Neat example of AI in the humanities. A Google model trained on Latin text fills in lost parts of Latin inscriptions & identifies related texts Historians increased their accuracy by 44% when working with the AI (Though AI alone beats historians, historian + AI was usually best)

The idea of utopia goes back centuries. It’s as misleading as dystopia. Nice ideal, but this bot is a bot and has no grounding in the reality of how sociocultural systems work.

I kinda like the new iOS 26 version. Apple is very good at easing their updates unlike Microsoft, which is kinda like blunt force trauma.

It’s quite fascinating seeing the myths people are making around LLMs and not understanding how they work, that they don’t “think”. they’re statistical tools. Nothing else.

Audited a failed marketing campaign this week from my digital anthropologist lens. Reason it failed? Too much yellow, when the target culture sees yellow associated with death. Always culture.

Watched a bunch of YouTube videos on using AI agents and well, a bit underwhelming really. I think more could be done, but most use cases are quite shallow.

I don’t think there’s any reasoning with LLMs, it’s just that they’re incredibly interesting statistical systems. Kind of like the ancient abacus on steroids.

Another thing LLMs don’t “understand” is culture and how it works.

The concept of AGI is quite fascinating, but not achievable given processor and energy constraints of today.

Digital nomads are basically the 21st century’s version of our hunter-gatherer ancestors. It’s just different resources they’re chasing. And slightly more comfy.

David Graeber on social relations under capitalism and the labor theory of value. A clip worth watching.

Always find it a bit funny how each AI app or tool “completely changes everything”. At scale, wouldn’t that mean everything is cancelled out and nothing does nothing?

🤣

Interesting, but I don’t think this applies to most tech, LLMs are an exception given the nature of the technology.

There’s a stark difference between the US and Chinese approach to AI and I don’t think the US is taking the right one with this technology.

This is quite a figure.

New Anthropic research: Persona vectors. Language models sometimes go haywire and slip into weird and unsettling personas. Why? In a new paper, we find “persona vectors"—neural activity patterns controlling traits like evil, sycophancy, or hallucination.

I’d say about 95% of the sales emails I get are irrelevant to my business. Why as a marketing agency owner, would I be interested in cow feed?

The Saudi’s are very digitally progressive. Some tax revenue but also more opportunity for influencers to get paid.

My issue with this kind of research is that it’s way too early to talk about these kinds of cognitive changes at scale. This takes decades to surface.