1
00:00:00,560 --> 00:00:02,520
This. 
Is the Elon Musk podcast. 

2
00:00:02,520 --> 00:00:06,400
Your daily hit of what is really
going on at Tesla, SpaceX X AI. 

3
00:00:06,680 --> 00:00:08,480
And the rest of the Musk 
universe. 

4
00:00:08,760 --> 00:00:11,520
I'm your host Will Walden and I 
have covered Elon Musk for more 

5
00:00:11,520 --> 00:00:15,120
than five years, spent a year on
the ground at SpaceX, Starbase 

6
00:00:15,120 --> 00:00:17,120
during. 
Early Starship development and 

7
00:00:17,120 --> 00:00:20,600
before this I spent my career as
a software developer working 

8
00:00:20,600 --> 00:00:22,160
with billion dollar. 
Companies. 

9
00:00:22,480 --> 00:00:26,840
I've also built and sold my own 
businesses and now I make. 

10
00:00:26,840 --> 00:00:28,280
Content and. 
Help other people. 

11
00:00:28,280 --> 00:00:29,920
Grow. 
Their companies now on this 

12
00:00:29,920 --> 00:00:33,800
show, I use that experience to 
break down the news, filter out 

13
00:00:33,920 --> 00:00:38,520
all the noise and give you clear
context you can actually use. 

14
00:00:39,720 --> 00:00:43,120
Tesla has begun pushing Full 
Self Driving Supervised version 

15
00:00:43,120 --> 00:00:47,040
14.2 to cars in the field, and 
the company is framing this 

16
00:00:47,040 --> 00:00:49,960
build as the step that enables A
wider release. 

17
00:00:50,360 --> 00:00:53,760
Now the update centers on a new,
higher resolution vision encoder

18
00:00:54,080 --> 00:00:57,240
and a collection of driving 
behavior refinements that aim to

19
00:00:57,240 --> 00:00:59,840
make the car smoother and more 
confident. 

20
00:01:00,200 --> 00:01:03,280
The big question is whether this
roll out stays limited to a 

21
00:01:03,280 --> 00:01:07,920
small set of vehicles or expands
quickly to much of the fleet, as

22
00:01:07,920 --> 00:01:10,720
the company hints. 
Now today we're going to walk 

23
00:01:10,720 --> 00:01:14,760
through what changed in V142, 
what hardware is receiving at 

24
00:01:14,760 --> 00:01:18,640
first, how the release notes 
describe new emergency vehicle 

25
00:01:18,640 --> 00:01:22,000
handling and routing, and why 
the parking and arrival options 

26
00:01:22,000 --> 00:01:25,200
are a practical shift. 
We'll also look at the October 

27
00:01:25,200 --> 00:01:28,120
comment that set expectations 
for broader availability. 

28
00:01:28,600 --> 00:01:32,680
Then we're going to talk about 
what it's going to be like and 

29
00:01:32,680 --> 00:01:36,600
what a measured, widespread push
likely looks like in practice. 

30
00:01:37,040 --> 00:01:40,960
Now, after the break, I'll reset
with the Corfax and then move 

31
00:01:41,120 --> 00:01:45,880
feature by feature with clear 
examples pulled from Tesla's 

32
00:01:45,880 --> 00:01:49,800
notes. 
Now, Tesla started rolling out 

33
00:01:49,800 --> 00:01:53,600
FSD supervised 14.2, and the 
company ties this build to the 

34
00:01:53,600 --> 00:01:55,880
moment when a broader release 
could begin. 

35
00:01:56,320 --> 00:01:59,560
The release notes focused on 
vision emergency vehicle 

36
00:01:59,560 --> 00:02:03,320
protocols, routing that adapts 
to blockages in a set of new 

37
00:02:03,320 --> 00:02:06,880
parking choices with several 
scenarios specific improvements 

38
00:02:06,880 --> 00:02:10,720
listed plainly. 
Now with that baseline, let's 

39
00:02:10,759 --> 00:02:16,160
unpack the details and what they
add up to for drivers giving FSD

40
00:02:16,240 --> 00:02:19,960
on daily routes. 
Now this is a fully upgraded 

41
00:02:19,960 --> 00:02:22,840
neural network vision encoder 
that reads higher resolution 

42
00:02:22,840 --> 00:02:26,280
features from the scene. 
Now, in practice, that means the

43
00:02:26,280 --> 00:02:30,480
system should capture finer 
details and reduce ambiguity in 

44
00:02:30,480 --> 00:02:33,560
cluttered environments, which 
supports more decisive actions. 

45
00:02:34,160 --> 00:02:37,920
Now this connects directly to 
better detection of emergency 

46
00:02:37,920 --> 00:02:41,120
vehicles, obstacles and even 
human gestures. 

47
00:02:41,120 --> 00:02:45,120
So the camera stack is not just 
seeing more pixels, it is trying

48
00:02:45,120 --> 00:02:48,920
to interpret intent in motion 
any more granular way. 

49
00:02:49,560 --> 00:02:52,200
Now emergency vehicle behavior 
gets specific treatment with 

50
00:02:52,200 --> 00:02:54,960
this one. 
Tesla describes added logic for 

51
00:02:54,960 --> 00:02:57,880
controlled pullovers in yielding
behavior when the system 

52
00:02:57,880 --> 00:03:01,480
encounters police cars, fire 
trucks, and ambulances. 

53
00:03:02,040 --> 00:03:04,960
That's a practical step because 
these encounters carried legal 

54
00:03:04,960 --> 00:03:08,040
expectations and social norms 
that people take for granted, 

55
00:03:08,120 --> 00:03:10,840
and the software now aims to 
follow those norms with 

56
00:03:10,840 --> 00:03:13,480
structured decisions. 
If the system recognizes 

57
00:03:13,480 --> 00:03:16,520
flashing lights and adjusts 
predictably, it reduces 

58
00:03:16,520 --> 00:03:19,840
confusion for other drivers and 
for the occupants of the Tesla. 

59
00:03:20,880 --> 00:03:23,560
Now routing and navigation moves
deeper into the vision network 

60
00:03:23,560 --> 00:03:26,720
itself, which matters when the 
roads are blocked or detours 

61
00:03:26,720 --> 00:03:30,400
appear with a little warning. 
And the notes say that the car 

62
00:03:30,400 --> 00:03:33,280
can respond to block roads and 
detours in real time because 

63
00:03:33,280 --> 00:03:37,440
planning is more tightly coupled
to perception, That architecture

64
00:03:37,440 --> 00:03:40,840
should help the system replan 
without awkward pauses when a 

65
00:03:40,840 --> 00:03:44,440
lane closes or when a route 
changes near construction, which

66
00:03:44,440 --> 00:03:46,880
is where earlier versions 
sometimes hesitated. 

67
00:03:47,920 --> 00:03:50,600
Now there's several driving 
scenarios that receive targeted 

68
00:03:50,600 --> 00:03:54,640
refinements in this build. 
The notes call out unprotected 

69
00:03:54,640 --> 00:03:58,280
turns, lane changes, vehicle 
cut, insurance, and interactions

70
00:03:58,280 --> 00:04:01,160
with school buses. 
Each of these is a stress test 

71
00:04:01,160 --> 00:04:04,080
for prediction and yield 
behavior, so tuning them 

72
00:04:04,080 --> 00:04:06,400
suggests work on both comfort 
and compliance. 

73
00:04:06,600 --> 00:04:09,960
And when the card judges gaps 
more clearly on an unprotected 

74
00:04:09,960 --> 00:04:13,840
left, commits earlier on a safe 
lane change, or respects school 

75
00:04:13,840 --> 00:04:16,640
bus rules without abrupt 
breaking, the occupant 

76
00:04:16,640 --> 00:04:20,560
experience improves while risk 
drops dramatically. 

77
00:04:21,519 --> 00:04:24,360
Now there's also parking and 
arrival options, and they've 

78
00:04:24,360 --> 00:04:26,480
added a new layer of user 
control to that. 

79
00:04:26,960 --> 00:04:31,320
The release lists choices for 
where FSD should park or drop 

80
00:04:31,320 --> 00:04:35,120
off, such as a lot, the street, 
a driveway, a garage, or the 

81
00:04:35,120 --> 00:04:38,040
curbside. 
Now that matters for the last 50

82
00:04:38,040 --> 00:04:41,120
meters of a trip. 
Or earlier versions could second

83
00:04:41,120 --> 00:04:44,320
guess a destination or stop in a
suboptimal place a couple blocks

84
00:04:44,320 --> 00:04:48,600
away and just expect you to get 
out and parked there. 

85
00:04:48,840 --> 00:04:50,440
Well, it's most of the time that
doesn't work. 

86
00:04:50,760 --> 00:04:54,200
And if the driver can pre select
the arrival style, the system 

87
00:04:54,200 --> 00:04:57,360
can plan the final approach with
fewer last second adjustments, 

88
00:04:57,400 --> 00:05:00,440
which should feel more natural 
when you arrive at your 

89
00:05:00,440 --> 00:05:03,520
destination. 
Now the hardware scope is very 

90
00:05:03,520 --> 00:05:05,280
clear in the early hours of the 
rollout. 

91
00:05:05,640 --> 00:05:08,680
Reports that the update landed 
first on Model Y vehicles in 

92
00:05:08,680 --> 00:05:13,920
California with the company's 
A14 hardware set the expectation

93
00:05:13,920 --> 00:05:16,680
that this is not yet across 
every single car. 

94
00:05:17,440 --> 00:05:20,080
The notes also indicate that 
this wave is limited in terms of

95
00:05:20,080 --> 00:05:24,080
included vehicles, which points 
to a staged approach as Tesla 

96
00:05:24,080 --> 00:05:27,280
validates the new model and a 
narrower slice of the fleet 

97
00:05:27,280 --> 00:05:30,440
before pushing it more widely. 
That matches how complex 

98
00:05:30,440 --> 00:05:34,440
perception changes typically 
ship, and the version string in 

99
00:05:34,440 --> 00:05:42,800
the notes reads 2 O 25.38.9.5 
with FSD supervised V 14.2 

100
00:05:42,800 --> 00:05:47,320
installed now alongside the big 
items, the list includes smaller

101
00:05:47,320 --> 00:05:50,760
quality of life changes, like an
alert for residue buildup on the

102
00:05:50,760 --> 00:05:54,200
interior windshield that could 
degrade front camera visibility.

103
00:05:54,480 --> 00:05:57,520
That line is a reminder that 
sensor health is not just about 

104
00:05:57,520 --> 00:06:00,240
software, though. 
Real world grime can still blunt

105
00:06:00,240 --> 00:06:05,840
the vision system. 
Snow, dirt, mud, a leaf falling 

106
00:06:05,840 --> 00:06:07,320
in front of your camera, 
anything. 

107
00:06:07,320 --> 00:06:10,840
And the car now warns if that 
happens, so drivers can clean it

108
00:06:11,160 --> 00:06:15,320
and restore fidelity on their 
own and not rely on the system. 

109
00:06:16,280 --> 00:06:19,680
Tesla also lays out a few 
upcoming improvements attached 

110
00:06:19,680 --> 00:06:22,280
to this train, including 
continued to work on overall 

111
00:06:22,280 --> 00:06:25,600
smoothness, on parking spot 
selection, and on parking 

112
00:06:25,600 --> 00:06:28,040
quality. 
The company positions these as 

113
00:06:28,040 --> 00:06:31,800
the next steps rather than 
finished work inside a 14.2, 

114
00:06:32,000 --> 00:06:35,200
which signals that parking 
remains an active area of tuning

115
00:06:35,560 --> 00:06:37,560
now for drivers. 
That means the car should get 

116
00:06:37,560 --> 00:06:41,680
better at choosing A stall or a 
curbside space and it completing

117
00:06:41,680 --> 00:06:44,840
the maneuver without hunting. 
Now, context for the widespread 

118
00:06:44,840 --> 00:06:48,240
framing comes from a comment in 
October responding to a well 

119
00:06:48,240 --> 00:06:52,480
known tester who praised a prior
14 point 1.2 build for reducing 

120
00:06:52,600 --> 00:06:54,640
indecisive lane changes and 
braking. 

121
00:06:55,040 --> 00:06:59,000
Now, Elon Musk said that 14.2 is
for broad use. 

122
00:06:59,360 --> 00:07:02,920
That comment sets a bar for 
today's release, because if 14.2

123
00:07:02,920 --> 00:07:06,120
begins on a narrow hardware 
subset and then steps into a 

124
00:07:06,120 --> 00:07:09,600
much larger pool to performance,
people experienced on 14 Point, 

125
00:07:09,600 --> 00:07:13,720
1.2 should carry forward just 
with the higher resolution 

126
00:07:13,720 --> 00:07:19,560
vision and added behaviors. 
Now the least notes repeat 

127
00:07:19,560 --> 00:07:22,920
several times in ways that show 
emphasis, including improvements

128
00:07:22,920 --> 00:07:26,360
to handling static and dynamic 
gates and offsetting for Rd. 

129
00:07:26,360 --> 00:07:28,760
debris such as tires, branches, 
and boxes. 

130
00:07:29,280 --> 00:07:32,360
And in daily driving, that 
translates to fewer awkward 

131
00:07:32,360 --> 00:07:36,240
slowdowns, an object sits 
partially in the lane and better

132
00:07:36,240 --> 00:07:38,240
lateral positioning to give 
space. 

133
00:07:38,800 --> 00:07:42,320
Those are subtle changes that 
add up over a commute, cutting 

134
00:07:42,320 --> 00:07:46,240
down on small corrections that 
used to break the flow of the 

135
00:07:46,240 --> 00:07:49,640
drive. 
Now decision making reliability 

136
00:07:49,640 --> 00:07:53,120
gets a nod with language about 
better management of system 

137
00:07:53,200 --> 00:07:56,280
faults and smoother recovery 
from degraded operation. 

138
00:07:56,560 --> 00:07:59,800
So in practical terms, that 
means if a sub component has a 

139
00:08:00,000 --> 00:08:03,520
momentary hiccup, the system 
should not overreact or exit 

140
00:08:04,480 --> 00:08:06,280
automation abruptly without 
need. 

141
00:08:06,800 --> 00:08:10,000
Smooth recovery reduces driver 
workload and builds trust 

142
00:08:10,240 --> 00:08:12,800
because consistent behavior is 
easier to supervise than 

143
00:08:12,800 --> 00:08:15,440
behavior that changes 
dramatically and drastically 

144
00:08:15,440 --> 00:08:19,800
under small disturbances. 
Now there is also the question 

145
00:08:19,800 --> 00:08:21,920
of hardware diversity in the 
existing fleet. 

146
00:08:22,920 --> 00:08:26,160
The company's wide FSD fleet 
still includes a large number of

147
00:08:26,280 --> 00:08:31,840
HW-3 vehicles, and it remains to
be seen how 14.2 maps out of 

148
00:08:31,840 --> 00:08:36,640
that mix. 
If early validation on AI 14 

149
00:08:37,120 --> 00:08:41,200
looks strong, Tesla will need to
communicate how the vision 

150
00:08:41,200 --> 00:08:44,680
encoder and the new behavior 
scaled down to earlier hardware,

151
00:08:45,000 --> 00:08:47,280
or whether some features arrive 
in a trimmed form. 

152
00:08:47,680 --> 00:08:50,400
That's a normal step is 
architectures evolve. 

153
00:08:51,520 --> 00:08:56,040
Now Tesla is rolling out FSD 
Supervised V 14.2 with a higher 

154
00:08:56,040 --> 00:08:58,880
resolution vision encoder, 
explicit emergency vehicle 

155
00:08:58,880 --> 00:09:02,000
protocols, routing that adapts 
to detours, new parking and 

156
00:09:02,000 --> 00:09:05,320
arrival options, and a list of 
scenario level improvements. 

157
00:09:05,320 --> 00:09:09,840
The first wave appears limited 
to AI 14 hardware, with the 

158
00:09:09,840 --> 00:09:13,640
company signaling that this is 
the build meant for wider use 

159
00:09:13,640 --> 00:09:16,800
once validation supports it. 
If the earlier behavior holds 

160
00:09:16,800 --> 00:09:19,720
across more cars and hardware 
versions, the path to a broader 

161
00:09:19,720 --> 00:09:23,600
release is straightforward. 
Keep the comfort gains from 14.1

162
00:09:23,600 --> 00:09:27,480
point X, apply the high 
resolution perception and 

163
00:09:27,480 --> 00:09:32,280
expand, expand, expand. 
Hey, thank you so much for 

164
00:09:32,280 --> 00:09:34,480
listening today. 
I really do appreciate your 

165
00:09:34,480 --> 00:09:35,120
support. 
If. 

166
00:09:35,120 --> 00:09:37,800
You could take a second and hit 
this subscribe or the follow 

167
00:09:37,800 --> 00:09:40,000
button on whatever podcast 
platform that you're listening 

168
00:09:40,000 --> 00:09:42,560
on. 
Right now I greatly appreciate. 

169
00:09:42,560 --> 00:09:44,240
It it helps out the show 
tremendously. 

170
00:09:44,560 --> 00:09:46,240
And you'll never. 
Miss an episode? 

171
00:09:46,480 --> 00:09:49,240
And each episode is about 10 
minutes or less. 

172
00:09:49,440 --> 00:09:52,880
To get you caught up quickly and
please if you want to support 

173
00:09:52,880 --> 00:09:54,920
the show even more. 
Go to Patreon dot. 

174
00:09:54,920 --> 00:09:58,080
Com slash. 
Stage 0. 

175
00:09:58,560 --> 00:10:00,760
And please take care of 
yourselves and each other, and 

176
00:10:00,760 --> 00:10:38,880
I'll see you tomorrow. 
OK, so what's going on?