David Alan Hjelle

I’m assuming this “just” is formalizing what’s been working already for some time?

I’ve heard things like SSO — needing the username to figure out what service to pass the password to and to display the correct logo or whatever.

Not really a sufficient reason in my mind. 🤷‍♂️

Are you using it integrated with the desktop app or have you turned that off? My problems with the Safari extension have mostly been authenticating (of all things). I have disabled the authentication integration with the desktop app for a couple days, and it seems better now? FWIW

Theodore Roosevelt National Park. It’s across the state, but gorgeous. If I can go the other direction — another state but closer — Glendalough State Park is another favorite.

Not to mention whether we are treating people as humans or robots.

I am philosophically/theological convinced that we aren’t on the road to AGI — but even a “good-enough” simulation really has some rough impacts on real people.

Trying to find the right balance in our org and even just for myself.

I share those concerns!

I wish I knew whether LLMs end up more like:

- machine code to higher-level languages, where (generally) skills didn’t atrophy, but did change significantly
- artisan craftspeople to factories, where we lost real skillsets

I don’t know!

It’s always a dance with tradeoffs. I work for a small company — and have stayed, in part, because my employer and I share a lot of values. I have probably limited my career in some sense by doing so, but I’d consider myself very fortunate.

It sounds like your experience has been more difficult?

The Floodlight of Christmas: politics-and-poetry.thehjellejar.com/the-floodlig...

I sure agree learning — in all forms — is high-leverage; it pays off in oft-unexpected ways.

On one hand, “productivity” is overrated in our culture, sure. On the other: I agreed to be paid by my employer for a job; it is *our* decision more than *my* decision what things are worth paying for.

I’m skeptical, but trying to be open-minded, of productivity gains in many uses of LLMs. Hard to figure out where the truth actually is given so many claims. (With my own likes and dislikes also getting thrown in the mix.)

This is one of my struggles, too. I’ve been trying to figure out — both in the _programming_ part and also in the detective-like _debugging_ part. I wish I knew how to measure productivity, though: my job is to accomplish a task, and have fun if I can. How far can I justify my preferred way?

My bedstand books represent an aspiration. I would like to have read them. Most I haven’t started. Hairshirt, yes.

I also have _How to Think_ and _Don’t Overthink It_ next to each other, whatever that says.

Pardon me…I have a detective story to finish reading…

@cferdinandi.bsky.social I noticed on kelpui.com/docs/compone... you mentioned that one ought to only use date pickers for dates that are in the near future or past. Is this primarily because of the amount of clicking to get to 50 years ago, or is there another usability issue I’m not thinking of?

As long as it is subsidiary to Parnas’s “what is going to change?” (paraphrased a bit…)

Amen!

The Re-Formation of Manners: politics-and-poetry.thehjellejar.com/the-re-forma...

Heavily reinforced in the human part of the training? And all the vendors use the same, cheapest, vendors?

(I actually have no idea.)

I tried hard to expand my musical tastes so I could enjoy music with my kids, no matter their interests. I did *not* account for Baby Shark.

You have my deepest condolences.

I don’t think it has to be that way. But good interfaces that respect their human users are hard in the best of scenarios, and much harder when we don’t fully understand the underlying technology.

…and sometimes due to almost incidental UI decisions. For instance, we worry about people making “friends” with LLMs…but the chat interface we give them is mimicking the same chat interface they talk to humans with.

I get utility out of LLMs! But we have given them such misleading interfaces.

ChatGPT’s UI says “What can I help with?” followed by “Ask anything”.

Yeah! I think this is my biggest frustrations with LLMs as we have them (after setting aside the hype-train): the language and interfaces we use to present these tools to people are actively misleading. Sometimes due to language (as previously discussed or as in, for instance, ChatGPT’s UI)…

I totally agree — I’ve been doing the same. It’s sometimes frustratingly difficult to not anthropomorphize these tools, but LLM is at least a start!

What features are you looking for? Chart of Accounts, check writing, credit cards, reconciliation? AR or AP? Payroll? Other?

If you ever come to Minnesota, there is an installation of his trolls in Detroit Lakes!

Have you found thoughtful Christian resources on the topic? Or do you find the topic uninteresting and unuseful?

Thank you!

Fair! Though I sure didn’t mean to sound complacent. Even though there is nothing new under the sun, I’m trying to figure out my dent to make, so to speak.

Do you have some links to the environment and human displacement issues? I haven’t dug into those as deeply as I probably should…

Not that it excuses anything! Just observing that it’s a problem not specific to AI, I think.

Seems pretty par for the course for humanity for millennia, unfortunately. :-( We want what we do not have, so we do anything to get it.

Amazon Programmers Say What Happened After Turn to AI Was Dark Amazon engineers saying they're being forced to use AI while being told to do more work on even tighter deadlines.

Humanity has a long history of adopting new technologies (defined broadly) without thinking through the implications.

What side effects do you have in mind? Stuff like futurism.com/amazon-progr... ?