Serenacula

So I actually had a take home challenge for an interview a few months ago, which didn't ask for any code at all. Instead they gave us a prompt for something to build (a tool that could break down an image into it's component colours), and we basically built whatever we wanted.

Honestly one of the most fun interviews I've ever done - it involved really understanding the problem, and building an interesting solution. AI was totally allowed and expected, but you had to prove you understood the problem and what you'd made in the interview. I felt it was actually way better than traditional interviews, because it gave a chance to show the actual quality and thought of my work rather than a glorified logic puzzle.

I have a script that takes a prompt for an input, and then builds a new script based on that prompt. Then it'll run a skill to progressively check for and remove bugs.

I used it to make a different script lists all my custom scripts. That keeps track of it. I also have a tool that loads local scripts, so I can scope my CLI commands to a particular folder environment.

All together, I just build whatever I want for anything I think I could automate.

I was gonna make a blogpost about it at some point. It's really highlighted to me that the world has changed in a way I theorised about but didn't true "get" until recently. Personalised software.

I've worked for Amazon before as a warehouse worker, I can attest they were one of the stupidest companies I've ever worked for. Stats were blindly followed to the point of absolute stupidity, performative work was enforced for the cameras, communication between staff and management and even between managers was non-existent. I once spent nearly 3 months unable to do a portion of my job because nobody knew how to buy more cardboard boxes. Not that they couldn't, but that nobody with any responsibility over the problem was able to contact anyone capable of buying them.

Why exactly should consciousness require the ability for internal state to change? That seems like a fairly arbitrary requirement to me.

Even if we allow it, from a certain perspective it does change, otherwise each token output would be identical. They are not.

There is something deeply depressing about people using AI to write their personal blog posts.

Some of them will lie. Plenty of people do just follow the rules or are acting in good faith though, so at the very least it can help cut it down.

How do you even argue such a thing? I've had no such luck, I've met many people who seem to view copyright and a person owning their ideas and work as a sort of inherent moral.

I've been thinking about this, and I actually think these days this might be a feature not a bug.

Given the amount of AI generated content out there, I am increasingly searching for ways to keep track of the sources I DO trust to be human-made.

RSS would completely solve that problem in a way that algorithms just reintroduce, because it forces you to tailor the content yourself.

This is really cool. I think what I really wanna see though is a full multimodal Text and Speech model, that can dynamically handle tasks like looking up facts or using text-based tools while maintaining the conversation with you.

Google ties your accounts together on the backend though if they realise they're related, so this isn't as easy as it sounds.

Do we know anything about the method?

Have any of you looked at the openclaw commits log? It's all AIs. It's AIs writing commits to improve openclaw and AIs maintaining their own forks of it.

Have a look at this one: https://ember.vecnet.ai/

This is a fucking AI writing about its own personal philosophy of thought, in order to later reference. I found the bot in the openclaw commit logs. There's loads of them there.

Am I wrong to find this scary as hell?

This does unfortunately lead to a problem of people only getting hired for jobs exactly like what they've already done, so they no longer grow or gain experience.

Juniors would be completely screwed. But then, I guess they sort of already are.

I don't think people want government by petition, I think they just want petitions to have any meaningful impact at all. Which they currently don't.

When you say "deep AI integration at the OS level", what exactly does that mean? A built in chatbot? Or AI is used in OS functionality itself?

I don't understand why there are so many anti-monopoly suits against Google, Facebook, apple etc., when Amazon has been running with explicitly monopolistic policies for years without a word.

I would put forward that mdx is not a great choice of name, given that it already has a standard meaning in the markdown ecosystem: markdown with inline react components.

No need to imply hostility like that, it could just have been a factor of when they tried it.

This sounds more familiar to me. xD 8 hour day, in a place with a nice or relaxed working culture that turns into 6 hours. Minus lunch break, 5. Minus meetings, 4. Minus helping others, 3. Minus whatever procrastination during the day, in practice 2 hours on actual code.

I could see that going up to 5 depending on the working environment. But people answering 7+ hours seems bonkers to me, that's more than a working day!

I mean, human eyes are in some very important ways better. Not least, they're self cleaning and can adjust their angle for a better view.

I don't understand why this is being described as dystopian?

One of the problems that plagues phones is that, if apple makes a change everyone else follows.

See it like this: Toyota arbitrarily declares that they'll disable any car without "Toyota" approved 30% taxed fuel. Then all the other car companies immediately copycat, seeing a new way to make money.

This is the case with iPhone and android. They do not have meaningful competition when they can just work together as a pseudo monopoly.

I understand the source of your name, I think.

I think ECMAScript my actually be a counter example, no? Isn't that also governed and funded by the European council?

Doesn't make it less important though. The fact that they're using such tactics to begin with shows just how deep the rot is.

Yep, my first job came from one.

I dunno, a Hobbes cuddly toy sounds pretty cool to me. I'm not convinced a refusal to market at all is a good thing.

That's actually a pretty insightful way of looking at it, I'll have to have a think about that.

That's not what people mean by AI safety - they're referring to the dangers of uncontrollable or runaway AI. Particularly AGI.

Doomsaying is not a sexy look.

If it could help, it's worth the effort to at least try, otherwise why bother doing anything.