Saturday, June 28, 2025

LLM AI = "people spirits", SW 3.0

very interesting an potentially influential view on AI and future

Andrej Karpathy on X: "Nice - my AI startup school talk is now up! Chapters: 0:00 Imo fair to say that software is changing quite fundamentally again. LLMs are a new kind of computer, and you program them *in English*. Hence I think they are well deserving of a major version upgrade in terms of" / X

"14:39 LLM psychology: LLMs = "people spirits", stochastic simulations of people,
where the simulator is an autoregressive Transformer.

Since they are trained on human data, they have a kind of emergent psychology, a
nd are simultaneously superhuman in some ways, but also fallible in many others.

Given this, how do we productively work with them hand in hand?"

Andrej Karpathy: Software Is Changing (Again) - YouTube


Karpathy vs. McKinsey: The Truth About AI Agents (Software 3.0) - YouTube

1. Software 3.0 Paradigm Shift: Andrej Karpathy argues the next “language” of coding is English, forcing teams to rethink every layer of software design—from data pipelines to deployment. 

2. LLMs as “People Spirits”: Large language models are stochastic simulations with jagged intelligence; they feel human but still need tight human supervision and constrained output. 

3. Human-in-the-Loop by Design: Success hinges on making validation frictionless and deliberately limiting AI generation so reviewers can keep up. 

4. Builder Honesty vs. Boardroom Hype: Karpathy names CI/CD gaps and edge-model limits, while McKinsey’s “agentic mesh” offers a seductive yet unbuildable fairy tale. 

5. Enterprise AI Reality Check: Plug-and-play agents and tiny edge models don’t exist; incremental crawl-walk-run adoption with clear culture change is mandatory. 

6. Risk of Consultant Oversimplification: CEO faith in word-salad frameworks stalls projects and wastes budgets—tech leaders must push for empirically grounded plans. 

7. Edge Computing Debate: Despite bets from Apple and others, large centralized models still outperform small edge deployments in 2025; prudence beats hype.


"enterprise consulting speak" (useless)

No comments: