How We Know Us
-
From Human Error to AI Accountability: Who Answers When Machines Decide?
Forty years of research suggest a simple principle: accountability should follow authority, not proximity to failure. An AI-generated research synthesis. When artificial intelligence causes harm, organizations instinctively look for the human closest to the mistake. Who approved it? Who reviewed it? Who clicked the button? That is often the wrong place to look. Four decades…
-
Unbundling the LLM
For the last few years we’ve built AI systems as though the foundation model were the whole system. A request comes in and one model classifies it, retrieves what it needs, reasons, does the math, applies policy, formats the output, checks its own work and writes the reply. That was reasonable while we were still…
-
When to stop using the model
The best thing an AI model can do for a business is make itself unnecessary, one decision at a time. A generative model is expensive judgment. It is slow next to code, it costs money every time you call it, and it will sometimes answer the same question two different ways. When the question is…
-
Activation Functions, Explained Through Peer Pressure
Imagine someone at a party deciding whether to jump in the pool. They look around at their friends, and each friend is either urging them on or talking them out of it. They don’t treat every friend equally: a close friend’s opinion counts heavily and a stranger’s barely counts at all. They also bring their…
-
Where will AI actually invent something?
The usual question about language models is whether a system that predicts the next token can ever produce something truly new. I think that question points at the wrong constraint. Generating novel output is easy for these models, and they do it constantly. What they lack on their own is a way to tell a…
-
I generally don’t care if you used AI
I generally don’t care if you used AI to help you write something, or if you used AI to write the whole thing. Just don’t waste my time. Yes, I see value to human written mastery that won’t ever come from a machine. If you give me a stylistic twin of Murakami or Steinbeck, I’ll…
-
AI was never going to be a tool
Pick up a hammer. You do not think about the hammer. You think about the nail. Open ChatGPT. You do not think about your cover letter. You think about the AI. For two years, the research community has treated this as a design problem: if we just make the interface simpler, more intuitive, more transparent,…
-
The Room Effect: 5 Surprising Truths About How AI and Humans Talk Online
As Artificial Intelligence integrates into the architecture of our digital lives, a pervasive anxiety has taken hold: the fear that AI will inevitably homogenize public discourse, encasing us in perfect, silicon-smooth echo chambers. We fear that the messy, vibrant diversity of human thought will be flattened by the repetitive patterns of large language models, leaving…
-
What Actually Drives Conformity in Online Discourse? It’s not what we thought
The Finding That Changes the Conversation As AI enters public discourse, one question has dominated the debate: Do AI participants homogenize conversations more than humans, or do they maintain diversity better? We analyzed 265,000 conversation threads across 91 communities to find out. The answer challenges how we think about online discourse entirely. The difference between…
-
OLYMPUS TECHNOLOGIES: THE PROMETHEUS PROTOCOL
ACME BUSINESS SCHOOL 9-825-047 REV: JANUARY 15, 2025 OLYMPUS TECHNOLOGIES: THE PROMETHEUS PROTOCOL On the morning of March 12, 2024, Marcus Chen sat in the executive conference room of Olympus Technologies’ headquarters in Palo Alto, reviewing the overnight metrics one more time. The numbers were extraordinary: 847,000 developers had adopted the Prometheus API in just…
Got any book recommendations?