Dave Guarino

Wow, using an AI transcription tool like Wispr Flow (which is what I use to brain dump ideas) might be the biggest breakthrough in my writing process. It just makes getting started so much easier.

If it’s the person’s actual case and it’s just a name misspelling I do not see any likely scenario where they would get in trouble. Most states’ cards don’t even have names on them.

One of my favorite nouns in public benefits land is "EDG"

I've heard more than once the "should we call people users? people" households? case? or..."

An "EDG" is: Eligibility Determination Group

I iterated on it across time, but the major components were giving it the core ingredients I used today, saying I'd like to optimize it for health and nutrition and nutritional balance (e.g. fat, protein, carbs), and give lots of context (e.g. I had a somewhat elevated but normal a1c)

File under "good uses of AI": optimizing meals for balance and health

I took my base preferred breakfast (high protein yogurt) and used AI to make it really well balanced (adding ground flaxseed and hemp hearts, in addition to berries and nuts)

Little-t tools for thought On folk software, plumber programming, and agency from making things that work for you

Enjoyed this by @allafarce.bsky.social

open.substack.com/pub/daveguar...

I do have to give it to Gen Z for Turnstile.

We tackled some of this in the Data Rights Protocol work I consulted on for Consumer Reports. IMO there are already legal frameworks for people to delegate certain tasks (such as filing DSARs) to companies acting as agents on their behalf, and this isn't much different if LLMs are added to the mix

We really need to start having the policy framework conversation around AI agent access to online government services.

This is going to happen no matter what. Not having even a basic policy scaffolding that sets norms will create a selection effect for less good actors.

Excellent bumper sticker.

You definitely know I occupy a very specific niche when I say to a co-worker, "The U.S. is in a down round on the safety net."

Deeply enjoyed Claude Code chastizing my colleague for having the local version of an alpha app talking to the production database.

(It was fine, but also funny to see Claude panicking about it.)

Oh very cool!

In my newsletter today I sketched out my own AI moonshot, a hyper niche one:

Can we use today’s AI coding capabilities not for greenfield app code generation, but rather for automated characterization test coverage on large legacy systems we’re afraid of changing?

A moment of appreciation for my morning walk.

A perhaps heterodox view in my world:

I don't think the 50 states building roughly the same system is a meaningful problem when the programs are administered at that state level.

In fact, if anything, AI increases the possibility of getting those systems more contextually right.

I need a political scientist to look into this, stat

You’re welcome! If you also want to test a domain you know well, the GitHub repo and spreadsheet are quite understandable and reusable.

I think for anyone whose heuristics are grounded in, say, ChatGPT 12-18 months ago, it’s well worth re-testing.

The most shocking thing of getting even semi-decent at mindfulness meditation is realizing just how damn anxious I have normally been in many contexts!

At the airport and being able to notice it come up is astounding!

Building a SNAP LLM eval: part 3 - testing nuanced capabilities (This is part 3 in a series. See part 1 where we discussed what an eval is and why building one is valuable, and part 2 where we discussed automation of writing factual test cases.) In this post, we’...

I’ve found pretty impressive results even out of the box without added policy context! With including policy in the context window, it’s usually pretty high accuracy at this point. www.propel.app/insights/bui...

Yes it is I! We are doing a lot with this. Will be posting results of a real pilot helping people navigate a lapse in SNAP benefits this week.

A kernel I need to write up:

- Abundance should take seriously AI as a lever on state capacity
- *Particularly* because much govt work today is *not* human judgment work
- What needs MORE space for human judgment is the management OF the automation, technology, and AI in govt

You can just… learn to cook.

(Sautéed zucchini with salt, pepper, olive oil, lemon juice, chili flakes, mint, cilantro)

I think this is an area where my thinking has shifted. I'm more of the belief now that, rather than government forms being an improvable vector, each form falls somewhere on a spectrum of "adversarial interaction" to "aligned interaction." As a result I believe *more* in third party representation.

Man, AI is not going to be kind to the remote software dev market…

Yup!!

Underneath this sign lies a rich and hilarious story of a no good, very bad afternoon someone had.

Where we need AI deployed is against the long, LONG list of tasks parents of young children have to wrangle. What exists here??

Filed under "reflecting on 40 years, what have I learned?"

I think I am very, very compelled at this point that instead of focusing on what you *don't* like, you should make more of what you like in the world — and give people more options to choose from.

Happy "hey are you at cfa" message week to those who celebrate!

(No, I'm not)