Heard some junk tips about using AI with kids this week… Here's a quick video on how I intro kids (and adults!) to LLMs to build strong mental models for how AI works. You need models that are small (they fail more often) and local (so the kids can tweak things). www.youtube.com/watch?v=fNkt...
Local LLMs are the Best Way to Teach Kids About AI… Here's How to Get Started
YouTube video by Drew Breunig
youtube.com
On Monday I'll be chatting about prompt debt, context fails and fixes, DSPy, evolving engineering practices, and more with Tim O'Reilly. www.oreilly.com/live/live-wi...
Built a thing to help you keep your project's context clean by analyzing your agent's tools and mcps: www.dbreunig.com/2026/07/24/m...
Manage Your Agent’s Loadout with Dr. Skill
Keep your context clean by keeping your agent’s skills and tools in check.
dbreunig.com
@tonyhschu.bsky.social Up to chat mech interp and visualizing LLMs over coffee sometime in the Bay? Would enjoy talking shop and trading notes.
It is so depressing that I have reached an age where the NYT is solving the mysteries of my children's speech. www.nytimes.com/2026/07/23/m...
Saying ‘Google It’ Makes You Sound Old. ‘Search It Up’ Instead.
As online research evolves, so has the language we use to describe it.
nytimes.com
Ran up against a few Skill conflicts in an env, so I built `brew doctor` for Skills and MCPs: github.com/dbreunig/drs...
GitHub - dbreunig/drskill: Making sure your context is clean.
Making sure your context is clean. Contribute to dbreunig/drskill development by creating an account on GitHub.
github.com
I really want to know what OpenAI's sandbox set up is like.
Where do people show off their information design work these days?
Read the research papers people circulate and write down questions you have as you go. It is shocking how many people don’t read the papers.
Casey Newton@caseynewton.bsky.social · 4w agoFwiw I have found myself subscribing to even more journalism and newsletters largely in the belief that simply reading long text documents now gives you a simple and definitive advantage in the world
After last week, have a couple more additions to this list. But before I do… Do you all have any suggestions for additional tips? www.dbreunig.com/2026/05/04/1...
10 Lessons for Agentic Coding
Guidelines for programming when code is cheap.
dbreunig.com
My number one tip for teaching younger kids AI is to let them play with local models. Let them change the system prompt, temperature, toggle thinking, and model. Suddenly it's not an abstract intelligence but just software that has a natural language interface.
Understanding the tensions building in the AI ecosystem with Stewart Brand's Pace Layers: www.dbreunig.com/2026/07/03/a...
Understanding the Dynamics of the AI Ecosystem with Pace Layers
Stewart Brand’s Pace Layers framework reveals why AI’s speed is straining the slower systems it depends on for support.
dbreunig.com
This episode of East Bay Insiders is INSANE. And it highlights the importance of supporting local beat reporters. Subscribe today! Andy Duong thinks it’s worth +$20k a month! substack.com/@eastbayinsi...
Ep. 136: The Story They Didn't Want East Bay Insiders to Write
Steve Tavares reveals how East Bay Insiders became part of the FBI corruption saga and examines new allegations that could widen the scandal.
substack.com
CA residents: Just a few clicks to let your rep know you don't want your 3D printer turned into a camera into your home. www.eff.org/deeplinks/20...
We Can Still Stop California’s 3D Printer Surveillance Scheme
Ignoring EFF’s warnings about the dangers and impossibility of implementing a new mandate for 3D print surveillance software, the California State Assembly has signed off on legislation to do just tha...
eff.org
Natural language is great for one-off tasks, but suboptimal for defining system behavior. Prompt debt creeps in with each small fix, until the prompt is brittle, you're afraid to touch it for fear of regression, and you're locked into one model. www.dbreunig.com/2026/06/22/t...
The Problem is Prompt Debt
The plain-English prompt that makes prototypes effortless turns out to be a poor way to specify how a system should behave, and the bill arrives slowly, disguised as ordinary progress, until the appli...
dbreunig.com
Everything about this story is just too heavy handed of a metaphor, right down to the name, "The Reflecting Pool"
This validates something I've long suspected: less adept users of Claude Code will hit their limits MUCH faster than skilled engineers. www.anthropic.com/research/cla...
The 3 claims here are both practical and imperative: 1 “Regulation will fail if expertise remains concentrated inside labs” 2 “Open science for AI needs reimagined institutions” 3 “The missing institutions should sit between secrecy and unrestricted release” pgasawa.bearblog.dev/unsafe-ai-or...
Unsafe AI or Consolidated Power: AI’s False Dichotomy
Parth Asawa and Joseph E. Gonzalez We’re excited about AI’s potential to improve society but we’re concerned that the AI community is being polarized into...
pgasawa.bearblog.dev
Everyone building programs with AI needs to now: 1. Be able to swap out models, in short order. 2. Be able to verify, regularly, that your output isn’t silently changing. (You should have been doing this for any production system, with evals, but now there’s no excuse.)
The imperfect and awkward ways Anthropic is using to control how their models are used (with Fable now, OpenClaw a bit ago) is a great example of the imprecision of natural language as an interface. The best model can’t differentiate a bio threat from an innocuous health or research question.
Apple announcements I’m most excited for, in order: 1. Screen time improvements 2. Liquid Glass slider
My recent obsession is diversity in LLM outputs, specifically from the frontier models. They are being post-trained so hard for code and agentic use (and to be _reliable_ at these things) that its shaving off the novelty, impacting other tasks. Have you experienced this? If so, how so?
When an agent can explain anything, what is the role of human-centric docs? www.dbreunig.com/2026/05/31/w...
What Do Humans Need From Docs?
When an agent can explain anything, should we keep writing docs for humans?
dbreunig.com
DSPy requires more upfront learning than just writing natural language prompts. But once you get it, it makes building, maintaining, & improving AI programs so much easier. We aim to soften the learning curve to make the benefits more accessible, starting with more new docs and front page. dspy.ai
DSPy
The framework for programming—rather than prompting—language models.
dspy.ai
Given the state and trajectory of agents today: if someone wanted to start learning how to be a software engineer today, what book would you recommend to them?
Wired's kicker on their story about the UFO releases is the thing to keep in mind. This is a good reason why they wouldn't hide things; and a better reason why they'd stir shit up: www.wired.com/story/pentag...
