqou>
QEMU/KVM can do that too
<rqou>
ime that's much easier to set up
<whitequark>
huh? what's hard about xen
<whitequark>
I literally just `apt-get install xen-system`
<rqou>
ime getting the boot chain to actually work is/was very hard
indy has joined ##openfpga
<rqou>
iirc there were a whole bunch of issues like "cannot agree who is supposed to call ExitBootServices"
<rqou>
maybe they fixed it by now
<rqou>
after I already switched to KVM
<awygle>
my xen problems were all because i was also using full-disc encryption and secure boot
<awygle>
it worked fine once i got it set up
<rqou>
i never got it to work at all
<whitequark>
nope, never hit anything like that
<rqou>
and using pcie passthrough seems to be much easier with KVM
<rqou>
well, i guess they finally fixed it
<awygle>
oh i was _also_ running lxc as well
<whitequark>
are you booting it directly from EFI?
<awygle>
i did a number of dumb things
<whitequark>
if you install it on Debian it gets booted via GRUB
<rqou>
i don't remember
<rqou>
probably booting through grub
<awygle>
i was booting it from grub. EFI->shim->grub->linux
<awygle>
iirc i had a big problem with EFI not finding the shim, i eventually had to name it something like "i_am_the_official_intel_boot_image_please_load_me.img"
<rqou>
anyways, maybe kvm wouldn't have required a patch for RSDP address?
<awygle>
this was ~5 years ago though
<rqou>
i was probably using EFI->rEFInd (maybe still rEFIt at that time)->grub->linux
<whitequark>
rqou: probably, but this was by no means major issue
<rqou>
and replacing linux with xen never worked
<whitequark>
was that on a mac?
<rqou>
why else would there be a rEFInd involved? :P
<rqou>
iirc this was on one of the macs that didn't default boot camp in EFI mode (and thus weren't 100% set up to boot "other" OSs via EFI)
<balrog>
rqou: you can thank ACPI for a lot of this
<rqou>
yeah
<rqou>
ACPI/EFI are both stupidly bloated
<balrog>
rqou: and you can thank Intel for that
<rqou>
whitequark: it seems interesting to me that you don't consider xen an inelegant mess
<rqou>
e.g. with all the stuff for handling the "no VT-x at all" use case and special dom0 modifications
<azonenberg_work>
rqou: why do they still have that
<azonenberg_work>
are there actually x86 processors without vt-x now?
<azonenberg_work>
my laptop from ten years ago had it
<rqou>
maybe?
<azonenberg_work>
i'm a big fan of aggressively deprecating cruft after a decade or so, if not sooner depending on the user base
<rqou>
the operating systems course lab TAs spent the entire first week showing people how to enable VT-x in the bios
<azonenberg_work>
So you don't end up like windows :p
<azonenberg_work>
o_O
<rqou>
and didn't succeed for everybody
<azonenberg_work>
there are still bioses without vt enabled out of the box?
<rqou>
yes
<azonenberg_work>
are there reasons to not enable it?
<rqou>
mumble mumble security?
<rqou>
don't want to get "blue pill" malware?
<rqou>
apparently some AV even did that to bypass patchguard
<rqou>
anyways, kvm has legacy too but arguably a lot more reasonable
<rqou>
kvm has cruft for handling lack of support for nested paging/real mode/"unrestricted guest"
<rqou>
azonenberg_work: apparently there are still some shitty Atoms that people might use that don't have VT-x
digshadow has quit [Ping timeout: 260 seconds]
<rqou>
but "unrestricted guest" has existed since Westmere so maybe we actually can start deprecating not having thay
<rqou>
*that
<daveshah>
was running an FPGA workshop using Vagrant so people could run the open source tools on Windoze
<daveshah>
Only one laptop didn't have VT-x
<daveshah>
and it was a 9-year-old low-end pentium
<rqou>
I've definitely seen "couldn't find bios option no matter how we looked"
<rqou>
maybe the bios was just borked
<rqou>
but yes, the operating systems course was basically doing the same thing (using vagrant to package a bunch of linux tools together for easy setup)
<rqou>
i personally didn't use any of that because the vagrant script just wrapped iirc puppet
<rqou>
so i just installed Ubuntu in lxc and ran the puppet script directly
<rqou>
hmm i just remembered that kvm also has a whole bunch of junk for handling MSRs and timers and stuff
<rqou>
not sure if that's the right place to put this stuff
<awygle>
i feel like the joyent folks had a talk about the internal horrors of kvm
<rqou>
i wonder how bhyve does it
<awygle>
they're fairly dramatic in general though
<rqou>
someone should try writing a "lhyve"
* awygle
still isn't convinced virtual machines are a good thing in the first palce
<rqou>
ime they can get really really lightweight
<rqou>
not sure if that's enough to make them a "good idea" :P
<awygle>
i feel like if linux containers weren't quite so broken, and if windows/mac weren't so protectionist, there'd be no reasons to run virtual machines
<rqou>
hey, xhyve is pretty cool
<awygle>
VMs, like Docker et al, feel like an arguably good solution to a problem that shouldn't exist
<rqou>
ime lxc containers work just fine
<awygle>
(the above does not apply to e.g. an ARM vm for developing on x86)
<awygle>
rqou: would you bet your security on one?
<rqou>
hmm... I'm not sure
<rqou>
i trust it more than docker :P
<awygle>
zones and jails are quite secure, they don't seem any more likely to be escaped than qemu
<awygle>
pretty sure the same is not true of linux containers
<rqou>
ok, user namespaces are a trashfire of vulns
<awygle>
mostly because linux doesn't believe in "containers" as an object (or didn't last time i checked)
<rqou>
but otherwise i don't see lxc as being worse than zones/jails
<rqou>
yeah, "containers" are built by carefully orchestrating namespaces and cgroups
<rqou>
also at this point i am betting security on lxc
<rqou>
since i run a build VM inside an lxc container
<awygle>
i mean, doesn't that mean you're at least as secure as the VM?
<rqou>
oh, not a real VM
<rqou>
just using lxc as if it were a VM
<awygle>
ah
<awygle>
anyway i'm basically ignorant on this issue, i'd be interested in hearing azonenberg_work's thoughts
<rqou>
azonenberg_work is weird and trusts VMware
<azonenberg_work>
rqou: i dont trust vmware exactly
<azonenberg_work>
i run a bunch of vmware instances on one computer
<azonenberg_work>
then i use my real computer to vnc into them
<azonenberg_work>
So there is potential if the hypervisor is popped to contaminate one vm to another
<azonenberg_work>
but it's a lot harder to exploit my client
<rqou>
wait, you didn't have the vnc step before?
<azonenberg_work>
No, i was running it locally
<azonenberg_work>
i plan to move from vmware to kvm or something as part of my IT revamp during the move
digshadow has joined ##openfpga
<azonenberg_work>
Since "good local 3D performance for gaming" was the main thing tying me to vmware before
<azonenberg_work>
that and unity mode, which i used heavily
<azonenberg_work>
they removed it in the recent versions
<azonenberg_work>
So basically now that i dont game in vms anymore, and unity mode is gone
<azonenberg_work>
i have no reason to use vmware anymore
<rqou>
pcie passthrough is really cool
<balrog>
vmware 3D performance has gone to shit?
<rqou>
can't use VMware for that though
<rqou>
it was historically full of vulns
<azonenberg_work>
pcie passthrough and 3d accel are always buggy
<azonenberg_work>
doesnt matter whose hypervisor you use
<azonenberg_work>
just one set of bugs or another
<rqou>
and (at the risk of making scanlime sad) in my experience the performance was never good
<rqou>
pcie passthrough seems to work for me once you set it up right
<rqou>
somebody eventually even fixed my workaround of (pretty much) reaching into the kernel and freeing a block of memory
<rqou>
inb4 gets a "wtf" and gets slapped by whitequark
<whitequark>
wtf
user10032 has joined ##openfpga
<rqou>
the issue was that you cannot pass through a device that's in use (obviously) so to pass through a gpu you need to make it unused
<rqou>
but if it's the boot gpu some memory gets claimed by the efi framebuffer driver
<rqou>
and (before they fixed it at least) no matter how you tried to disable the framebuffer driver it would still mark the memory region as used for io
<rqou>
even if the framebuffer driver doesn't load
<rqou>
so the hack was that i wrote a kernel driver that just called the function to deallocate io reserved regions
<rqou>
and just told it to deallocate the efi framebuffer region
<whitequark>
ok that's not too bad
<whitequark>
oh wtf
<whitequark>
xen doesn't really play well with overclocking
<whitequark>
well that means I have to go to kvm
fouric has quit [Quit: WeeChat 2.1]
fouric has joined ##openfpga
digshadow has quit [Quit: Leaving.]
digshadow has joined ##openfpga
fouric has quit [Client Quit]
fouric has joined ##openfpga
fouric has quit [Client Quit]
fouric has joined ##openfpga
<whitequark>
xen also adds major syscall overhead for builds in dom0
<gruetzkopf>
q3k: come over to cccac
pie_ has joined ##openfpga
<gruetzkopf>
G33KatWork: you too
<awygle>
today i encountered firewire for the first time in like a decade
<balrog>
awygle: what context?
<awygle>
i work at a company that's >40 years old, apparently at least one of our projects still uses it
user10032 has quit [Quit: Leaving]
<pie_>
rqou: high expectations asian *crosses out father* programmer: B-Tree? Why not A-Tree?
<rqou>
lol nice meme
<rqou>
weren't you here for the discussion where awygle pointed out that white people parents do that too? :P
<pie_>
i came up with it myself :D
<rqou>
I'm pretty sure I heard it before
<pie_>
aww...well its obvious enough i guess
<awygle>
is there some legal/NDA/whatever reason why BIOSes aren't open source?
<awygle>
(with some exceptions of course)
<jn__>
intel has NDAs for chipset documentation
<jn__>
/cpu
<awygle>
it just seems like nobody benefits really. the IHVs want to be V-ing H not writing code, the OSes have to deal with terrible firmware. is it the processor vendors that are the problems? seems like more and better support for their processors should be A Good
<jn__>
IP vendors are dicks about releasing documentation
<awygle>
i'm always sort of suspicious when the answer seems to be "these capitalists are too dumb to be optimal capitalists"
<jn__>
i don't even have proper documentation for the DRAM controller that SiFive used
<awygle>
hard to imagine what trade secrets "here is how to boot the thing" could reveal but i guess intel sees it differently
<pie_>
jn__, wow
<pie_>
jn__, fwiw i didnt actually think about it but in context, that sucks
<pie_>
just because the isa is oss, doesnt mean the hardware is :D (fml)
<jn__>
pie_: true, very true
<pie_>
hence my disappointment when i found out riscv is just the isa, but hey, whatcha gonna do.
<jn__>
the public documentation is quite usable as far SiFive can control it, but they didn't design their own DRAM controller
<awygle>
write an open source riscv core (and then get sad about PDKs)
<jn__>
that's so deep in the stack that i didn't think about it :)
<awygle>
apparently there are open PDKs that intentionally suck to hide process secrets but i've never bothered to look into it much
<awygle>
what level of DRAM does the sifive support?
<awygle>
DDR3?
<jn__>
DDR4, afaik
<awygle>
and is this the Super Linux-y One or the MCU?
<jn__>
the one that actually has DRAM ;)
<awygle>
ah :)
<awygle>
some MCUs have dram support now! M7's iirc
<jn__>
rather than "16KiB of SRAM, now BE HAPPY!"
<awygle>
not the 310 though apparently
<jn__>
i can understand if that's just because they didn't want to bother with NDA'd third-party IP unless necessary
<jn__>
for their first and second SoC…
<lain>
I was amused to see some new Intel stuff where they now allow you to use the large on-die caches as ram during startup
<lain>
so the uefi can come up, THEN bring up sdram, in case the dram has issues or etc
<jn__>
cache-as-ram is still an ugly hack :)
<lain>
I sort of want to make a board that uses an intel soc with no dram just because of this :P
<jn__>
heh
<lain>
but I don't have *that* much time to burn lol
<jn__>
or POWER9, if you're rich
<lain>
yesss
<lain>
POWER9 looks neat
<jn__>
(but i hear that their hostboot firmware is a pain to look at, too)
<lain>
haha
<awygle>
that's how the rpi works, isn't it?
<awygle>
except there's a crazy detour through, like, the GPU
<awygle>
"that" == "cache as ram"
<lain>
the rpi is a mess :D
<lain>
but yeah I have no idea
<awygle>
huh, it looks like the power company wants me to join a program where i pay more for power but it's all from renewable sources?
<awygle>
that's kind of cool, if somewhat weird...
<pie_>
maybe its not a scam
<awygle>
it looks like it would only cost me 3$ more a month, so i'll probably do it
<pie_>
for every 1$ they hook a battery potato into the power grid :D
<daveshah>
Plans like that are fairly common in the UK
<daveshah>
I think there was an all-nuclear option at one point too, if I recall
<pie_>
anyway, woo go renewables
<pie_>
can renewables have negative ecological effects? presumably still a lot better than nonrenewables
<pie_>
like, side effects of increasing albedo for solar, and i dunno?
<daveshah>
The impacts, whether significant or not, are from the production
<daveshah>
Albedo is very hard to change, no solar array is going to affect it realistically
<pie_>
windmills chopping up birds en masse? (cooking birds cartoon style with a solar array would bring a new meaning to fast food...hmmm)
bitd has quit [Quit: Leaving]
<awygle>
well hydropower tends to change the environment pretty drastically
<awygle>
because you build a dam and a reservoir
<awygle>
and there's all the side effects of making solar panels which i have to assume is a pretty nasty industrial process
<whitequark>
good thing it's all in china
<whitequark>
"cache as ram" is super common
<whitequark>
I think all mips boards use it
<whitequark>
I don't really think it's a hack though? I mean, cache *is* ram
<awygle>
yeah i don't have any particular problem with it
<lain>
I don't understand why companies are still trying to make hydrogen cars happen
<lain>
it's possible I've missed something, but last I checked hydrogen is insanely low energy density :P
<whitequark>
yup
<whitequark>
they should be using ammonia
<lain>
and I really wonder about environmental impact, because they're typically making it with a petroleum process iirc, or via electrolysis, which is offensively inefficient
<lain>
though I guess recently they've been using PEMs for the electrolysis which increases efficiency? I don't actually know how much..
<lain>
(and that electrolysis is probably powered by coal power, I would guess)
<awygle>
huh ammonia is surprisingly good
<lain>
:o
<lain>
methanal was a hot topic like 13 years ago in the fuel cell world, I wonder if anybody ever got it to work
<awygle>
ammonia is almost as good as methanol, volumetrically
<lain>
it was supposed to go through a whole bunch of stages, each one providing another electron or two, but they could never figure out the right catalyst(s) (this was circa 2005)
<awygle>
still less than half of gasoline of course
<awygle>
we should be using fission of U-235 :p
<lain>
haha
<lain>
we should each just have our own personal thorium reactor
<awygle>
(that's the top of this chart and the only one in scientific notation)
<lain>
that's safe, right?
<lain>
:P
<lain>
haha
<awygle>
only 8 orders of magnitude more dense than diesel!
<lain>
in /theory/ I like the idea of personal thorium reactors if they can run on a small enough quantity of fuel as to not be a MASSIVE DANGER TO PUBLIC SAFETY when joe blow decides to overclock/dismantle/otherwise damage their reactor :P
<rqou>
lain: don't become a boy scout tearing down smoke detectors and ashing lantern mantles :P
<rqou>
also RIP
<lain>
rqou: hahaha
<lain>
yeah that was sad, alcoholism iirc?
<rqou>
yeah
<rqou>
pros: became famous cons: ruined life
genii has quit [Remote host closed the connection]
<rqou>
also lain: how come nobody has freaked out about cesium/cobalt sources?
<rqou>
and actual incidents have happened with those
<lain>
doesn't make a catchy headline I guess?
<lain>
it's so freaking nice out.
<lain>
just thought y'all should know.
<awygle>
lain: :(
<lain>
hm.. maybe some time this weekend I can find time to debug a weird graphical corruption issue on this GPD Pocket
<whitequark>
awygle: I wonder if we made a mistake going for UP5K
<awygle>
whitequark: why's that? slow fabric?
<whitequark>
yes
<whitequark>
I guess we can heavily register everything
<awygle>
i don't really see that we had an alternative, other than the colossal QFP
<whitequark>
BGA
<whitequark>
sorear: re genome: nope, they removed all my data already
<awygle>
last i checked the 0.8 doesn't exist, we'd have ended up with 0.5
<whitequark>
and no such clauses in the contract
<rqou>
you can try cheating the tolerances
<awygle>
oh. it exists now. never mind.
<rqou>
whitequark: is there any way for me to mail you a blood sample and USD$3k and get a genome back?
<whitequark>
rqou: that was a discount provided by a startup for... I actually don't know why.
<whitequark>
maybe they just happened to have a sequencer and needed money?
<rqou>
lol
<whitequark>
it was limited to 2 people and I only got one there
<whitequark>
but if you shell up like 2x more, I think I can do that
<rqou>
hmm so USD$6k? i might consider it in the future
<whitequark>
more like $5k
<whitequark>
let me check the prices
<whitequark>
oh wtf
<whitequark>
genomed.ru does a full sequence for 100000 RUB
<whitequark>
that's 1600 USD
<rqou>
how?!
<whitequark>
90 day latency
<whitequark>
and you need to be there in person
<rqou>
and they sell your data? :P
<whitequark>
I don't think so
<whitequark>
also, Illumina declared "$1000 per genome" done in 2014
<rqou>
for raw reads like you did?
<rqou>
also, $100/genome when? :P
<awygle>
you just have to share your genome with 10 other people, and you don't get any I/O
<rqou>
awygle get out :P
<rqou>
also, human cloning when? :P
<rqou>
"herp derp wat r an ethics?"
* pie_
adds a watchlist
<awygle>
should make bingo cards
<rqou>
lol
<pie_>
watchlist bingo? thats great
<rqou>
whitequark: more seriously, do you have any thoughts as to when cloning/CRISPR will become accessible to a group of "rogue" hobbyists/biohackers?
<pie_>
rqou, you really want those catgirls dont you
<rqou>
lolol
<lain>
I am amused. the kanji (japanese) for caterpillar is fur and bug: 毛虫
<rqou>
no, there's an old whitequark thread on birbsite that you don't even need CRISPR to make catgirls
<rqou>
current plastic surgery technology is sufficient
<rqou>
pie_: you have seen the old meme, right? "every dollar spent on the war on drugs is a a dollar not spent on genetic engineering of catgirls <strike>for domestic ownership</strike>"
<pie_>
old signs: "support transhumanism for a better world!" new signs: "support transhumanism! make anime real!" "make catgirls not
<pie_>
ah well you were faster
<pie_>
and i lost the last half
<pie_>
if only genetic engineering catgirls wasnt morally wrong on so many levels :D
<rqou>
i wonder how many TLA analysts are now lurking here :P
<pie_>
they want to steal our catgirl propaganda and prospective methods
<pie_>
"hmmm when was azonenberg planning to have kids? :P"
* pie_
gets back to linear algebra...cant have catgirls without math
<rqou>
so a catgirl military might just work :P :P :P
<pie_>
i dont quite follow
<rqou>
make enemy soldiers too "excited" to actually fight effectively :P
<pie_>
OH
* pie_
leaves the beginnings of a felanthropist pun for someone else to finish
<rqou>
"make love, not war"
<pie_>
rqou, meanwhile, HERESY
<rqou>
?
<whitequark>
rqou: um
<whitequark>
the US military literally had that idea
<whitequark>
google "gay bomb"
<rqou>
i literally just posted that
<rqou>
scroll up :P
<whitequark>
oh
<whitequark>
oops
<rqou>
[14:28] (rqou) whitequark: more seriously, do you have any thoughts as to when cloning/CRISPR will become accessible to a group of "rogue" hobbyists/biohackers?
<whitequark>
not really
<whitequark>
but... most molecular biology is far more ghetto than you would expect
<whitequark>
btw if you want superconducting wire there's a vendor in russia
<whitequark>
We are the first producers of the 2 Generation High Temperature Superconductor (2G HTS) tape located in Moscow (Russia) and Tokyo (Japan).
<whitequark>
oh
<rqou>
whitequark: at this point in time, do you think a sufficiently "unethical" group of biohackers can already try things like CRISPR on humans? is the technology _already_ accessible?
<whitequark>
I think I know this company
<rqou>
also, that was also linked here just yesterday or so
<whitequark>
yes
<whitequark>
I'm reading the backlog
<whitequark>
define biohackers
<rqou>
idk, some group of motivated eccentric people with a BS degree?
<whitequark>
what about a lab?
<whitequark>
degree not necessary for sure lol
<rqou>
a ghetto home/garage lab like what the people in this channel could put together?
<pie_>
who would be crazy enough to *try* implanting themselves with an embryo hacked together in a garage
<whitequark>
you need to convince someone to do IVF
<whitequark>
and implantation yes
<lain>
I mean
<pie_>
well thats basically what i meant xD
<pie_>
all that predicating that you can even hack an embryo together in a garage
<whitequark>
if you mean trying CRISPR on somatic cells and not ova...
<whitequark>
the issue is that humans have an innate immune response to Cas9
<pie_>
*predicating on
<lain>
I read an article a long time ago about people getting "plastic surgery" to increase butt volume.. in someone's garage.. the person was arrested because 1/ that's not a legitimate medical practice, and 2/ they were actually injecting random crap like fix-a-flat and instant cement. which is all kinds of horrifying to think about :|
<whitequark>
since it's a bacterial complex
<Bike>
there was like a whole subreddit for transcranial stim so yeah someone's gonna inject themselves with whatever
<rqou>
no, on ova. but i was mostly thinking of whether you could even do the gene edit at all
<pie_>
lain, what the fuuuuuck
<lain>
so the problem of finding people crazy enough to consent to the experiments is probably a non-issue >_>
<lain>
<_____<
<awygle>
oh yeah i heard about that too
<rqou>
not the "convince a woman to try it" part
<whitequark>
Bike: well the question here is whether they can successfully do it
<lain>
just put out a craigslist ad
<lain>
:P
<Bike>
holding out for a DIY krukenberg procedure
<pie_>
Bike, i meant someone that isnt an idiot, but then again
<Bike>
whitequark: yeah i guess that's a higher standard
<Bike>
pie_: smart people inject themselves with dumb shit a fair bit
<lain>
cue*
<pie_>
lain, oh my god xD
<whitequark>
rqou: hard to say
<whitequark>
things are generally harder in humans than in even mice
<pie_>
oh wow krukenberg is both weird cool awesome and horrifying
<pie_>
uh...both = 4
<Bike>
right?
<rqou>
whitequark: what about an "unethical" group trying to clone a human?
<lain>
oh god why did I google that
<pie_>
lain, ok i dont think its thaaaaat bad
<Bike>
cloning doesn't seem inherently less ethical than just procreating old skool, assuming you're not messing up the kid's sonic hedgehog orwhatnot
<lain>
pie_: I mean, it's neat, but it's also probably going to haunt my dreams
<pie_>
maybe i just didnt look at it enoug
<pie_>
yeah thats kind of some silent hill shit lol
<whitequark>
what Bike says
<pie_>
Bike, messing up the kid's sonic hedgehog?
<whitequark>
it's a gene
<pie_>
oh
<Bike>
causes cyclopia
<whitequark>
if it's broken you get horrible deformities
* pie_
out of the loop
<Bike>
another thing you should seriously reconsider googling
<pie_>
yeaaaahh ill pass on that one
<whitequark>
and then have to explain to the parents why their child has "a problem with their sonic hedgehog"
<whitequark>
of course the gene was named by the fruit fly people
<awygle>
that's pretty horrifying
<whitequark>
if you ever find a gene with a fucked up name it's the fruit fly people. every time.
<awygle>
(also i thought bike meant circumcision)
<pie_>
lol
<whitequark>
uh
<awygle>
i didn't say i thought it made _sense_
<Bike>
bringing up circumcision in proximity to silent hill causes bad memories
<whitequark>
uhhhh
<pie_>
...i havent actually watched most of any of those movies
<awygle>
complicated bris, eh?
<rqou>
in terms of feasibility though, is cloning a human accessible to a group of hobbyists?
<pie_>
awygle, look, people make all kinds of weird allusions to dicks, so if you ever dont understand something, i suppose "dicks" is a good first guess
<whitequark>
rqou: "hobbyists" is a really wide class
<pie_>
some hobbyists have a lot of money
<whitequark>
i'm sure there are hobbyists equipped better than some proper labs
<rqou>
like azonenberg? :P
<whitequark>
yes
<pie_>
like that one guy with a sem and other things in his garage
<whitequark>
sam zeloof
<Bike>
i've taken a few trips to a lab with mosquitos and found that you feed them with your arm
<Bike>
but they have an airlock, so it's still science
<rqou>
ugh what
* awygle
shudders
<whitequark>
cheap, efficient
<pie_>
i guess theyre sterile mosquitos
<awygle>
lucky star taught me that if you clench up you can lock a mosquito in your arm long enough to kill it
<Bike>
obviously they're not infected with anything
<whitequark>
also you can use grad students as food literally
<whitequark>
what's not to love
<rqou>
lolol
<Bike>
yes, i know one of the food
<pie_>
mosquitos i guess
<whitequark>
awygle: now i have to rewatch lucky star
<awygle>
rqou: better hurry up and graduate :p
<whitequark>
good job
<awygle>
whitequark: lol i just finished
<pie_>
well if they werent infected before they fed on you...
<whitequark>
awygle: do you have a torrent
<Bike>
this is before we get into the odor study which reminds climbing into a plastic bag and laying on the ground as fans move your odor towards various bugs
<awygle>
i hadn't seen it in like eight years, it's Good
<rqou>
awygle: konata :3
<Bike>
which entails*
<awygle>
whitequark: not ready to hand, no, sorry
<pie_>
just pipe your /dev/sda over the internet
<rqou>
what could go wrong? :P
<whitequark>
there's rutracker but it doesn't have a bdremux
<awygle>
you know how WebRTC has TURN and STUN servers?
<pie_>
Bike, i fart in your general direction
<awygle>
i wonder what percentage of connections can be handled by STUN-like brokering vs requiring an actual third party
<pie_>
you know what could suck with cat ears? ....getting it caught on stuff
<rqou>
um...
<rqou>
real cats don't have this problem
<lain>
ahh man, I need to rewatch Lucky Star
<lain>
it's been so long
<whitequark>
##animefpga
<lain>
rqou: real cats are small
<pie_>
^ i was going to say that
<whitequark>
lain: um
<whitequark>
there are some pretty big cats
<lain>
well ok I guess there's like, big cats
<pie_>
then again
<lain>
but are their ears as big as catgirls/bois typically are depicted?
<lain>
maybe they are actually, hm
<pie_>
p r o p o r t i o n
* lain
shrugs
<pie_>
man now i need cuddles...fuck this
* pie_
goes back to math
Bike has quit [Ping timeout: 260 seconds]
* sorear
already barely fits in doorways
<rqou>
iirc whitequark mentioned a problem that we don't really have enough muscles we can hijack to control cat ears
<lain>
I should get back to work
<rqou>
working on making catgirls/catboys? :P
<lain>
I wish
<lain>
need to finish weeb practice, then work on some code
<pie_>
so....important question....functional or cosmetic cat ears
<whitequark>
you're not getting anywhere with functional
<whitequark>
without significant impairments
<whitequark>
you basically have to use a cochlear implant then
<pie_>
i didnt think it would work
<whitequark>
and those aren't very good
<rqou>
cosmetic but consciously controllable?
<whitequark>
yes, we still can't get a cochlear implant to work well, and some nutjobs are out there talking about uploading
<rqou>
wait what
<rqou>
i thought this already worked?
<whitequark>
they aren't as good as your natural hearing
<whitequark>
not even close
<rqou>
hmm
<whitequark>
not even the latest models with a shitton of DSP in them
<rqou>
but this is supposed to be an "easy" problem
<whitequark>
precisely
<rqou>
wtf
<whitequark>
this is easy compared to, like, pretty much anything else
<whitequark>
retina: mwahahahahaha
<whitequark>
your optic nerve is a part of your visual processing pipeline, and it does a nontrivial amount of it
<whitequark>
retina, too
<whitequark>
e.g. edge detection is already done there
<lain>
I'm kind of in shock right now because I was typing and felt something under my thumb when I smacked the spacebar, turns out it was a bug of some kind that I had just smashed (with the spacebar). I mayyyy have a panic response to bugs suddenly appearing very near / on me. so then I go to take a deep breath and I see this little pseudoscorpion hitching a ride on a fly... even more distressing :P
<awygle>
(warning, the page image is very good at eliciting the feeling described)
<lain>
I'm indoors why are these things happening :P
<whitequark>
oh, I was thinking about tryphophobia but haven't realized pepper seeds are realted
<pie_>
but what does that have to do with pepper seeds
<whitequark>
awygle: do you know why trypophobia exists
<whitequark>
basically your brain does a fourier transform because snakes have a high spatial frequency component
<whitequark>
it also happens to trigger on holes because it's not selective enough
<awygle>
seriously? but i like snakes! stupid brain...
<whitequark>
let me grab the paper
<lain>
dangit brain
<lain>
I wish I knew why I had an irrational fear of insects
<lain>
with very few exceptions
<awygle>
and yes, pepper seeds trigger it reliably in me, which has made my recent explorations of meal delivery services challenging
<pie_>
think about cute catgirlsthink about cute catgirls think about cute catgirls
<lain>
^^
<awygle>
yeah i'm super creeped out by insects too
* awygle
is a scardy-cat
<whitequark>
cute
<lain>
yeah they freak me right the heck out
<awygle>
i'm legitimately getting uncomfortable with these conversations actually lol. so i am going to take a for-real irc break. back later y'all.
<azonenberg_work>
lain: did you see the ADC i was talking about in ##fpga the other day?
<azonenberg_work>
ADC12DJ2700
<azonenberg_work>
2.7/5.4 Gsps dual/single channel, 12 bit, JESD204, about $1.8K each
<azonenberg_work>
So fairly straightforward high-end ADC
<azonenberg_work>
.... but
<lain>
whitequark: thanks. was taking out the recycle, saw that lil guy in there, had to drop everything and fetch my camera gear lol
<azonenberg_work>
it has the same bug as the altera serdes
<lain>
azonenberg_work: oh?
<azonenberg_work>
If you power down the device / a serdes channel for too long
<azonenberg_work>
the channel starts to fry
<lain>
D:
<azonenberg_work>
i'm VERY curious how this happens
<azonenberg_work>
because this implies a common physical mechanism
<azonenberg_work>
i doubt that altera is licensing somebody else's serdes, or selling their fpga transceiver
<azonenberg_work>
and its not even the same process, i dont think ti's adcs are made on an intel digital process
<lain>
huuuuh.
<azonenberg_work>
and yet they both have the EXACT same issue, unused serdes channels fail over time
<azonenberg_work>
"This pin disablesall analogcircuitsand serializeroutputswhenset high for temperaturediodecalibrationonly.Do not use this pin to powerdownthe devicefor powersavings.Tie this pin toGNDduringnormaloperation.For informationregardingreliableserializeroperation,see footnote(1)in thePin Functionstable"
<azonenberg_work>
(wow spacing fail)
<lain>
somehow I parsed that
<awygle>
azonenberg_work: decap and SEM? only a 1600$ investment to satisfy your curiosity :p
<q3k>
gruetzkopf: -ENOTIME :(
<azonenberg_work>
awygle: lolol
<azonenberg_work>
nooooope
<azonenberg_work>
i fully plan to buy several of those and make a DSO out of them
<azonenberg_work>
But not any time soon
<awygle>
why is it -ENOTIME instead of just ENOTIME?