adding blog post for dh2025

Author: gofilipa

public/cat-bag.jpg

@@ -0,0 +1,106 @@
+---
+title: Cat's Out of the Bag
+description: 'Reflections from ADHO 2025'
+pubDate: 'July 22, 2025'
+heroImage: '/codes/cat-bag.jpg'
+---
+
+The cat's out of the bag. But that doesn't mean we cannot trap it in
+the bathroom.
+
+At the ADHO (Alliance of Digital Humanities Organizations) 2025
+conference last week, I gave two presentations about AI and how it's
+affecting the ways [we teach
+programming](https://github.com/gofilipa/codes/tree/master/writings/talks/2025/dh2025/dh2025_teaching_python.pdf)
+and [share open data and open source
+projects](https://github.com/gofilipa/codes/tree/master/writings/talks/2025/dh2025/dh2025_open_data.pdf).
+In both of my presentations, I tried to give the sense that, although
+the "cat's out of the bag" with Machine Learning tools, and there's no
+going back to a world without it, we need to be really careful and
+thoughtful about how we use it. Because doing so make us complicit in
+the massive levels of environment, labor, and creative exploitation
+which is being carried out by big tech.
+
+While giving my remarks, especially at the panel about teaching with
+coding tools, I really felt like I was moving against the grain. There
+is so much widespread acceptance of ML in academia and industry. And I
+am not blameless. I also use it. I use it to write assignment
+descriptions for my teaching and to draft abstracts and to distill
+ideas from my own research. I use it to debug my code and sometimes,
+though less often, to generate code. I try to minimize my use, in the
+same way I try to minimize my meat intake, hoping conscientiousness
+will bring me toward something like ethical and sustainable use.
+
+My position is really the minority, though. Most people, even the ones
+who know about the massive and irreversable costs of developing ML
+technology on a large scale, don't think about it when they use it.
+For so many people, it doesn't affect their lives. They are what I
+call the technology's "final user," disconnected from the chain of
+production that brought it to them.
+
+Beneath the screen where the final user interacts with these tools,
+there's a massive stack of computational processes---hardware
+sourcing, labor exploitation, and transportation of materials---that
+pass underneath her perception. They are disconnected from her. And
+this is a kind of [screen
+essentialism](https://nickm.com/writing/essays/continuous_paper_mla.html),
+but it's also different. Because screen essentialism is about
+analyzing computational processes as material processes, bound to the
+physical and logical constraints of hardware and software. But this is
+about something else, about something which makes us forget, perhaps,
+the material constraints of the physical world in first place, what we
+already have.
+
+In [my talk about teaching
+Python](https://github.com/gofilipa/codes/tree/master/writings/talks/2025/dh2025/dh2025_teaching_python.pdf),
+I brought in a quote from Sam Altman (which comes from an interview
+with *TIME Magazine*), where he elaborates on the immense potential of
+AI to improve things like of productivity and expertise.
+
+> "If you think about how different the world can be, not only when
+> every person has... ChatGPT... but next they have the world's best
+> chief of staff. And then after that, every person has a company of
+> 20 or 50 experts that can work super well together. And then after
+> that, everybody has a company of 10,000 experts in every field that
+> can work super well together. And if someone wants to go focus on
+> curing disease, they can do that. And if someone wants to focus on
+> making great art, they can do that."
+> 
+> \- Sam Altman
+
+When I look at this, the thing that sticks out is this almost paranoid
+desire for more. More expertise, more workers, more art, more science.
+And the number of experts keeps increasing, from 20 to 50 to 10,000.
+It's as if there is no limit to what one person needs.
+
+On the surface, this is a claim about abundance. That "intelligence"
+enables abundance, will allow us to reach a future of abundance.
+However, in projecting this future of abundance, there's actually a
+more subtle claim, which is implied, going relatively unnoticed, about
+scarcity.
+
+People like Altman and other CEOs want us to believe that we need
+more. But more of what? To make things more perfectly or quickly?
+Would it really make our lives better? Or to put it differently, whose
+lives will be better? Will the benefits of AI trickle to people
+starving and dying in Gaza, in Sudan, to human suffering? Will it
+finally right the scales of inequality? It's absurd to think that,
+given *we already have* the money, knowledge, and tools to solve
+inequality, like world hunger for example, that this new tool is what
+will finally get us there.
+
+The truth is, we don't need "intelligence" to create things in order
+to have a world of abundance. The world is already abundant. What we
+actually need, and what I tried to get at, in both of my talks, is to
+slow down and think about how to make the already existing abundance
+accessible to the most amount of people possible.
+
+For my teaching and research, that means using ML tools in a critical
+way, to expose the social biases (like anti-trans bias) in popular
+discourse, or for one of my students, who built a ["men's mag" text
+generator](https://github.com/jfung53/mensmagbot) trained on men's
+magazines, with the goal of studying language aimed at male audiences.
+But for others, it might be something different, I don't know what.
+But as long as we build carefully, critically, with some reflection, I
+think that's much better than using ML just for the sake of creating
+another project.