Assorted links — 2026-08-05
I am experimenting with a new way to collect and create the things I read on the web. I like the link posts from https://marginalrevolution.com and so am going to try to do something similar over the next few months. I'm going to start out by sending these to my subscribers, but if you don't like to get these in your inbox let me know and I'll stop including these kinds of posts in my email alerts. The "AI Observatory" is a tool I'm working on to help manage this.
The “In brief” summaries below are AI-generated from the page text captured by the AI Observatory.
Describe the world in poetry. — poetry.camera. This is a great companion device to Poem/1 https://poem.town in a kind of way.
How a Beam Engine Works — An Interactive Guide — Gary Linscott.
Experimenting with the new pure CSS random() function | Polypane — polypane.app. This is cool and beautiful.
GitHub - mathiasschilling/A-B-Test-Using-Generated-Personas-and-LLMs · GitHub — GitHub. Just another example of how LLMs are moving into the product development life cycle.
Bending Spoons to buy Airtable for $1.28B | TechCrunch — TechCrunch. via Daring Fireball, this is interesting because BendingSpoons knows how to turn around “distressed assets” and operating them well. Things to ponder - will we see more of this kind of activity in data/no code/low code/AI businesses - I’m sure we will. Secondly are there operations like this in scholarly publishing? If they emerge will that indicate that this is a moment where scholarly publishing is overvalued, equally if they don’s is that an indication that scholarly publishing companies are not overvalued?
[2604.11184] Taking a Pulse on How Generative AI is Reshaping the Software Engineering Research Landscape — arXiv.org. This paper sounds more interesting than it is, they looked at how researchers who study GenAI are using GenAi, and found that they are indeed using GenAI, but agree that things are not very stable.
In brief: This paper surveys 457 software engineering researchers to characterize how generative AI is reshaping academic practice. It finds widespread adoption concentrated in writing tasks, alongside significant concerns regarding accuracy, bias, and the need for clearer governance. The study establishes an empirical baseline for the responsible integration of these tools into research.
Regarding governance, almost all respondents agreed that some regulation is needed, though they stressed it should focus on principles rather than specific models. Peer review emerged as a particularly divisive topic, with views ranging from outright prohibition to conditional acceptance with safeguards. These positions reflect the difficulty of establishing norms, also because GenAI technology is still evolving rapidly.