Digital garden

“What's the difference between water as irrigation and water as a flood? In the first, the human is determining where to send it. In the second, the water is making the determination. Sure, it can find its own path, but that path can be destructive and you often won't know which it will be until the damage is done. When we let AI do work for us, we are allowing it to flood our lives. It produces slop, keeps us from learning, voids trust, etc. But when we harness it, it can be extremely powerful. I've been thinking about the idea of AI music. It feels icky, unless you think about a music producer who knows how to do everything it takes to produce great songs, but they can't sing. They can hire a singer, but even then, there's a limit to what they can control. AI music generation allows them to fill this gap, [Harnessed by their expertise](Harness%20your%20expertise.md).”

Harness your expertise

There are three distinct focuses on design today–sensory (TV), software (UX), systems (Architecture)

“Why does nobody care about anything? The world is full of stuff that could be excellent with just 1% more effort. But people don't care.”

“A bullshit job is one that even the person doing it secretly believes need not, or should not, exist.”

“I would hope that everyone, regardless of their background political convictions, should aspire to be less full of shit.”

“any graduate student with a clipboard can demonstrate that reasonable-seeming people are often totally irrational”

“One implication of the naturalness with which we divide cognitive labor, is that there’s no sharp boundary between one person’s ideas and knowledge those of other members of the group.”

“Agentic systems make it possible to offload larger, more repeatable, and more modular units of work, potentially reducing the importance of some routine execution tasks while increasing the importance of judgment, oversight, coordination, and review.”

“Artificial intelligence can do more every day, but deciding what it should do is up to us: individuals, organizations, humanity as a whole. These decisions require knowledge and judgment that people acquire through continuous contact with the work, increasingly done alongside AI. Shaping the goals of advanced intelligence is also a continuous process of feedback, learning, and realignment.”

“Even with the best intentions, a model shaped in one place inevitably encodes the values of its owner, not the individual users it serves.”

“The model you choose to serve your users affects not just your P&L but your product quality, which affects your competitive position. Use a cheaper model and you protect margin but risk losing users to a competitor with a better experience. Use a frontier model[3] and you win on product but bleed on economics.”

“whatever potential hazards and unintended consequences flow from AI’s mass adoption, my own experience using it contradicted much of what I was being told. One fear is that these systems will strip prose of style, idiosyncrasy, and that elusive element “voice,” but I found that they could do the opposite, introducing the unexpected, the “wrong,” the uncanny disjuncture between the realistic and the oneiric.”

“The emergence of world models points to a broader shift in AI. The systems that define the current era operate primarily through language. Multimodal systems expand that capability to images, audio, and video, but most still cannot track a coherent environment over time. World models, by contrast, aim to maintain a coherent representation of an environment over time and predict how it may change in response to action. That capacity — to understand a physical environment and use that information to guide action — is spatial intelligence. World models are one technical pathway toward achieving that capability.”

“If you expect a language model to derive all of its answers from a single, monolithic, coherent internal model of the outside world, all the behavior described above is evidence that this world model doesn’t exist.”