This document discusses issues related to open source software (OSS) in the era of cloud computing and machine learning. It addresses topics like sustainability of OSS projects, maintainer burnout, and commercial exploitation of OSS. It argues that many of these issues are really "business model" problems rather than technical problems. The document also discusses how OSS communities value empowering people to innovate through open collaboration and aligning various stakeholders. It emphasizes that open APIs and non-proprietary standards are important to preserve user choice and control as software becomes more distributed through APIs and services.
7. • Not about
licenses
• Empowering
people &
communities
to innovate
• Aligns users,
customers,
innovators
8. Code —> APIs
• Software is becoming more about APIs.
• Fewer and fewer projects are huge monoliths.
• There is gnashing of teeth, but what younger devs
don't understand is that this is the "gig-ification" of
the software industry.
• Leads to darker pattern: industry using faux-OSS
to establish de facto standards that benefit
proprietary platforms (“gaming the bazaar”)
9. These are not “open source” problems.
These are “business model” problems.
10. Stable, open APIs are a commons; a
utility; a market between:
• Maintainers
• Users
• Innovators
They preserve choice for the user.
11. Non-binary Openness: How To Measure It?
• Reproducibility of running deployment (ie cost
to reproduce)?
• Cost to change a core schema or app logic and
redeploy and migrate users over (cost to fork)?
• Downside risk, e.g. "cost or impact to users if
the service terminated”?
13. OSS, Two Ways
“Classical OSS”:
• By devs, for devs
• “Scratch my own itch”
Scipy / PyData:
• “Scratch my own
itch”
• …b/c CS majors
don’t know
differential equations
14. Commoditization vs Innovation
• Classic, “infrastructure” OSS tends to solve “well
known problems”
• Data Science, Numerical, Scientific OSS tends to
solve advanced problems that appear to be niche
(until they’re not)
In both cases: OSS facilitates an open marketplace
• The cost of maintenance can be honestly traded off
against the risk/upside of future innovation.
• Free for network intelligence to explore a much bigger
solution space, in parallel.
15. The PyData & SciPy Innovation Community
PyData & Scipy have successfully pursued the art of the
possible for 20 years
Successful projects tend to:
• Limit scope
• Deliver concrete utility and innovation
• Play well with others
• Have reasonable leadership
This collaboration community has out-innovated
& better leveraged human labor than capitalism.
16. What Do Businesses Value in OSS?
• Not about cost of software (CapEx)
• Not even about maintenance of software
(OpEx)
• Core business goals:
• Avoid lock-in
• Harness innovation
17. • Gift economy - sharing is caring!
• Culture of Participation
• Pride in the craft
• Oftentimes, the software artifact is almost
ancillary — it’s a by-product
• Its actual utility for end-users
• The predictability, stability, efficiency of the
development process
What Does the OSS Community Value?
18. As OSS communities mature, how we play
together becomes as important as the music itself
19. • Software is a relationship: your code, solving someone
else’s problem
• Dynamic, emergent semi-stable pattern emerging from
underlying complexity
• Grows by consuming ideas that better adapt it for
some function within an external context
• can look at health based on factors like internal
entropy, the external environment in which it lives, etc.
•
What They Both Miss: Software Isn’t Code
20. • Open Source Software is a
special type of un-property
• Confuses most economic
engines
• Intrinsically anti-rivalrous
• Sharing increases value
• Forking decreases value
Open Source Software Isn’t Property
OSS flourishes when approached with an
generative, abundance mentality
23. Clay Shirky: Here Comes Everybody
1. “Me-first collaboration”
2. Conversation – people coming together to learn
3. Collaboration – a group forms with a shared purpose
“It requires a division of labor, and teamwork. It can
often be characterized by people wanting to fix a
market failure, and is motivated by increasing
accessibility.”
4. Collective action – “the fate of the group as a whole
becomes important”
24. Participation vs Labor
• Participatory cultural activity vs. free stuff. Fruit vs.
tree.
• Participation gives you a voice and preserves your
agency.
• To accept other people into the participation
culture, requires developing trust
• It’s hard or impossible to trust a corporation or
brand. Trust inherently is still tied to individuals.
27. • What are the values that underpin our open
source communities?
• “GNU philosophy for ML era"?
• What is "access"?
• What is libre, what is compelled, what is de
facto in our modern era?
Goals and Values
29. An Innovation Commons
• Most commons are about
defending a scarce resource from
rivalrous competition and depletion
• Innovation & Creation are
generative, non-rivalrous. So what
commons needs defending?
The Attention and Time of
the developer community
(locally and globally)
31. Recap
Open
Is a value regarding contribution, utilization, and
user freedom
Source
Is an outdated term, from when just “reading
code” guaranteed user sovereignty and freedom
Software
Is a valuable artifact for goal-directed businesses;
But developing it well is a primary value for OSS
communities
Community
The social environment in which we can live our
values.
34. How can automata harm human
society?
“Automatons must not be taken
for granted… the machines may
be able to escape human control
if humans do not continue proper
supervision of them. We might
become entirely dependent on
them, or even controlled by them.
There is danger in trusting
decisions to something which
cannot think abstractly, and may
therefore be unlikely to identify
with intellectual human values
which are not purely utilitarian.”
35. Jacques Ellul, The Technological Society
The rationality of technique
enforces logical and
mechanical organization
through division of labor, the
setting of production
standards, etc. And it creates
an artificial system which
"eliminates or subordinates
the natural world.”
36. “…those applied sciences that are amenable
to the aims of collective economic production
(be it capitalist, socialist, or communist),
have been elevated to the position of sacred
in Western culture.”
37.
38. “Modern technology has become a total
phenomenon for civilization, the defining force
of a new social order in which efficiency is no
longer an option but a necessity imposed on
all human activity.”
39. Towering stacks of
complex software & algos
create a technological
caste system.
ML & automated
inference will grow to
consume all surplus.
They will also reinforce
existing structures of
power & regulatory
capture.
40.
41. Python Is Different. Python Is for Humans.
• Python is not technology for technology's sake.
• It's about empowering regular people.
• In this history of tech, actual democratization &
user empowerment is rare
• Language is a human instinct and is a natural path
to insight
• Computer languages are thoughtware, not software
42. Open Data Science == Sovereignty, Freedom
• Open for innovation
• Defends fork-right
• Empowers people to not become
informational vassals
I’m not trying to be the new RMS…
…but I will be if I have to be!