Slide 1
Slide 1 text
Holly Cummins
IBM + Red Hat
Things You Thought
You Didn’t Need To
Care About That Have a
Big Impact On Your Job
Slide 2
Slide 2 text
@hollycummins.com #Quarkus #IBM
invisible forces shaping your job
Slide 3
Slide 3 text
why is everything hard?
Slide 4
Slide 4 text
why are computers terrible?
Slide 5
Slide 5 text
geography
101
Slide 6
Slide 6 text
norway
geography
101
Slide 7
Slide 7 text
@hollycummins.com #Quarkus #IBM
Slide 8
Slide 8 text
@hollycummins.com #Quarkus #IBM
ski_regions:
Slide 9
Slide 9 text
@hollycummins.com #Quarkus #IBM
ski_regions:
- fr
Slide 10
Slide 10 text
@hollycummins.com #Quarkus #IBM
ski_regions:
- fr
- ch
Slide 11
Slide 11 text
@hollycummins.com #Quarkus #IBM
ski_regions:
- fr
- ch
- se
Slide 12
Slide 12 text
@hollycummins.com #Quarkus #IBM
ski_regions:
- fr
- ch
- se
- no
Slide 13
Slide 13 text
@hollycummins.com #Quarkus #IBM
ski_regions:
- fr
- ch
- se
- no
Slide 14
Slide 14 text
@hollycummins.com #Quarkus #IBM
ski_regions:
- fr
- ch
- se
- no
{"ski_regions": ["fr", "ch", "se", false]}
Slide 15
Slide 15 text
@hollycummins.com #Quarkus #IBM
ski_regions:
- fr
- ch
- se
- no
{"ski_regions": ["fr", "ch", "se", false]}
Slide 16
Slide 16 text
@hollycummins.com #Quarkus #IBM
ski_regions:
- fr
- ch
- se
- no
{"ski_regions": ["fr", "ch", "se", false]}
Slide 17
Slide 17 text
@hollycummins.com #Quarkus #IBM
ski_regions:
- fr
- ch
- se
- no
{"ski_regions": ["fr", "ch", "se", false]}
Slide 18
Slide 18 text
@hollycummins.com #Quarkus #IBM
ski_regions:
- fr
- ch
- se
- no
{"ski_regions": ["fr", "ch", "se", false]}
Slide 19
Slide 19 text
@hollycummins.com #Quarkus #IBM
ski_regions:
- fr
- ch
- se
- no
{"ski_regions": ["fr", "ch", "se", false]}
false
- no
Slide 20
Slide 20 text
@hollycummins.com #Quarkus #IBM
sensible decision had
bad outcome
Slide 21
Slide 21 text
how can we know what
outcomes our decisions
will have?
Slide 22
Slide 22 text
how can we make decisions
in a complex world?
Slide 23
Slide 23 text
@hollycummins.com #Quarkus #IBM
systems thinking
Slide 24
Slide 24 text
how can we make fast
decisions in an uncertain world?
Slide 25
Slide 25 text
statistics
101
Slide 26
Slide 26 text
@hollycummins.com #Quarkus #IBM
what do you call a
statistician who lives
in san francisco?
Slide 27
Slide 27 text
geography
201
Slide 28
Slide 28 text
@hollycummins.com #Quarkus #IBM
what do you call a statistician who lives in san francisco?
Slide 29
Slide 29 text
@hollycummins.com #Quarkus #IBM
what do you call a statistician who lives in san francisco?
a data scientist.
Slide 30
Slide 30 text
@hollycummins.com #Quarkus #IBM
statistics
Slide 31
Slide 31 text
@hollycummins.com #Quarkus #IBM
data science
statistics
Slide 32
Slide 32 text
@hollycummins.com #Quarkus #IBM
data science
statistics
machine learning
Slide 33
Slide 33 text
@hollycummins.com #Quarkus #IBM
data science
statistics
AI
machine learning
Slide 34
Slide 34 text
@hollycummins.com #Quarkus #IBM
even if you’re not doing AI,
statistics is important
Slide 35
Slide 35 text
Bayes’
theorem
statistics
101
Slide 36
Slide 36 text
how can we detect spam?
Slide 37
Slide 37 text
@hollycummins.com #Quarkus #IBM
“spamicity”
Slide 38
Slide 38 text
No content
Slide 39
Slide 39 text
Exciting offer just for you!
Deep discounts on Viagra.
Limited time offer, act now.
Slide 40
Slide 40 text
Exciting offer just for you!
Deep discounts on Viagra.
Limited time offer, act now.
Slide 41
Slide 41 text
Exciting offer just for you!
Deep discounts on Ozempic.
Limited time offer, act now.
Slide 42
Slide 42 text
Exciting offer just for you!
Deep discounts on Ozempic.
Limited time offer, act now.
Slide 43
Slide 43 text
No content
Slide 44
Slide 44 text
banned-words:
Slide 45
Slide 45 text
banned-words:
- Ozempic
Slide 46
Slide 46 text
banned-words:
- Ozempic
Slide 47
Slide 47 text
banned-words:
- Ozempic
Slide 48
Slide 48 text
Exciting offer just for you!
Deep discounts on Ozemp1c.
Limited time offer, act now.
Slide 49
Slide 49 text
Exciting offer just for you!
Deep discounts on 0zempic.
Limited time offer, act now.
Slide 50
Slide 50 text
Exciting offer just for you!
Deep discounts on Ozempick.
Limited time offer, act now.
Slide 51
Slide 51 text
No content
Slide 52
Slide 52 text
banned-words:
Slide 53
Slide 53 text
banned-words:
- Ozempic
Slide 54
Slide 54 text
banned-words:
- Ozempic
- Ozempick
Slide 55
Slide 55 text
banned-words:
- Ozempic
- Ozempick
- Ozempik
Slide 56
Slide 56 text
banned-words:
- Ozempic
- Ozempick
- Ozempik
- 0zempic
Slide 57
Slide 57 text
banned-words:
- Ozempic
- Ozempick
- Ozempik
- 0zempic
- Ozemp1c
Slide 58
Slide 58 text
banned-words:
- Ozempic
- Ozempick
- Ozempik
- 0zempic
- Ozemp1c
- Øzempic
Slide 59
Slide 59 text
banned-words:
- Ozempic
- Ozempick
- Ozempik
- 0zempic
- Ozemp1c
- Øzempic
- 0zempick
Slide 60
Slide 60 text
banned-words:
- Ozempic
- Ozempick
- Ozempik
- 0zempic
- Ozemp1c
- Øzempic
- 0zempick
- …
Slide 61
Slide 61 text
banned-words:
- Ozempic
- Ozempick
- Ozempik
- 0zempic
- Ozemp1c
- Øzempic
- 0zempick
- …
Slide 62
Slide 62 text
Dear friend,
Exciting offer just for you! Deep discounts on Ozemp1c.
Limited time offer, act now.
Slide 63
Slide 63 text
Dear friend,
Exciting offer just for you! Deep discounts on Ozemp1c.
Limited time offer, act now.
Hey Alice,
Are you still good for next week? We had drama yesterday
because the dog threw up, on the clean dishes in the
dishwasher! On the bright side, at least it was easy to
clean. I started my Ozempic on Sunday, and so far so
good. I can see my appetite is less already.
Slide 64
Slide 64 text
Dear friend,
Exciting offer just for you! Deep discounts on Ozemp1c.
Limited time offer, act now.
Hey Alice,
Are you still good for next week? We had drama yesterday
because the dog threw up, on the clean dishes in the
dishwasher! On the bright side, at least it was easy to
clean. I started my Ozempic on Sunday, and so far so
good. I can see my appetite is less already.
Ozemp1c
Slide 65
Slide 65 text
Dear friend,
Exciting offer just for you! Deep discounts on Ozemp1c.
Limited time offer, act now.
Hey Alice,
Are you still good for next week? We had drama yesterday
because the dog threw up, on the clean dishes in the
dishwasher! On the bright side, at least it was easy to
clean. I started my Ozempic on Sunday, and so far so
good. I can see my appetite is less already.
Ozemp1c
offer
Slide 66
Slide 66 text
Dear friend,
Exciting offer just for you! Deep discounts on Ozemp1c.
Limited time offer, act now.
Hey Alice,
Are you still good for next week? We had drama yesterday
because the dog threw up, on the clean dishes in the
dishwasher! On the bright side, at least it was easy to
clean. I started my Ozempic on Sunday, and so far so
good. I can see my appetite is less already.
Ozemp1c
yesterday
offer
Slide 67
Slide 67 text
Dear friend,
Exciting offer just for you! Deep discounts on Ozemp1c.
Limited time offer, act now.
Hey Alice,
Are you still good for next week? We had drama yesterday
because the dog threw up, on the clean dishes in the
dishwasher! On the bright side, at least it was easy to
clean. I started my Ozempic on Sunday, and so far so
good. I can see my appetite is less already.
Ozemp1c
dishwasher
yesterday
offer
Slide 68
Slide 68 text
Dear friend,
Exciting offer just for you! Deep discounts on Ozemp1c.
Limited time offer, act now.
Hey Alice,
Are you still good for next week? We had drama yesterday
because the dog threw up, on the clean dishes in the
dishwasher! On the bright side, at least it was easy to
clean. I started my Ozempic on Sunday, and so far so
good. I can see my appetite is less already.
Ozemp1c
dishwasher
yesterday
offer
Slide 69
Slide 69 text
Dear friend,
Exciting offer just for you! Deep discounts on Ozemp1c.
Limited time offer, act now.
Hey Alice,
Are you still good for next week? We had drama yesterday
because the dog threw up, on the clean dishes in the
dishwasher! On the bright side, at least it was easy to
clean. I started my Ozempic on Sunday, and so far so
good. I can see my appetite is less already.
Ozemp1c
dishwasher
yesterday
offer
Slide 70
Slide 70 text
offer
act
discount
Ozemp1c
dog
dishwasher
yesterday
Ozempic
Slide 71
Slide 71 text
dog
dishwasher
yesterday
Ozempic
Slide 72
Slide 72 text
No content
Slide 73
Slide 73 text
@hollycummins.com #Quarkus #IBM
P(spam | 0zemp1c)
Slide 74
Slide 74 text
@hollycummins.com #Quarkus #IBM
P(spam | 0zemp1c)
“probability of”
Slide 75
Slide 75 text
@hollycummins.com #Quarkus #IBM
P(spam | 0zemp1c)
“given”
“probability of”
Slide 76
Slide 76 text
@hollycummins.com #Quarkus #IBM
P(A | B) =
P(B | A) · P(A)
P(B)
Slide 77
Slide 77 text
@hollycummins.com #Quarkus #IBM
probability of A, given B
P(A | B) =
P(B | A) · P(A)
P(B)
Slide 78
Slide 78 text
@hollycummins.com #Quarkus #IBM
probability of A, given B probability of
B, given A
P(A | B) =
P(B | A) · P(A)
P(B)
Slide 79
Slide 79 text
@hollycummins.com #Quarkus #IBM
probability of A, given B probability of
B, given A probability of A
P(A | B) =
P(B | A) · P(A)
P(B)
Slide 80
Slide 80 text
@hollycummins.com #Quarkus #IBM
probability of A, given B probability of
B, given A probability of A
probability of B
P(A | B) =
P(B | A) · P(A)
P(B)
Slide 81
Slide 81 text
@hollycummins.com #Quarkus #IBM
probability of A, given B probability of
B, given A probability of A
probability of B
guidance for a decision
P(A | B) =
P(B | A) · P(A)
P(B)
Slide 82
Slide 82 text
@hollycummins.com #Quarkus #IBM
P(spam | Ozemp1c) =
P(Ozemp1c | spam) · P(spam)
P(Ozemp1c)
Slide 83
Slide 83 text
@hollycummins.com #Quarkus #IBM
is this email spam?
P(spam | Ozemp1c) =
P(Ozemp1c | spam) · P(spam)
P(Ozemp1c)
Slide 84
Slide 84 text
@hollycummins.com #Quarkus #IBM
is this email spam?
what proportion of
spam talks about
Ozemp1c?
P(spam | Ozemp1c) =
P(Ozemp1c | spam) · P(spam)
P(Ozemp1c)
Slide 85
Slide 85 text
@hollycummins.com #Quarkus #IBM
is this email spam?
what proportion of
spam talks about
Ozemp1c?
what proportion
of emails are
spam?
P(spam | Ozemp1c) =
P(Ozemp1c | spam) · P(spam)
P(Ozemp1c)
Slide 86
Slide 86 text
@hollycummins.com #Quarkus #IBM
is this email spam?
what proportion of
spam talks about
Ozemp1c?
what proportion
of emails are
spam?
what proportion of all
emails mention Ozemp1c?
P(spam | Ozemp1c) =
P(Ozemp1c | spam) · P(spam)
P(Ozemp1c)
Slide 87
Slide 87 text
@hollycummins.com #Quarkus #IBM
P(spam | Ozemp1c) =
P(Ozemp1c | spam) · P(spam)
P(Ozemp1c)
Slide 88
Slide 88 text
@hollycummins.com #Quarkus #IBM
P(spam | Ozemp1c) =
P(Ozemp1c | spam) · P(spam)
P(Ozemp1c)
1%
Slide 89
Slide 89 text
@hollycummins.com #Quarkus #IBM
P(spam | Ozemp1c) =
P(Ozemp1c | spam) · P(spam)
P(Ozemp1c)
1%
0.001%
Slide 90
Slide 90 text
@hollycummins.com #Quarkus #IBM
P(spam | Ozemp1c) =
P(Ozemp1c | spam) · P(spam)
P(Ozemp1c)
1% 6%
0.001%
Slide 91
Slide 91 text
@hollycummins.com #Quarkus #IBM
P(spam | Ozemp1c) =
P(Ozemp1c | spam) · P(spam)
P(Ozemp1c)
1% 6%
0.001%
60%
Slide 92
Slide 92 text
@hollycummins.com #Quarkus #IBM
P(A | B) =
P(B | A) · P(A)
P(B)
Slide 93
Slide 93 text
@hollycummins.com #Quarkus #IBM
P(A | B)
what if I don’t know any of these numbers??
=
P(B | A) · P(A)
P(B)
Slide 94
Slide 94 text
@hollycummins.com #Quarkus #IBM
P(A | B)
what if I don’t know any of these numbers??
or this one?
=
P(B | A) · P(A)
P(B)
Slide 95
Slide 95 text
@hollycummins.com #Quarkus #IBM
P(A | B)
what if I don’t know any of these numbers??
or this one?
most useless
formula ever?
=
P(B | A) · P(A)
P(B)
Slide 96
Slide 96 text
@hollycummins.com #Quarkus #IBM
P(A | B) =
P(B | A) · P(A)
P(B)
Slide 97
Slide 97 text
@hollycummins.com #Quarkus #IBM
guess!
P(A | B) =
P(B | A) · P(A)
P(B)
Slide 98
Slide 98 text
@hollycummins.com #Quarkus #IBM
guess! guess!
P(A | B) =
P(B | A) · P(A)
P(B)
Slide 99
Slide 99 text
@hollycummins.com #Quarkus #IBM
guess! guess!
guess!
or ignore, because it’s a normalizing factor
P(A | B) =
P(B | A) · P(A)
P(B)
Slide 100
Slide 100 text
@hollycummins.com #Quarkus #IBM
P(A | B) =
P(B | A) · P(A)
P(B)
P(A
Slide 101
Slide 101 text
@hollycummins.com #Quarkus #IBM
P(A | B) =
P(B | A) · P(A)
P(B)
P(A
Slide 102
Slide 102 text
@hollycummins.com #Quarkus #IBM
recursion
P(A | B) =
P(B | A) · P(A)
P(B)
P(A
Slide 103
Slide 103 text
@hollycummins.com #Quarkus #IBM
recursion
P(A | B) =
P(B | A) · P(A)
P(B)
P(A
(cheating)
P(A
P(A)
Slide 104
Slide 104 text
@hollycummins.com #Quarkus #IBM
recursion
P(A | B) =
P(B | A) · P(A)
P(B)
P(A
(cheating)
(iteration)
Slide 105
Slide 105 text
@hollycummins.com #Quarkus #IBM
Slide 106
Slide 106 text
@hollycummins.com #Quarkus #IBM
"ignorance... coined into science."
Slide 107
Slide 107 text
@hollycummins.com #Quarkus #IBM
"ignorance... coined into science."
“a staggering falsity”
Slide 108
Slide 108 text
@hollycummins.com #Quarkus #IBM
"ignorance... coined into science."
“a staggering falsity”
“perhaps the only error to which
the mathematical world has so
deeply committed itself”
Slide 109
Slide 109 text
@hollycummins.com #Quarkus #IBM
Slide 110
Slide 110 text
@hollycummins.com #Quarkus #IBM
Slide 111
Slide 111 text
@hollycummins.com #Quarkus #IBM
cowpox?
Slide 112
Slide 112 text
@hollycummins.com #Quarkus #IBM
cowpox?
chickenpox?
Slide 113
Slide 113 text
@hollycummins.com #Quarkus #IBM
P(chickenpox | spots)
P(cowpox | spots)
=
=
P(spots | chickenpox) · P(chickenpox)
P(spots)
P(spots | cowpox) · P(cowpox)
P(spots)
Slide 114
Slide 114 text
@hollycummins.com #Quarkus #IBM
P(chickenpox | spots)
P(cowpox | spots)
=
=
P(spots | chickenpox) · P(chickenpox)
P(spots)
P(spots | cowpox) · P(cowpox)
P(spots)
Slide 115
Slide 115 text
@hollycummins.com #Quarkus #IBM
P(chickenpox | spots)
P(cowpox | spots)
=
=
P(spots | chickenpox) · P(chickenpox)
P(spots)
P(spots | cowpox) · P(cowpox)
P(spots)
85%
Slide 116
Slide 116 text
@hollycummins.com #Quarkus #IBM
P(chickenpox | spots)
P(cowpox | spots)
=
=
P(spots | chickenpox) · P(chickenpox)
P(spots)
P(spots | cowpox) · P(cowpox)
P(spots)
85% 1%
Slide 117
Slide 117 text
@hollycummins.com #Quarkus #IBM
P(chickenpox | spots)
P(cowpox | spots)
=
=
P(spots | chickenpox) · P(chickenpox)
P(spots)
P(spots | cowpox) · P(cowpox)
P(spots)
85% 1%
90%
Slide 118
Slide 118 text
@hollycummins.com #Quarkus #IBM
P(chickenpox | spots)
P(cowpox | spots)
=
=
P(spots | chickenpox) · P(chickenpox)
P(spots)
P(spots | cowpox) · P(cowpox)
P(spots)
85% 1%
90% 0.0001%
Slide 119
Slide 119 text
@hollycummins.com #Quarkus #IBM
P(chickenpox | spots)
P(cowpox | spots)
=
=
P(spots | chickenpox) · P(chickenpox)
P(spots)
P(spots | cowpox) · P(cowpox)
P(spots)
85% 1%
90% 0.0001%
0.85%
Slide 120
Slide 120 text
@hollycummins.com #Quarkus #IBM
P(chickenpox | spots)
P(cowpox | spots)
=
=
P(spots | chickenpox) · P(chickenpox)
P(spots)
P(spots | cowpox) · P(cowpox)
P(spots)
85% 1%
90% 0.0001%
0.85%
0.0009%
Slide 121
Slide 121 text
@hollycummins.com #Quarkus #IBM
P(chickenpox | spots)
P(cowpox | spots)
=
=
P(spots | chickenpox) · P(chickenpox)
P(spots)
P(spots | cowpox) · P(cowpox)
P(spots)
85% 1%
90% 0.0001%
0.85%
0.0009%
Slide 122
Slide 122 text
@hollycummins.com #Quarkus #IBM
it’s chickenpox
Slide 123
Slide 123 text
@hollycummins.com #Quarkus #IBM
common sense
it’s chickenpox
Slide 124
Slide 124 text
@hollycummins.com #Quarkus #IBM
magic static analysis tool
99% accurate
Slide 125
Slide 125 text
@hollycummins.com #Quarkus #IBM
✓ code is clean ☺
Slide 126
Slide 126 text
@hollycummins.com #Quarkus #IBM
what’s the chance the code is clean?
✓ code is clean ☺
Slide 127
Slide 127 text
@hollycummins.com #Quarkus #IBM
P(clean code | clean report) =
P(clean report | clean code) · P(clean code)
P(clean report)
Slide 128
Slide 128 text
@hollycummins.com #Quarkus #IBM
P(clean code | clean report) =
P(clean report | clean code) · P(clean code)
P(clean report)
99%
Slide 129
Slide 129 text
@hollycummins.com #Quarkus #IBM
P(clean code | clean report) =
P(clean report | clean code) · P(clean code)
P(clean report)
99%
1.98%
Slide 130
Slide 130 text
@hollycummins.com #Quarkus #IBM
P(clean code | clean report) =
P(clean report | clean code) · P(clean code)
P(clean report)
99% 1%
1.98%
Slide 131
Slide 131 text
@hollycummins.com #Quarkus #IBM
P(clean code | clean report) =
P(clean report | clean code) · P(clean code)
P(clean report)
99% 1%
1.98%
50%
Slide 132
Slide 132 text
@hollycummins.com #Quarkus #IBM
not common-sense
Slide 133
Slide 133 text
@hollycummins.com #Quarkus #IBM
bayes is everywhere
Slide 134
Slide 134 text
@hollycummins.com #Quarkus #IBM
artificial intelligence
Slide 135
Slide 135 text
@hollycummins.com #Quarkus #IBM
medicine
Slide 136
Slide 136 text
@hollycummins.com #Quarkus #IBM
top tip
Slide 137
Slide 137 text
@hollycummins.com #Quarkus #IBM
tired: “we have no idea”
top tip
Slide 138
Slide 138 text
@hollycummins.com #Quarkus #IBM
tired: “we have no idea”
wired: “we are adopting bayesian thinking”
top tip
Slide 139
Slide 139 text
why does concurrency
make my head hurt?
Slide 140
Slide 140 text
@hollycummins.com #Quarkus #IBM
concurrency is hard
Slide 141
Slide 141 text
@hollycummins.com #Quarkus #IBM
amdahl’s law
Slide 142
Slide 142 text
@hollycummins.com #Quarkus #IBM
𝖯
(
𝖼𝗈 𝗇𝖼
𝗎 𝗋
𝗋𝖾 𝗇𝖼
𝗒
∣
𝗁
𝖾𝖺𝖽𝖺 𝖼𝗁𝖾
) =
𝖯
(
𝗁 𝖾 𝖺
𝖽 𝖺𝖼
𝗁𝖾
∣
𝖼𝗈𝗇
𝖼 𝗎𝗋𝗋
𝖾𝗇
𝖼𝗒
) ⋅
𝖯
(
𝖼𝗈𝗇
𝖼𝗎𝗋𝗋
𝖾𝗇
𝖼𝗒
)
𝖯
(
𝗁
𝖾𝖺𝖽
𝖺 𝖼𝗁𝖾
)
Slide 143
Slide 143 text
@hollycummins.com #Quarkus #IBM
𝖯
(
𝖼𝗈 𝗇𝖼
𝗎 𝗋
𝗋𝖾 𝗇𝖼
𝗒
∣
𝗁
𝖾𝖺𝖽𝖺 𝖼𝗁𝖾
) =
𝖯
(
𝗁 𝖾 𝖺
𝖽 𝖺𝖼
𝗁𝖾
∣
𝖼𝗈𝗇
𝖼 𝗎𝗋𝗋
𝖾𝗇
𝖼𝗒
) ⋅
𝖯
(
𝖼𝗈𝗇
𝖼𝗎𝗋𝗋
𝖾𝗇
𝖼𝗒
)
𝖯
(
𝗁
𝖾𝖺𝖽
𝖺 𝖼𝗁𝖾
)
can I blame concurrency for
my headache?
Slide 144
Slide 144 text
@hollycummins.com #Quarkus #IBM
𝖯
(
𝖼𝗈 𝗇𝖼
𝗎 𝗋
𝗋𝖾 𝗇𝖼
𝗒
∣
𝗁
𝖾𝖺𝖽𝖺 𝖼𝗁𝖾
) =
𝖯
(
𝗁 𝖾 𝖺
𝖽 𝖺𝖼
𝗁𝖾
∣
𝖼𝗈𝗇
𝖼 𝗎𝗋𝗋
𝖾𝗇
𝖼𝗒
) ⋅
𝖯
(
𝖼𝗈𝗇
𝖼𝗎𝗋𝗋
𝖾𝗇
𝖼𝗒
)
𝖯
(
𝗁
𝖾𝖺𝖽
𝖺 𝖼𝗁𝖾
)
can I blame concurrency for
my headache?
how many devs can cope
with concurrency?
Slide 145
Slide 145 text
@hollycummins.com #Quarkus #IBM
𝖯
(
𝖼𝗈 𝗇𝖼
𝗎 𝗋
𝗋𝖾 𝗇𝖼
𝗒
∣
𝗁
𝖾𝖺𝖽𝖺 𝖼𝗁𝖾
) =
𝖯
(
𝗁 𝖾 𝖺
𝖽 𝖺𝖼
𝗁𝖾
∣
𝖼𝗈𝗇
𝖼 𝗎𝗋𝗋
𝖾𝗇
𝖼𝗒
) ⋅
𝖯
(
𝖼𝗈𝗇
𝖼𝗎𝗋𝗋
𝖾𝗇
𝖼𝗒
)
𝖯
(
𝗁
𝖾𝖺𝖽
𝖺 𝖼𝗁𝖾
)
can I blame concurrency for
my headache?
how many devs can cope
with concurrency?
how much
concurrency is there?
Slide 146
Slide 146 text
@hollycummins.com #Quarkus #IBM
𝖯
(
𝖼𝗈 𝗇𝖼
𝗎 𝗋
𝗋𝖾 𝗇𝖼
𝗒
∣
𝗁
𝖾𝖺𝖽𝖺 𝖼𝗁𝖾
) =
𝖯
(
𝗁 𝖾 𝖺
𝖽 𝖺𝖼
𝗁𝖾
∣
𝖼𝗈𝗇
𝖼 𝗎𝗋𝗋
𝖾𝗇
𝖼𝗒
) ⋅
𝖯
(
𝖼𝗈𝗇
𝖼𝗎𝗋𝗋
𝖾𝗇
𝖼𝗒
)
𝖯
(
𝗁
𝖾𝖺𝖽
𝖺 𝖼𝗁𝖾
)
can I blame concurrency for
my headache?
how many devs can cope
with concurrency?
how often do heads hurt?
how much
concurrency is there?
Slide 147
Slide 147 text
@hollycummins.com #Quarkus #IBM
𝖯
(
𝖼𝗈 𝗇𝖼
𝗎 𝗋
𝗋𝖾 𝗇𝖼
𝗒
∣
𝗁
𝖾𝖺𝖽𝖺 𝖼𝗁𝖾
) =
𝖯
(
𝗁 𝖾 𝖺
𝖽 𝖺𝖼
𝗁𝖾
∣
𝖼𝗈𝗇
𝖼 𝗎𝗋𝗋
𝖾𝗇
𝖼𝗒
) ⋅
𝖯
(
𝖼𝗈𝗇
𝖼𝗎𝗋𝗋
𝖾𝗇
𝖼𝗒
)
𝖯
(
𝗁
𝖾𝖺𝖽
𝖺 𝖼𝗁𝖾
)
can I blame concurrency for
my headache?
how many devs can cope
with concurrency?
how often do heads hurt?
how much
concurrency is there?
Slide 148
Slide 148 text
chip design
electrical
engineering 101
Slide 149
Slide 149 text
@hollycummins.com #Quarkus #IBM
cores per cpu
amd + intel
sources: https://www.reddit.com/r/Amd/comments/6cu5ss/highest_amount_of_cores_per_cpu_amd_vs_intel_year/
and https://www.servethehome.com/amd-and-intel-2p-server-core-count-growth-2010-2022/
Slide 150
Slide 150 text
@hollycummins.com #Quarkus #IBM
cores per cpu
amd + intel
sources: https://www.reddit.com/r/Amd/comments/6cu5ss/highest_amount_of_cores_per_cpu_amd_vs_intel_year/
and https://www.servethehome.com/amd-and-intel-2p-server-core-count-growth-2010-2022/
1
Slide 151
Slide 151 text
@hollycummins.com #Quarkus #IBM
cores per cpu
amd + intel
sources: https://www.reddit.com/r/Amd/comments/6cu5ss/highest_amount_of_cores_per_cpu_amd_vs_intel_year/
and https://www.servethehome.com/amd-and-intel-2p-server-core-count-growth-2010-2022/
1 2
Slide 152
Slide 152 text
@hollycummins.com #Quarkus #IBM
cores per cpu
amd + intel
sources: https://www.reddit.com/r/Amd/comments/6cu5ss/highest_amount_of_cores_per_cpu_amd_vs_intel_year/
and https://www.servethehome.com/amd-and-intel-2p-server-core-count-growth-2010-2022/
1 2
48
Slide 153
Slide 153 text
@hollycummins.com #Quarkus #IBM
cores per cpu
amd + intel
sources: https://www.reddit.com/r/Amd/comments/6cu5ss/highest_amount_of_cores_per_cpu_amd_vs_intel_year/
and https://www.servethehome.com/amd-and-intel-2p-server-core-count-growth-2010-2022/
1 2
48
128
Slide 154
Slide 154 text
@hollycummins.com #Quarkus #IBM
sources: https://www.reddit.com/r/Amd/comments/6cu5ss/highest_amount_of_cores_per_cpu_amd_vs_intel_year/
and https://www.servethehome.com/amd-and-intel-2p-server-core-count-growth-2010-2022/
1 2
48
128
P(concurrency)
Slide 155
Slide 155 text
@hollycummins.com #Quarkus #IBM
Slide 156
Slide 156 text
@hollycummins.com #Quarkus #IBM
why?
Slide 157
Slide 157 text
@hollycummins.com #Quarkus #IBM
why?
single-processor performance
has plateaued
Slide 158
Slide 158 text
@hollycummins.com #Quarkus #IBM
we have to get good
at concurrency
Slide 159
Slide 159 text
@hollycummins.com #Quarkus #IBM
but not all of us
have to :)
Slide 160
Slide 160 text
@hollycummins.com #Quarkus #IBM
the quarkus team:
P(headache | concurrency) < 0.10
Slide 161
Slide 161 text
@hollycummins.com #Quarkus #IBM
2022
Slide 162
Slide 162 text
@hollycummins.com #Quarkus #IBM
2022
Slide 163
Slide 163 text
@hollycummins.com #Quarkus #IBM
2022
people thought they had
to do reactive programming
Slide 164
Slide 164 text
@hollycummins.com #Quarkus #IBM
2025
Slide 165
Slide 165 text
@hollycummins.com #Quarkus #IBM
framework
Slide 166
Slide 166 text
@hollycummins.com #Quarkus #IBM
framework
fast reactive core
… if it’s quarkus :)
Slide 167
Slide 167 text
@hollycummins.com #Quarkus #IBM
framework
fast reactive core
… if it’s quarkus :)
user code
Slide 168
Slide 168 text
@hollycummins.com #Quarkus #IBM
framework
fast reactive core
… if it’s quarkus :)
user code
reactive programming model
Slide 169
Slide 169 text
@hollycummins.com #Quarkus #IBM
framework
fast reactive core
… if it’s quarkus :)
user code
reactive programming model
virtual threads
Slide 170
Slide 170 text
@hollycummins.com #Quarkus #IBM
framework
fast reactive core
… if it’s quarkus :)
user code
reactive programming model
virtual threads
structured concurrency
Slide 171
Slide 171 text
@hollycummins.com #Quarkus #IBM
quarkus programming model
@RunOnVirtualThread
Slide 172
Slide 172 text
why did green threads flop,
and loom succeed?
Slide 173
Slide 173 text
@hollycummins.com #Quarkus #IBM
sources: https://www.reddit.com/r/Amd/comments/6cu5ss/highest_amount_of_cores_per_cpu_amd_vs_intel_year/
and https://www.servethehome.com/amd-and-intel-2p-server-core-count-growth-2010-2022/
1 2
48
128
why did green threads flop,
and loom succeed?
Slide 174
Slide 174 text
why was AOT a bad idea,
but GraalVM is a good one?
Slide 175
Slide 175 text
meteorology
101
Slide 176
Slide 176 text
meteorology
101
Slide 177
Slide 177 text
cloud
meteorology
101
Slide 178
Slide 178 text
why is there all this
hype, always?
Slide 179
Slide 179 text
business
studies 101
Slide 180
Slide 180 text
venture capital
business
studies 101
Slide 181
Slide 181 text
@hollycummins.com #Quarkus #IBM
Slide 182
Slide 182 text
@hollycummins.com #Quarkus #IBM
baby company
Slide 183
Slide 183 text
@hollycummins.com #Quarkus #IBM
baby company
Slide 184
Slide 184 text
@hollycummins.com #Quarkus #IBM
baby company
Slide 185
Slide 185 text
@hollycummins.com #Quarkus #IBM
funding rounds
baby company
Slide 186
Slide 186 text
@hollycummins.com #Quarkus #IBM
funding rounds
baby company
attracted
investment
Slide 187
Slide 187 text
@hollycummins.com #Quarkus #IBM
funding rounds
baby company
attracted
investment
Slide 188
Slide 188 text
@hollycummins.com #Quarkus #IBM
funding rounds
baby company
attracted
investment
Slide 189
Slide 189 text
@hollycummins.com #Quarkus #IBM
funding rounds
baby company
attracted
investment
Slide 190
Slide 190 text
@hollycummins.com #Quarkus #IBM
funding rounds
baby company
attracted
investment
Slide 191
Slide 191 text
@hollycummins.com #Quarkus #IBM
funding rounds
baby company
attracted
investment
Slide 192
Slide 192 text
@hollycummins.com #Quarkus #IBM
funding rounds
baby company
attracted
investment
Slide 193
Slide 193 text
@hollycummins.com #Quarkus #IBM
funding rounds
baby company
attracted
investment
Slide 194
Slide 194 text
@hollycummins.com #Quarkus #IBM
“exit”
funding rounds
baby company
attracted
investment
Slide 195
Slide 195 text
@hollycummins.com #Quarkus #IBM
“exit”
funding rounds
baby company
IPO
attracted
investment
Slide 196
Slide 196 text
@hollycummins.com #Quarkus #IBM
“exit”
funding rounds
baby company
IPO
acquisition
attracted
investment
Slide 197
Slide 197 text
@hollycummins.com #Quarkus #IBM
“exit”
funding rounds
baby company
IPO
acquisition
attracted
investment
Slide 198
Slide 198 text
@hollycummins.com #Quarkus #IBM
what attracts
investment?
Slide 199
Slide 199 text
@hollycummins.com #Quarkus #IBM
what attracts
investment?
stability?
Slide 200
Slide 200 text
@hollycummins.com #Quarkus #IBM
what attracts
investment?
stability?
profitability?
Slide 201
Slide 201 text
@hollycummins.com #Quarkus #IBM
what attracts
investment?
stability?
profitability?
revenue?
Slide 202
Slide 202 text
@hollycummins.com #Quarkus #IBM
what attracts
investment?
growth
stability?
profitability?
revenue?
Slide 203
Slide 203 text
@hollycummins.com #Quarkus #IBM
what attracts
investment?
growth
stability?
profitability?
excitement
revenue?
Slide 204
Slide 204 text
@hollycummins.com #Quarkus #IBM
what attracts
investment?
growth
stability?
profitability?
excitement
revenue?
likelihood of
selling stake
at a profit
Slide 205
Slide 205 text
@hollycummins.com #Quarkus #IBM
“It’s not simply that one piece of technology is
overhyped, it’s that hype is a necessary ingredient of
the current business ecosystem of the tech industry.”
– Meredith Whittaker
Slide 206
Slide 206 text
@hollycummins.com #Quarkus #IBM
“It’s not simply that one piece of technology is
overhyped, it’s that hype is a necessary ingredient of
the current business ecosystem of the tech industry.”
– Meredith Whittaker
Slide 207
Slide 207 text
why can’t I find a job?
Slide 208
Slide 208 text
why can’t I find a new job?
Slide 209
Slide 209 text
@hollycummins.com #Quarkus #IBM
Source: indeed.com via https://newsletter.pragmaticengineer.com/p/software-engineering-job-openings?publication_id=458709
usa developer vacancies
Slide 210
Slide 210 text
@hollycummins.com #Quarkus #IBM
Source: indeed.com via https://newsletter.pragmaticengineer.com/p/software-engineering-job-openings?publication_id=458709
uk developer vacancies
Slide 211
Slide 211 text
interest rates
macroeconomics
101
Slide 212
Slide 212 text
@hollycummins.com #Quarkus #IBM
ZIRP
Slide 213
Slide 213 text
@hollycummins.com #Quarkus #IBM
ZIRP
“zero interest-rate phenomenon”
Slide 214
Slide 214 text
@hollycummins.com #Quarkus #IBM
0%
Slide 215
Slide 215 text
@hollycummins.com #Quarkus #IBM
<1%
Slide 216
Slide 216 text
@hollycummins.com #Quarkus #IBM
Source: Bank of England
UK interest rates
Slide 217
Slide 217 text
@hollycummins.com #Quarkus #IBM
Source: Bank of England
averting
subprime
collapse
UK interest rates
Slide 218
Slide 218 text
@hollycummins.com #Quarkus #IBM
Source: Bank of England
averting
subprime
collapse
covid
UK interest rates
Slide 219
Slide 219 text
@hollycummins.com #Quarkus #IBM
Source: Bank of England
averting
subprime
collapse
covid
0.1% (!)
UK interest rates
Slide 220
Slide 220 text
@hollycummins.com #Quarkus #IBM
Source: Bank of England
averting
subprime
collapse
covid
0.1% (!)
inflation
shortages
war
UK interest rates
Slide 221
Slide 221 text
@hollycummins.com #Quarkus #IBM
Source: Bank of England
UK interest rates
Slide 222
Slide 222 text
@hollycummins.com #Quarkus #IBM
exit
attracted investment
baby company
Slide 223
Slide 223 text
@hollycummins.com #Quarkus #IBM
exit
attracted investment
baby company
Slide 224
Slide 224 text
@hollycummins.com #Quarkus #IBM
exit
attracted investment
baby company
Slide 225
Slide 225 text
@hollycummins.com #Quarkus #IBM
exit
attracted investment
baby company
risk
Slide 226
Slide 226 text
@hollycummins.com #Quarkus #IBM
ZIRP
Slide 227
Slide 227 text
@hollycummins.com #Quarkus #IBM
software company
ZIRP
Slide 228
Slide 228 text
@hollycummins.com #Quarkus #IBM
government bond
software company
ZIRP
Slide 229
Slide 229 text
@hollycummins.com #Quarkus #IBM
government bond
risk
return
software company
ZIRP
Slide 230
Slide 230 text
@hollycummins.com #Quarkus #IBM
government bond
risk
return
software company
high
ZIRP
Slide 231
Slide 231 text
@hollycummins.com #Quarkus #IBM
government bond
risk
return
software company
high
20% - 500%
ZIRP
Slide 232
Slide 232 text
@hollycummins.com #Quarkus #IBM
government bond
low
risk
return
software company
high
20% - 500%
ZIRP
Slide 233
Slide 233 text
@hollycummins.com #Quarkus #IBM
government bond
low
0%
risk
return
software company
high
20% - 500%
ZIRP
Slide 234
Slide 234 text
@hollycummins.com #Quarkus #IBM
government bond
low
0%
risk
return
software company
high
20% - 500%
ZIRP
Slide 235
Slide 235 text
@hollycummins.com #Quarkus #IBM
government bond
low
4%
risk
return
software company
high
20% - 500%
post-ZIRP
Slide 236
Slide 236 text
@hollycummins.com #Quarkus #IBM
government bond
low
4%
risk
return
software company
high
20% - 500%
post-ZIRP
Slide 237
Slide 237 text
@hollycummins.com #Quarkus #IBM
less investment means less hiring
Slide 238
Slide 238 text
@hollycummins.com #Quarkus #IBM
less investment means less hiring
Slide 239
Slide 239 text
what about the ai-lephant in the room?
Slide 240
Slide 240 text
is AI killing jobs?
Slide 241
Slide 241 text
AI is attracting heavy investment
exit
hype generation
Slide 242
Slide 242 text
No content
Slide 243
Slide 243 text
@hollycummins.com #Quarkus #IBM
“our AI strategy is working”
Slide 244
Slide 244 text
just because AI makes developers more productive,
doesn’t mean it will reduce the number of developers
Slide 245
Slide 245 text
jevon’s paradox
economics
101
Slide 246
Slide 246 text
@holly_cummins #Quarkus #IBM
jevon’s paradox
Slide 247
Slide 247 text
@holly_cummins #Quarkus #IBM
jevon’s paradox
“efficiency improvements can
lead to increased consumption”
Slide 248
Slide 248 text
@holly_cummins #Quarkus #IBM
what we imagine
when we widen roads
Slide 249
Slide 249 text
@holly_cummins #Quarkus #IBM
what we get
Slide 250
Slide 250 text
@holly_cummins #Quarkus #IBM
Slide 251
Slide 251 text
@holly_cummins #Quarkus #IBM
Slide 252
Slide 252 text
@holly_cummins #Quarkus #IBM
professions using ai worked
3.15
hours
more per week
https://cepr.org/voxeu/columns/ais-power-grows-so-does-our-workday
Slide 253
Slide 253 text
@hollycummins.com #Quarkus #IBM
software is not going away
Slide 254
Slide 254 text
@hollycummins.com #Quarkus #IBM
1800s
Slide 255
Slide 255 text
@hollycummins.com #Quarkus #IBM
“knocker-up”
1800s
Slide 256
Slide 256 text
@hollycummins.com #Quarkus #IBM
“knocker-up”
1800s
human alarm clock
Slide 257
Slide 257 text
@hollycummins.com #Quarkus #IBM
“knocker-up”
1800s - 1940s
human alarm clock
Slide 258
Slide 258 text
@hollycummins.com #Quarkus #IBM
“knocker-up”
1800s
finite demand
- 1940s
human alarm clock
Slide 259
Slide 259 text
@hollycummins.com #Quarkus #IBM
“knocker-up”
1800s
finite demand
only employees need waking up
- 1940s
human alarm clock
Slide 260
Slide 260 text
@hollycummins.com #Quarkus #IBM
“knocker-up”
1800s
finite demand
only employees need waking up
no one needs waking up more than once
- 1940s
human alarm clock
Slide 261
Slide 261 text
@hollycummins.com #Quarkus #IBM
the more software we have,
the more we want
Slide 262
Slide 262 text
@hollycummins.com #Quarkus #IBM
Slide 263
Slide 263 text
@hollycummins.com #Quarkus #IBM
software to do stuff
Slide 264
Slide 264 text
@hollycummins.com #Quarkus #IBM
software to do stuff
software to do more stuff
Slide 265
Slide 265 text
@hollycummins.com #Quarkus #IBM
software to do stuff
software to do more stuff
software to tell people about software
Slide 266
Slide 266 text
@hollycummins.com #Quarkus #IBM
software to do stuff
software to do more stuff
software to tell people about software
software to bypass advertisements in the software
Slide 267
Slide 267 text
@hollycummins.com #Quarkus #IBM
software to do stuff
software to do more stuff
software to tell people about software
software to bypass advertisements in the software
software to bypass the bypass software so ads still show
Slide 268
Slide 268 text
@hollycummins.com #Quarkus #IBM
software to do stuff
software to do more stuff
software to tell people about software
software to bypass advertisements in the software
software to bypass the bypass software so ads still show
software to gamify software
Slide 269
Slide 269 text
@hollycummins.com #Quarkus #IBM
software to do stuff
software to do more stuff
software to tell people about software
software to bypass advertisements in the software
software to bypass the bypass software so ads still show
software to gamify software
software to monitor the software
Slide 270
Slide 270 text
@hollycummins.com #Quarkus #IBM
software to do stuff
software to do more stuff
software to tell people about software
software to bypass advertisements in the software
software to bypass the bypass software so ads still show
software to gamify software
software to monitor the software
software to debug the vibe-coded software
Slide 271
Slide 271 text
@hollycummins.com #Quarkus #IBM
software to do stuff
software to do more stuff
software to tell people about software
software to bypass advertisements in the software
software to bypass the bypass software so ads still show
software to gamify software
software to monitor the software
software to debug the vibe-coded software
software to manage the software
Slide 272
Slide 272 text
@hollycummins.com #Quarkus #IBM
software to do stuff
software to do more stuff
software to tell people about software
software to bypass advertisements in the software
software to bypass the bypass software so ads still show
software to gamify software
software to monitor the software
software to debug the vibe-coded software
software to manage the software
software to write the software
Slide 273
Slide 273 text
@hollycummins.com #Quarkus #IBM
software to do stuff
software to do more stuff
software to tell people about software
software to bypass advertisements in the software
software to bypass the bypass software so ads still show
software to gamify software
software to monitor the software
software to debug the vibe-coded software
software to manage the software
software to write the software
software to manage the software that writes the software
Slide 274
Slide 274 text
@hollycummins.com #Quarkus #IBM
software to do stuff
software to do more stuff
software to tell people about software
software to bypass advertisements in the software
software to bypass the bypass software so ads still show
software to gamify software
software to monitor the software
software to debug the vibe-coded software
software to manage the software
software to write the software
software to manage the software that writes the software
software to try and avoid skynet
Slide 275
Slide 275 text
@hollycummins.com #Quarkus #IBM
software to do stuff
software to do more stuff
software to tell people about software
software to bypass advertisements in the software
software to bypass the bypass software so ads still show
software to gamify software
software to monitor the software
software to debug the vibe-coded software
software to manage the software
software to write the software
software to manage the software that writes the software
software to try and avoid skynet
software to work out the meaning of life
Slide 276
Slide 276 text
@hollycummins.com #Quarkus #IBM
software to do stuff
software to do more stuff
software to tell people about software
software to bypass advertisements in the software
software to bypass the bypass software so ads still show
software to gamify software
software to monitor the software
software to debug the vibe-coded software
software to manage the software
software to write the software
software to manage the software that writes the software
software to try and avoid skynet
software to work out the meaning of life
…
Slide 277
Slide 277 text
@hollycummins.com #Quarkus #IBM
Slide 278
Slide 278 text
@hollycummins.com #Quarkus #IBM
Source: indeed.com via https://newsletter.pragmaticengineer.com/p/software-engineering-job-openings?publication_id=458709
usa developer vacancies
Slide 279
Slide 279 text
@hollycummins.com #Quarkus #IBM
Source: indeed.com via https://newsletter.pragmaticengineer.com/p/software-engineering-job-openings?publication_id=458709
usa developer vacancies
chatgpt
released
Slide 280
Slide 280 text
@holly_cummins #Quarkus #IBM
tl;dpa
(too long; didn’t pay attention)
Slide 281
Slide 281 text
@holly_cummins #Quarkus #IBM
tl;dpa
⁃ statistics is (annoyingly) important, especially bayes
(too long; didn’t pay attention)
Slide 282
Slide 282 text
@holly_cummins #Quarkus #IBM
tl;dpa
⁃ statistics is (annoyingly) important, especially bayes
⁃ to achieve more power, computers have to add more cores
(too long; didn’t pay attention)
Slide 283
Slide 283 text
@holly_cummins #Quarkus #IBM
tl;dpa
⁃ statistics is (annoyingly) important, especially bayes
⁃ to achieve more power, computers have to add more cores
⁃ software is not very good at taking advantage of these cores
(too long; didn’t pay attention)
Slide 284
Slide 284 text
@holly_cummins #Quarkus #IBM
tl;dpa
⁃ statistics is (annoyingly) important, especially bayes
⁃ to achieve more power, computers have to add more cores
⁃ software is not very good at taking advantage of these cores
⁃ the rise of interest rates means a tougher job market
(too long; didn’t pay attention)
Slide 285
Slide 285 text
@holly_cummins #Quarkus #IBM
tl;dpa
⁃ statistics is (annoyingly) important, especially bayes
⁃ to achieve more power, computers have to add more cores
⁃ software is not very good at taking advantage of these cores
⁃ the rise of interest rates means a tougher job market
⁃ jevon’s paradox means demand for software will go up even
if ai makes producing software cheaper
(too long; didn’t pay attention)
Slide 286
Slide 286 text
http://hollycummins.com/things-to-care-about-devoxx-be/