You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: content/posts/open_source_ai.md
+6-46Lines changed: 6 additions & 46 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -16,58 +16,18 @@ series: ["Open Cybernetics"]
16
16
17
17
**Author: Lum Ramabaja**
18
18
19
-
This essay was originally published on [Open Cybernetics](https://opncbr.com) & [ChainlessCoder](https://chainlesscoder.com). It can also be read as a PDF on [arXiv]().
19
+
This essay was originally published as a blog post on \href{https://blog.opncbr.com/post/open_source_ai/}{Open Cybernetics} \& \href{https://chainlesscoder.com/posts/open-source-ai-and-the-challenges-ahead/}{ChainlessCoder}. It can also be read as a PDF on [arXiv]().
20
20
21
21
---
22
22
<!--more-->
23
23
24
24
# The price of AI
25
25
26
-
While I remain confident that open source AI will continue to bloom, I
27
-
believe that the incentives for corporations and governments to limit
28
-
access to novel neural network architectures and techniques will only
29
-
intensify in the coming years. In this essay, I aim to predict and
30
-
outline the potential disruptions that may occur over the course of this
31
-
decade, and to offer some ideas and solutions to help the open source
32
-
community adapt to this hypothetical future. Predicting the future is of
33
-
course difficult, and I realize that many of the predictions I make may
34
-
not come to pass or may unfold entirely differently. However, this does
35
-
not render the exercise of thinking about the future futile. Rather,
36
-
contemplating possible futures can help us adapt proactively to an
37
-
ever-changing world. As our world continues to evolve at a rapid pace,
38
-
the ability to envision and prepare for possible new worlds will become
39
-
ever more important.
40
-
41
-
In the near future, the way we interact with people, the way we exchange
42
-
and perform labor, even the way societies are organized will change in
43
-
fundamental ways. It appears highly probable that each person will have
44
-
personalized AI entities, which I refer to as "Ghosts", that will link
45
-
them to a global network of other AI systems. These AI systems will
46
-
perform many services for us. Rather than perceiving them as mere
47
-
assistants, it may be more fitting to regard them as cognitive
48
-
extensions of ourselves. Firms and organizations will most likely also
49
-
have their own Ghosts, enabling more effective collaboration among
50
-
members. Beyond the social aspects, associative memory networks with
51
-
recurrent connections might enable AI systems to form a memory. These
52
-
stand-alone Ghosts might even develop an identity. Furthermore, AI
53
-
systems leveraging consensus algorithms may be developed, potentially
54
-
giving rise to decentralized autonomous AIs. Although this future has
55
-
yet to materialize, we can already envision some of the economic
56
-
dynamics that will unfold.
57
-
58
-
The exchange-value of AI-generated services will be equal to the energy
59
-
required to provide the service (i.e. the cost of running the relevant
60
-
model(s)), plus the AI's information asymmetry within the market.
61
-
Services that are relatively simple to replicate using AIs will
62
-
inevitably have low exchange-values, leading to diminished surplus-value
63
-
for the AI owners. This trend will particularly impact service-based
64
-
economies, which can expect a significant reduction in surplus-value for
65
-
most services. Consequently, we can expect that for a period, the
66
-
populations and governments of many Western nations will react similarly
67
-
to the Luddites of the 19th century. Driven by the diminished returns
68
-
experienced across sectors, many regions of the world will resort to
69
-
monopolistic licensing strategies and autocratic rules to impede AI
70
-
access.
26
+
While I remain confident that open source AI will continue to bloom, I believe that the incentives for corporations and governments to limit access to novel neural network architectures and techniques will only intensify in the coming years. In this essay, I aim to predict and outline the potential disruptions that may occur over the course of this decade, and to offer some ideas and solutions to help the open source community adapt to this hypothetical future. Predicting the future is of course difficult, and I realize that many of the predictions I make may not come to pass or may unfold entirely differently. However, this does not render the exercise of thinking about the future futile. Rather, contemplating possible futures can help us adapt proactively to an ever-changing world. As our world continues to evolve at a rapid pace, the ability to envision and prepare for possible new worlds will become ever more important.
27
+
28
+
In the near future, the way we interact with people, the way we exchange and perform labor, even the way societies are organized will change in fundamental ways. It appears highly probable that each person will have personalized AI entities, which I refer to as "Ghosts", that will link them to a global network of other AI systems. These AI systems will perform many services for us. Rather than perceiving them as mere assistants, it may be more fitting to regard them as cognitive extensions of ourselves. Firms and organizations will most likely also have their own Ghosts, enabling more effective collaboration among members. Beyond the social aspects, associative memory networks with recurrent connections might enable AI systems to form a memory. These stand-alone Ghosts might even develop an identity. Furthermore, AI systems leveraging consensus algorithms may be developed, potentially giving rise to decentralized autonomous AIs. Although this future has yet to materialize, we can already envision some of the economic dynamics that will unfold.
29
+
30
+
The exchange-value of AI-generated services will be equal to the energy required to provide the service (i.e. the cost of running the relevant model(s)), plus the AI's information asymmetry within the market. Services that are relatively simple to replicate using AIs will inevitably have low exchange-values, leading to diminished surplus-value for the AI owners. This trend will particularly impact service-based economies, which can expect a significant reduction in surplus-value for most services. Consequently, we can expect that for a period, the populations and governments of many Western nations will react similarly to the Luddites of the 19th century. Driven by the diminished returns experienced across sectors, many regions of the world will resort to monopolistic licensing strategies and autocratic rules to impede AI access.
71
31
72
32
In more AI-progressive regions of the world, open source AI will most
73
33
likely flourish for services with low exchange-value, but high use-value
This essay was originally published on Open Cybernetics & ChainlessCoder. It can also be read as a PDF on arXiv.
89
+
This essay was originally published as a blog post on \href{https://blog.opncbr.com/post/open_source_ai/}{Open Cybernetics} & \href{https://chainlesscoder.com/posts/open-source-ai-and-the-challenges-ahead/}{ChainlessCoder}. It can also be read as a PDF on arXiv.
<p>This essay was originally published on <a href="https://opncbr.com" target="_blank" rel="noopener">Open Cybernetics</a>&amp; <a href="https://chainlesscoder.com" target="_blank" rel="noopener">ChainlessCoder</a>. It can also be read as a PDF on <a href="">arXiv</a>.</p>
20
+
<p>This essay was originally published as a blog post on \href{https://blog.opncbr.com/post/open_source_ai/}{Open Cybernetics}&amp; \href{https://chainlesscoder.com/posts/open-source-ai-and-the-challenges-ahead/}{ChainlessCoder}. It can also be read as a PDF on <a href="">arXiv</a>.</p>
<p>This essay was originally published on <a href="https://opncbr.com" target="_blank" rel="noopener">Open Cybernetics</a>&amp; <a href="https://chainlesscoder.com" target="_blank" rel="noopener">ChainlessCoder</a>. It can also be read as a PDF on <a href="">arXiv</a>.</p>
20
+
<p>This essay was originally published as a blog post on \href{https://blog.opncbr.com/post/open_source_ai/}{Open Cybernetics}&amp; \href{https://chainlesscoder.com/posts/open-source-ai-and-the-challenges-ahead/}{ChainlessCoder}. It can also be read as a PDF on <a href="">arXiv</a>.</p>
<p>This essay was originally published on <ahref="https://opncbr.com" target="_blank" rel="noopener">Open Cybernetics</a> & <ahref="https://chainlesscoder.com" target="_blank" rel="noopener">ChainlessCoder</a>. It can also be read as a PDF on <ahref="">arXiv</a>.</p>
91
+
<p>This essay was originally published as a blog post on \href{https://blog.opncbr.com/post/open_source_ai/}{Open Cybernetics} & \href{https://chainlesscoder.com/posts/open-source-ai-and-the-challenges-ahead/}{ChainlessCoder}. It can also be read as a PDF on <ahref="">arXiv</a>.</p>
92
92
<hr>
93
-
<h1id="the-price-of-ai">The price of AI <ahref="#the-price-of-ai" class="anchor">🔗</a></h1><p>While I remain confident that open source AI will continue to bloom, I
94
-
believe that the incentives for corporations and governments to limit
95
-
access to novel neural network architectures and techniques will only
96
-
intensify in the coming years. In this essay, I aim to predict and
97
-
outline the potential disruptions that may occur over the course of this
98
-
decade, and to offer some ideas and solutions to help the open source
99
-
community adapt to this hypothetical future. Predicting the future is of
100
-
course difficult, and I realize that many of the predictions I make may
101
-
not come to pass or may unfold entirely differently. However, this does
102
-
not render the exercise of thinking about the future futile. Rather,
103
-
contemplating possible futures can help us adapt proactively to an
104
-
ever-changing world. As our world continues to evolve at a rapid pace,
105
-
the ability to envision and prepare for possible new worlds will become
106
-
ever more important.</p>
107
-
<p>In the near future, the way we interact with people, the way we exchange
108
-
and perform labor, even the way societies are organized will change in
109
-
fundamental ways. It appears highly probable that each person will have
110
-
personalized AI entities, which I refer to as “Ghosts”, that will link
111
-
them to a global network of other AI systems. These AI systems will
112
-
perform many services for us. Rather than perceiving them as mere
113
-
assistants, it may be more fitting to regard them as cognitive
114
-
extensions of ourselves. Firms and organizations will most likely also
115
-
have their own Ghosts, enabling more effective collaboration among
116
-
members. Beyond the social aspects, associative memory networks with
117
-
recurrent connections might enable AI systems to form a memory. These
118
-
stand-alone Ghosts might even develop an identity. Furthermore, AI
119
-
systems leveraging consensus algorithms may be developed, potentially
120
-
giving rise to decentralized autonomous AIs. Although this future has
121
-
yet to materialize, we can already envision some of the economic
122
-
dynamics that will unfold.</p>
123
-
<p>The exchange-value of AI-generated services will be equal to the energy
124
-
required to provide the service (i.e. the cost of running the relevant
125
-
model(s)), plus the AI’s information asymmetry within the market.
126
-
Services that are relatively simple to replicate using AIs will
127
-
inevitably have low exchange-values, leading to diminished surplus-value
128
-
for the AI owners. This trend will particularly impact service-based
129
-
economies, which can expect a significant reduction in surplus-value for
130
-
most services. Consequently, we can expect that for a period, the
131
-
populations and governments of many Western nations will react similarly
132
-
to the Luddites of the 19th century. Driven by the diminished returns
133
-
experienced across sectors, many regions of the world will resort to
134
-
monopolistic licensing strategies and autocratic rules to impede AI
135
-
access.</p>
93
+
<h1id="the-price-of-ai">The price of AI <ahref="#the-price-of-ai" class="anchor">🔗</a></h1><p>While I remain confident that open source AI will continue to bloom, I believe that the incentives for corporations and governments to limit access to novel neural network architectures and techniques will only intensify in the coming years. In this essay, I aim to predict and outline the potential disruptions that may occur over the course of this decade, and to offer some ideas and solutions to help the open source community adapt to this hypothetical future. Predicting the future is of course difficult, and I realize that many of the predictions I make may not come to pass or may unfold entirely differently. However, this does not render the exercise of thinking about the future futile. Rather, contemplating possible futures can help us adapt proactively to an ever-changing world. As our world continues to evolve at a rapid pace, the ability to envision and prepare for possible new worlds will become ever more important.</p>
94
+
<p>In the near future, the way we interact with people, the way we exchange and perform labor, even the way societies are organized will change in fundamental ways. It appears highly probable that each person will have personalized AI entities, which I refer to as “Ghosts”, that will link them to a global network of other AI systems. These AI systems will perform many services for us. Rather than perceiving them as mere assistants, it may be more fitting to regard them as cognitive extensions of ourselves. Firms and organizations will most likely also have their own Ghosts, enabling more effective collaboration among members. Beyond the social aspects, associative memory networks with recurrent connections might enable AI systems to form a memory. These stand-alone Ghosts might even develop an identity. Furthermore, AI systems leveraging consensus algorithms may be developed, potentially giving rise to decentralized autonomous AIs. Although this future has yet to materialize, we can already envision some of the economic dynamics that will unfold.</p>
95
+
<p>The exchange-value of AI-generated services will be equal to the energy required to provide the service (i.e. the cost of running the relevant model(s)), plus the AI’s information asymmetry within the market. Services that are relatively simple to replicate using AIs will inevitably have low exchange-values, leading to diminished surplus-value for the AI owners. This trend will particularly impact service-based economies, which can expect a significant reduction in surplus-value for most services. Consequently, we can expect that for a period, the populations and governments of many Western nations will react similarly to the Luddites of the 19th century. Driven by the diminished returns experienced across sectors, many regions of the world will resort to monopolistic licensing strategies and autocratic rules to impede AI access.</p>
136
96
<p>In more AI-progressive regions of the world, open source AI will most
137
97
likely flourish for services with low exchange-value, but high use-value
138
98
(i.e. services that are relatively simple to replicate using AI).
<p>This essay was originally published on <a href="https://opncbr.com" target="_blank" rel="noopener">Open Cybernetics</a>&amp; <a href="https://chainlesscoder.com" target="_blank" rel="noopener">ChainlessCoder</a>. It can also be read as a PDF on <a href="">arXiv</a>.</p>
20
+
<p>This essay was originally published as a blog post on \href{https://blog.opncbr.com/post/open_source_ai/}{Open Cybernetics}&amp; \href{https://chainlesscoder.com/posts/open-source-ai-and-the-challenges-ahead/}{ChainlessCoder}. It can also be read as a PDF on <a href="">arXiv</a>.</p>
<p>This essay was originally published on <a href="https://opncbr.com" target="_blank" rel="noopener">Open Cybernetics</a>&amp; <a href="https://chainlesscoder.com" target="_blank" rel="noopener">ChainlessCoder</a>. It can also be read as a PDF on <a href="">arXiv</a>.</p>
20
+
<p>This essay was originally published as a blog post on \href{https://blog.opncbr.com/post/open_source_ai/}{Open Cybernetics}&amp; \href{https://chainlesscoder.com/posts/open-source-ai-and-the-challenges-ahead/}{ChainlessCoder}. It can also be read as a PDF on <a href="">arXiv</a>.</p>
<p>This essay was originally published on <a href="https://opncbr.com" target="_blank" rel="noopener">Open Cybernetics</a>&amp; <a href="https://chainlesscoder.com" target="_blank" rel="noopener">ChainlessCoder</a>. It can also be read as a PDF on <a href="">arXiv</a>.</p>
20
+
<p>This essay was originally published as a blog post on \href{https://blog.opncbr.com/post/open_source_ai/}{Open Cybernetics}&amp; \href{https://chainlesscoder.com/posts/open-source-ai-and-the-challenges-ahead/}{ChainlessCoder}. It can also be read as a PDF on <a href="">arXiv</a>.</p>
0 commit comments