Ross's Micro Blog

Anyway, I’ve mostly been ill for the last two days and all that has, apparently, been brewing in my neurons while my immune system worked (and boy was it working hard!!)

🤷

I do really lament the apparent lack of (or possibly mis-directed?) curiosity that LLMs seem to have created.

Sensible assessment is mostly drowned out by hype.

And I really worry that we’re handing over critical thinking to corporate-controlled machines that may not be there in the future.

But mostly I’m thinking about how there is this AI divide in tech.

I’m sure there is something about neurotypes or personality types involved.

I’m a moderate, curious thinker. And I wonder how we can bridge this divide and find some helpful common ground. 😕🤔

Only atoms and lists, as far as I recall.

But that’s a good foundation right there!!

(I reckon this was my first academic textbook)

https://infosec.exchange/@0xabad1dea/117360462307042201

I properly read this post about the “gestalt” of tools yesterday. I think it’s incredibly insightful.

Tools are part of systems.

There’s something about systems thinking and friction-as-a-good-thing that I’ve not pieced together properly yet.

https://deadsimpletech.com/blog/no-such-thing-as-just-a-tool

Like, I was _genuinely_ excited about learning how to do a new thing the other day. I like learning how to do new things. That’s exciting.

But the world thinks it’s more exciting that a computer did a thing for them and that they had to do nothing!

“Look at all the learning I’m NOT doing!!”

Sigh

I’m also hearing, and deeply thinking about, the relentless (and possibly toxic) positivity of the pro-AI crowd.

I get why one could find this era attractive and exciting.

But how can the alternatives be presented in a positive light without sounding glib?

Thing that I’ve not seen discussed anywhere:

Chat interfaces must be a real accessibility boon for some.

But do they exclude people who struggle with words?

Like, is this LLM world really problematic for people with dyslexia?

Is literacy going to become even more important if LLMs persist?

I’m also thinking about chat as an interface.

Text chat is a crap interface for many tasks. It has no affordances. It’s often very slow. There’s cognitive load. And with LLMs involved there’s the possibility of the wrong thing being done.

Plus… like… just LEARN TO DO THINGS!

New models arrive every week. Just because we can. And orgs want to own to own the news cycle.

And what of these things will be here next year? MS’s Recall is dead. Sora video is dead.

The “more” of everything is relentless and stupid.

Today I’m thinking about product release cycles.

How Jev arrived and it will quickly be ripped off by one of the LLM houses.

How Meta’s Muse arrived and we already have OpenAI’s Dots as a response.

Everyone has to do the same thing as quickly as possible. Just because we can!

I should probably write down more of my thoughts. They’re really quite interesting. Well, I think so, at least.

The best I can say about how I feel today is: gosh I’m glad my wife and I didn’t have this virus at the same time!!! 🤒😭

UK WordPress folks… 😃

https://toot.wales/@rhyswynne/117347634436986062

This is an excellent and reasonable take on programming with LLMs. Sums up a lot of what I think about it.

https://mastodon.social/@adactio/117332468073001930

This is a very good point. Anthropic are arguing for AI rights. But we should do human rights properly first.

https://bsky.app/profile/did:plc:ovmoyxicp4hramrokx4vk3o5/post/3mwfsuwwhp226

This is such a brilliant idea. I’m gonna do this for sure.

https://social.lol/@annie/117333782061414131

I’m open minded and want to hear different points of view. So I watched DHH at his conference.

He seems to have lost any critical thinking that he once had.

OR

He’s deliberately ignoring some stuff.

You can’t present the “logical choice” when you’ve not considered any downsides in depth.

Yes yes yes! Saw someone bragging this morning that they shipped 70 days out of the last 100 or something.

WHY??? What did you make that needs a new release on 7 out of 10 days? I don’t want that.

Think first. Build later. Stop when you’re done.

https://tldr.nettime.org/@tante/117313713101758213