Links: Software Engineering in the face of AI

[sonia]

Your phone edits all your photos with AI - is it changing your view of reality? by Thomas Germain. "From simple enhancements to hallucinated facial features, modern phones choose how our memories will look."

No. You can't tell it was written by AI by Segun Famisa.

In this essay, I will argue that, your favourite “tells” that a document was produced by AI, at best, is wrong, and depending on your position, in life, at worst, is dangerous and harmful.[...]
So who trained [AI]? A lot of the early training, data annotations and other manual processes, happened with cheap labour in African countries. There are multiple sources that have revealed the hidden economy of workers that big-tech outsources these kinds of tasks to African countries with unstable political situations, weaker workers rights, and cheap labour.

Curious about how LLM's actually work? So What's The Next Word Then? by Matthias Kainer does a good job of explaining it, with diagrams. Via Martin Fowler's blog.

Acting ethically in an imperfect world by Jürgen Geuter describes and addresses Cory Doctorow's defensiveness about using LLMs.

I appreciate a lot of work Cory Doctorow has done in the last decades. But the arguments he presents here to defend his usage of LLMs for this rather trivial task (which TBH could probably be done reasonably well with traditional means) are part of why the Internet – and therefore the world – looks like it does right now. It’s a set of arguments that wants to delegitimize political and moral actions based on libertarian and utilitarian thinking.

GenAI has an Alignment Problem by Richard George.

But the mundane reality is much simpler: LLMs fail to effectively solve the problems we have, while creating a vast new class of problems to be solved. They are, ultimately, completely mis-aligned with our needs, and incompatible with the society we live in.

Relatedly, why AI isn't actually helping software companies. Dax Raad just dropped the most honest take on AI productivity written up by JP Caparas.

everyone's talking about their teams like they were at the peak of efficiency and bottlenecked by ability to produce code
here's what things actually look like:
- your org rarely has good ideas. ideas being expensive to implement was actually helping
- majority of workers have no reason to be super motivated, they want to do their 9-5 and get back to their life
- they're not using AI to be 10x more effective they're using it to churn out their tasks with less energy spend
- the 2 people on your team that actually tried are now flattened by the slop code everyone is producing, they will quit soon
- even when you produce work faster you're still bottlenecked by bureaucracy and the dozen other realities of shipping something real
- your CFO is like what do you mean each engineer now costs $2000 extra per month in LLM bills"

The only developer productivity metrics that matter by John SJ Anderson.

1. How often does the team routinely ship new versions of the software they build?
2. How often do things break when the team ships a new version?

Giving University Exams in the Age of Chatbots by Lionel Dricot.

A programmer's loss of identity by Dave Gauer.

The social group I still identify with shares my values. We value learning. We value the merits of language design, type systems, software maintenance, levels of abstraction, and yeah, if I’m honest, minute syntactical differences, the color of the bike shed, and the best way to get that perfectly smooth shave on a yak. I’m not sure what we’re called now, "heirloom programmers"?

"Acoustic" programmers (like guitars)? "Thought-powered" programmers (like gas-powered cars)? I'm not ready to be an heirloom yet!

AI Data Centers: Power-Hungry, Water-Thirsty, and Rare-Earth Reliant by Daniel.

Comments

[asakiyume]

Re: AI and phones, it reminds me of all the photos people post of the auroras, when that's not what we actually see with our eyes. Something like it, maybe, but not hot pink or neon turquoise. They're lovely photos ... but not a record of what was actually visible. It feels to me like the beauty and the mystery of the real thing get lost.

I wanted to take a picture of a very gray sky the other day, and the phone turned it blue. I adjusted it back to gray.

But that's a minor side point. Thanks for the links. I loathe generative AI and think it's hastening collapse that really doesn't need any hastening.