Ship failed to survive reentry. Not a surprise. Maybe lost too many tiles.
Overall, a real success! Some milestones achieved, others not.
And of course it is Pi Day today, so for all geeks everywhere…
View attachment 1027863
Hello from a denizen from other TMC forums, such as the TSLA Investors Discussion thread,
and occasionally subthreads of "AI, Autopilot, & Autonomous/FSD". I was quite impressed
by the launch today, and now see noted that Pi Day marks the founding of SpaceX. This is
clearly the place which has the most Pi Day related posts, so please accept my humble but
geeky contribution.
______
I've been thinking about Tesla's "march of nines" highlighted in earnings calls and other presentations
since 2020, largely applicable to autonomous driving,
https://www.carswithcords.net/2020/07/tesla-and-long-march-of-nines-to-full.html
At first glance the reference seems fanciful, since in much engineering work we’re
happy with an improvement of 10%, let alone a bigger breakthru of 90%, or one “nine”.
The term is squishy, but it’s been referenced in other fields. E.g. in well-controlled
clinical drug trials, having a “p-value” of < 0.01 augurs good statistical significance —
that is “two nines”. In physics there is an “n-sigma” event. For the discovery of the
Higgs boson, a 5-sigma experiment corresponds to a p-value of 3 x 10 to the minus 7,
or about 1 in 3.5 million, or in between six and seven nines.
https://www.scientificamerican.com/blog/observations/five-sigmawhats-that/
Relevant to FSD, there’s an injury collision about once every million driving miles,
and a fatality about once per 100M miles, so six to eight nines. But if you are
talking about the percentage of Tesla drive time one might actually use FSD,
it’s lower. I’m somewhere between one and two nines right now.
_______
OK, so now an anecdote about the great physicist Richard Feynman, who
used the notion of six nines, 999999, in a humorous way. This has to do
with the number pi, so happy pi day, today, March 14 …
Because pi appears random (every digit statistically seems to occur equally
often, and further testing on multi-digit sequences hints at “normality”),
Feynman was struck by the fact that 999999 occurs at position 762 of pi’s
decimal expansion. Known as the “Feyman point”, joker Feynman wanted
to memorize pi to that point, reciting the digits in a lecture and then after 762
say, “… nine, nine, nine, nine, nine, nine, and so on…” implying that it would
repeat nines forever, implying that pi is a rational number, ha!
Over at X née Twitter, the anecdote is here, with a reply thread:
A natural question is “what about seven 9’s or eight 9’s , or …?"
Some wag replied that he used Mathematica to pin down positions
in pi of 7- and 8- nines. I (being “loquitur” on that forum, too) replied that
you don’t even need to program anything to find out about nine nines--
just use ‘grep’ on any old Mac.
To wit, grab a sufficiently large number of digits of pi (I chose a
billion, but 1GB is little these days) from somewhere such as:
22.4 Trillion Digits of My Pi World Record
and then run a Unix command from the terminal, like
“fold -w 1000 bigfileofpi | grep -b —colour -n 999999999”
to see the march-of-nine nines light up in red! My primitive
command is not sophisticated enough to specify exactly which
position in pi a given digit string occurs (due to the vagaries of
‘fold’ inserting newlines), but also because a given string may
overlap the end of a text line and the beginning of the next.
Such is my story of nine nines on pi day.
—loquitur
P.S. There’s also the related recreational math activity of searching
for a name in pi, perhaps by using a suitably large encoding of
pi in base 27, allowing for English letters and a space.
P.P.S. Here’s a pic of the now-defunct Pi Bar near my neighborhood
in San Francisco:
