scrappyjoe

Park it in BRK?

I set up a little neighbourhood pbx this year on an oracle cloud always free instance. Took a couple of days.

Any family can buy a WiFi-enabled office phone and I’ll set up an extension for them. It’s working great! My six year old had a 15 minute chat with classmate while we were making dinner today; they have arranged a play date for next Monday.

A couple of weeks ago a 5 year old invented prank calls. Every now and then the phone will ring and we’ll pick up and she’ll sing a a couple of lines out of Frozen before hanging up. It’s made our community much closer.

I’ve had a couple of weeks of downtime at work, so I decided to incorporate agents into my work processes - things like note taking, task tracking, document management.

Your comment EXACTLY mirrors my experience. Week 1 was ever expanding prompts, and degrading performance. Week 2 has been all about actually defining the objects precisely (notes, tasks, projects, people etc) and defining methods for performing well defined operations against these objects. The agent surface has, as you rightly point out, shrunk to a translation layer that converts natural language to commands and args that pass the input validator.

Cool, I'm already digging into your stuff, thanks for posting it.

That's a really interesting perspective.

I've always taken the elite overproduction thing as an _analytical tool_ to help us make sense of why we have experienced the rise of an oppositional anti intellectual position in contemporary culture.

But you make the good point that it can also be a _weapon_, leveraged by those oppositional groups, to justify their oppositional position.

Perhaps this seeming tautology can be resolved with some systems thinking. Maybe there's some insight in the elite overproduction analysis, but that means that, as an argument for further polarising society it's a pretty effective tool. It's actually reinforcing the feedback loop! A fascinating example of a self fulfilling prophecy.

> I associate this phrase with losers

Completely separate from the substance of your point, this sort of language does not encourage constructive dialog, it frames the discussion in such a way that you are either going to get

a. People who agree with you, resulting in you not learning anything b. People who are triggered into fighting with you, once again, resulting in you not learning anything c. People ignoring you, resulting in you not learning anything.

My constructive suggestion to you is that you simply don't write that first sentence. I suspect you (and everyone else!) will have a much more fruitful time online as a result!

Edit: Spelling correction

Maybe this is how we get code versioning for Excel.

Git LFS for workbook + the following prompt :

“Create a commit explains what has changed in the workbook since the last commit. Be brief, but explain the change in business terms as well as code change terms.”

I know an Rmarkdown rendered script when I see one!

Jokes aside, I have been aware of System Cards for a while, but found them sort of irrelevant. As the tide of AI has started lapping at my profession, and I have found myself suddenly spending 1/4 of my day trying out Shiny New Stuff, Model Cards suddenly seem immensely interesting and useful.

I hope that the industry keeps publishing them, even if there is a muted response, because they are quite valuable snapshots of the state of the art, as measured against interesting dimensions, at a particular point in time.

Someone is going to collate all these model cards on a timeline and plot some charts into a _killer_ blog post in about 5 years.

Doesn’t renv do that? What need does renv not meet for you?

I've got one of these, and it works great! The core developers behind this work very, very hard on it, and I am amazed that it works to the degree that it does. Like, there are cars on the road running with this VCU.

Two weaknesses of the project are spotty / outdated / unclear documentation and support. This is not because the community doesn't care, it's because keeping the documentation up to date is hard. And support is hard, because you have like 3 core commenters on the forum answering 90% of questions. They know the stuff inside out, but they take strain from having to answer 'my hello world is broken' type questions over and over again.

It's hard to demonstrate this to software people. You know the endless 'works on my machine' swamps people would get stuck in in the bad old days? Imagine that, but you also have no assurance that the machine the person is using has a hard drive connected, or a cooling fan. You have to cover _all_ the hardware related root causes before you can even begin to address and software related issues people run into.

I think the best support this community could get is more resources dedicated to documentation and support. That would also serve to make the learning curve less steep.

EDIT: I think it's also worth saying that once you set one of these once, it becomes very easy. It's just figuring it out the first time that's hard. If anyone buys one and has trouble, hit me up and maybe I can help.

Well, if you modify a vehicle, generally you need to take it back to your DMV / TUV / MOT for inspection. Those are the bodies responsible for determining that your car is road legal. And if you don't have a proper BMS, emergency cutoff, adequate cooling and engineering, it won't be certified road legal. Not sure why an EV conversion should be treated any differently to an ICE conversion.