tmsr - Search: " T"

453700+ entries in 0.288s

asciilifeform: (at present everyone who reads it, also reads #b-a, but this is not guaranteed)

asciilifeform: phf: very interesting, and probably deserves to be posted to therealbitcoin mailinglist.

phf: asciilifeform: you might find this interesting http://glyf.org/tmp/bastard-delay.png. red dots are bastards when they are finally accepted, green dots are corresponding failed attempts

asciilifeform: no one in natoreich has so much as smelled $bil except by mercy of the crown. ☟︎

decimation: which is why it's amusing that the current generation thinks that launching smartphones into space is a great idea

asciilifeform: ... and this is just to get the 5-8 y. expected life.

decimation: yeah it's sop to overdesign satellites and expect failure

asciilifeform: but 'usable for the purpose for which was built' - different matter.

asciilifeform: i bet a good share of the scrap metal floating about in space is 'functional' by this definition

decimation: although nearly useless because of the decay in their radiothermal units

decimation: also I think both 'voyager' probes are still functional

decimation: it's mostly broken, but sometimes the batteries unshort and you can get hemisephric comms

decimation: but you don't want to be outside leo either, because of the radiation

decimation: atmospheric drag takes sats down

asciilifeform: important to remember that sat constellation is not a permanent animal in any sense

decimation: which I think is now actually owned by usg

decimation: maybe you are thinking of iridium

decimation: "An 840-satellite constellation named Teledesic went bust in 2002 before its first launch, even with the backing of Bill Gates, cell phone tycoon Craig McCaw and Saudi Arabian Prince Alwaleed bin Talal."

asciilifeform: perhaps somebody ~else~ is the mark this time

asciilifeform: perhaps i'm thinking of a previous iteration of this sc4m

decimation: same idea though

asciilifeform: ultimately they are gunning for u.s. state dept (yes) funding

assbot: OneWeb selects Airbus to build 900 Internet satellites | Spaceflight Now ... ( http://bit.ly/1Jkchbl )

asciilifeform: (repurposed ~rockets~ rather than nuke per se, yes)

asciilifeform: and - afaik - they're through

asciilifeform: but this was not a sufficiently 'mass' phenomenon to make a real difference

asciilifeform: there was a brief time in '90s when 'discount' launches using repurposed ru nukes on subs were to be had

assbot: Qualcomm teams with Branson on OneWeb satellite venture as SpaceX CEO pursues rival constellation - FierceWirelessTech ... ( http://bit.ly/1Jkc8EI )

asciilifeform: the actual economics of launches (and the basic tech) haven't changed in decades

decimation: viasat guy is pointing out that everyone who has tried - failed

asciilifeform: how this wins, other than by scamming usg, is unclear to me

decimation: oneweb is trying to make mega-constellation of leo sats for 'free internets fer all!'

asciilifeform: decimation: the sat folks are attempting a kind of chumpatron where 'let's build tiny sats that last 6 months'

decimation: http://spacenews.com/viasats-dankberg-unfazed-by-mega-constellation-hoopla/ < ViaSat Chief Executive Mark D. Dankberg said in a conference call with investors. ?There?s a question of whether making hundreds or thousands of satellites is a feature or a bug.?

decimation: asciilifeform: there's a similar effect in the space industry now adays

asciilifeform: yes, because they did not invent monetize-consumer-chumps as the mega-weapon

asciilifeform: esp. given as in '80s, 'plenty of folks risked, they all died'

decimation: asciilifeform: nearly all the 'evolution' in modern cpus can be considered to be reincarnating things ibm did in the 60's

asciilifeform: because literally no one can afford to take a risk

asciilifeform: where archs that would look entirely familiar to folks in 1960s are still considered 'state of the art'

asciilifeform: in practice, it means that, e.g., cpu world is as conservative as ancient egypt

decimation: so everyone derps in playground that can be had for $1k

decimation: yeah, it is a real problem for society when you literally need 10's to 100's of $bil to even try to 'make new ideas'

asciilifeform: no one makes 'intel 4004' in their kitchen any more than he can make 'pentium'

asciilifeform: ic put the mega-org (in practice, arm of the state) in the driver's seat in much the same way as nukes once did (if you ask mircea_popescu) or still do

decimation: yeah and the 'die shrink' has only made things worse

asciilifeform: this is why i believe that the discovery of integrated circuit was a serious 'degutenbergization' from which we have not even begun to recover. ☟︎

decimation: it becomes a vanity project for those who can print money

asciilifeform: 'wintel' gets to exist for strictly, strictly that

asciilifeform: because it is necessary to enable other rackets

decimation: it's the hardware equivilant for the software phenomenon - first turd to market wins

asciilifeform: only thing it can beat it on is cost, and the folks who give a fuck re: cost can't afford ~it~ either.

decimation: so for any kind of an open standard - it is impossible for 'generic' solution to beat specialized asic

decimation: yeah he said that his business is very difficult because any kind of processing that he could do with his chip - can be done with a specialized asic

asciilifeform: ideally - a scam directly on the folks who print money

asciilifeform: because there is no way to effectively amortize the astronomical costs without running a scam of one kind or another. ☟︎

decimation: plus third party 'ip cores'

asciilifeform: and god help you if your thing has to interact with components that you can't simulate (because you didn't design them, and what, other vendor is going to share its logic? dream)

decimation: and note that doesn't help you completely with silicon, as you have pointed out before

asciilifeform: not really sold in the usual sense. more of a lease, like the old ibm mainframes

decimation: which cost ~$100k themselves

asciilifeform: up to 'refrigerator' size

decimation: apparently they make massive 64-way fgpa boards to simulate a full chip

decimation: tried to simulate on fpga as best as he could

asciilifeform: the other things, that $m is pure sunk cost if what you get in the post is a dud that doesn't even power

decimation: apparently his eda tools were a special 'deal' for startups, with the understanding that he pays later

decimation: and he did confirm that it costs about $1m to do a chip rev, if you go through the fab's preferred 'agents'

decimation: asciilifeform: the parallella guy was complaining that fpgas have 'eaten' the dsp market

asciilifeform: but stranger things have happened

asciilifeform: i'd be surprised if archaeologists ever turn up a 'greenarrays' in usg gear

trinque: asciilifeform: it just clicked into place that these computers are precisely those field radios, as you pointed out.

decimation: mainly through its defense-contract departments like lockheed, etc

decimation: usg 'collects' technology in exactly this way

asciilifeform: (there's one lab at ft. something-or-other that has ~eternal~ support contract with symbolics co.)

decimation: nah, probably because someone who knew what they were doing made it work

asciilifeform: if usg, probably just buying it to sink in the ocean

asciilifeform: ('not shakespeare, but man by the same name')

asciilifeform: perhaps mircea_popescu, to run his bitcoind-in-forth

decimation: who knows how long they will last before they dry up and blow away

decimation: asciilifeform: the greenarrays site seems to imply they are in production

decimation: he said that he felt when writing anything higher-level than assembly, he feels like the code 'isn't his own'

decimation: yeah the same podcast interviewed vic antonic fellow, who was serb who made his own 8-bit computer back in the 80's

decimation: I tried to use it once, confusing as hell and also barely documented

decimation: I think chuck moore's 'colorforth' runs on windows

asciilifeform: then turd firmly cemented in.

asciilifeform: (other than chuck moore. who did this, yes. but he is old, and not sure if his thing is even still made)

asciilifeform: wake me up when somebody bakes a turing-complete gadget with no winblows used ANYWHERE in the brain-to-silicon road

decimation: the guy was a dsp designer and became disillusioned by all the shit that was stuffed into dsp chip

decimation: but they have their own silicon to perform dsp 'MAC' operation

decimation: asciilifeform: no, they use arm for board i/o control

asciilifeform: but just 'riff on old tune'

asciilifeform: afaik those folks are arm licensees

decimation: tried to sell 'new processor' idea on kickstarter, ended up hating kickstarter

decimation: he describes this process

assbot: An Interview with Andreas Olofsson - Adateva's Ampliative Abacus | The Amp Hour Electronics Podcast ... ( http://bit.ly/1Jkakvz )

decimation: asciilifeform: you might find this guy mildly interesting http://www.theamphour.com/254-an-interview-with-andreas-olofsson-adatevas-ampliative-abacus/

asciilifeform: typically takes a dozen or more.

decimation: and that's for one iteration

decimation: well, they are also useless if you don't have $1mil to get 'note from stalin' to make chips

decimation: asciilifeform: kinda related are the eda tools for designing chips