Harsh Thoughts

Sign in Subscribe

Harsh S

Intelligence Is Commoditizing. Trust Is Not.

Intelligence Is Commoditizing. Trust Is Not.

Model intelligence is becoming abundant. Trustworthy task completion remains scarce. I have built enough hobby projects with AI to recognize a recurring pattern. You discover that a frontier model can seemingly build an entire application on its own, so you try to one-shot bigger and bigger pieces of the

Kids Understand AI Better Than Most Founders Do

Kids Understand AI Better Than Most Founders Do

Three in four teens have used generative AI. The adoption number is unsurprising. What matters is how they use it. In one study of fifth graders, kids consistently distinguished three roles: AI as companion, as collaborator, as task automator. They were uncomfortable with the third. They worried that if AI

The Hard Work Happens Before the Prompt

The Hard Work Happens Before the Prompt

A friend sent me Andrej Karpathy's neural networks course last week. We spent an evening working through the first few lectures together. What struck me wasn't the content itself, embeddings and tokenization are familiar territory for most people building AI systems. What struck me was that

The Agent Was Authorized. The Purchase Was Still Wrong.

The Agent Was Authorized. The Purchase Was Still Wrong.

Payment networks solve three problems: identity, authorization, and settlement. They confirm who made the transaction, whether they had permission, and whether the money arrived. That design worked when humans made purchasing decisions. It does not work when agents do. An agent can pass all three checks and still create an

AI Doesn't Need New Rules Yet. It Needs an Audit Trail

AI Doesn't Need New Rules Yet. It Needs an Audit Trail

Steven Sinofsky recently argued that many AI harms may already be covered by existing law. Fraud, bias, negligence, copyright violations: all illegal before the first transformer was trained. He's right, but incomplete. The immediate problem is not missing prohibitions. It is missing evidence. Existing laws become harder to

Kimi-K3 and the End of Model Selection as Strategy

Kimi-K3 and the End of Model Selection as Strategy

Yesterday at Atlanta Tech Week, I met two founders building custom AI models for companies. Their pitch: pay them a five-figure fee, they train a model on your data for your use case, trading algorithms, customer service, compliance review, and you own the weights. You run it on your

Every Traditional Hiring Signal Is Breaking Simultaneously

Every Traditional Hiring Signal Is Breaking Simultaneously

Run three technical screens this month and you will likely see the same pattern: comprehensive, technically accurate answers to questions that should require synthesis and tradeoff judgment. The responses are perfect. And completely hollow. No experienced engineer talks that way when explaining a system they've actually built. The