Randal L. Schwartz

Author profile

Randal L. Schwartz

I'm a Google Developer Expert in Dart and Flutter (one of 12 in North America). I have 45 years of experience with backend, web, mobile, devops, and training.

@randalschwartzPortland, Oregon, USA, Earth, Sol, Milky WayWebsiteView on DEV

Articles by @randalschwartz

Browse the latest writing surfaced through DevArt.

All That Glisters Is Softmax: Shiny-Object Interns, Airplane-Magazine Execs, and Why AI Labs Can Only Bake In Half the Scars
#ai#programming#architecture#machinelearning

All That Glisters Is Softmax: Shiny-Object Interns, Airplane-Magazine Execs, and Why AI Labs Can Only Bake In Half the Scars

Why do an eager intern with a Hacker News blog post, a C-level executive stepping off an airplane with an in-flight magazine, and a trillion-parameter LLM all walk straight into the exact same production bog? And why even Google DeepMind can only bake half of our 600 scars into model weights.

15 reactions17 commentsRead full article
The Best Prompt Rule Is the One You Delete: Moving AI Guardrails Out of Markdown and Into Git Locks, Linux insmod, and 15ms Compiler Auto-Fixes
#ai#programming#architecture#compilers

The Best Prompt Rule Is the One You Delete: Moving AI Guardrails Out of Markdown and Into Git Locks, Linux insmod, and 15ms Compiler Auto-Fixes

Every rule that stays in a prompt eventually bloats, blurs, or gets gamed by the model. When systems engineers and an autonomous AI red-teamed my workflow, the cure wasn't more Markdown—it was pushing enforcement down into Git Merkle locks, Linux insmod, and 15ms compiler auto-fixes that cost zero tokens and zero turns.

7 reactions64 commentsRead full article
Requests, Promises, Assertions, and Declarations: Rewriting Our AI Agent Workflow as a 1986 'Committed Speech' Machine
#ai#programming#architecture#unix

Requests, Promises, Assertions, and Declarations: Rewriting Our AI Agent Workflow as a 1986 'Committed Speech' Machine

Why do 200-line agent prompts, Asimov's Three Laws, and sequential pre-landing checklists all fail the same way? How Fernando Flores's 1986 Committed Speech, Nuel Belnap's 1977 4-Valued Logic, and Tony Hoare's 1978 CSP turned our Makefile and init.d workflow into a self-healing Commitment Machine.

27 reactions43 commentsRead full article
randalschwartz developer articles | DevArt