TheOtherHobbes

Ask your LLM to hack it.

There's plenty of concrete evidence. But - for some reason - Trump's government has refused to release it, defying multiple legal judgements.

Good effort, but the ultimate goal has to be replacing the entire planet's economic stack with AI, from janitors to principal engineers to the entire board to dictators to Elon Musk's drug dealer, and creating a closed-bottle AIconomy trading entirely in AI speculation on the blockchain.

I was always amused that the rationalists threw themselves behind a professional trickster with absolutely no scientific credentials.

And Dawkins never had much of a research career.

My problem with the "skeptics" was that the paranormal is an easy target and politically it's punching down.

There are plenty of other irrationalities to consider - like how advertising and propaganda work, how cults and religions actually operate beyond "irrational bad", and how trad econ believes many things that are questionable at best - but the skeptics never dared go anywhere near those.

So skepticism turned into this kind of knockabout point-scoring media circus noise. And in the end the sneering and self-congratulatory condescension actually helped entrench modern conspiranoia instead of deflating it.

I'm torn between ugh and ick.

It's poorly implemented and looks crude and unpolished.

There might be an idea there, for limited applications, but this isn't it.

I think the whole idea of people running models is wrong. Models will run models, training will become distributed, and people will ask ModelNet to do whatever.

Computing tends to oscillate between centralised and decentralised models. It also oscillates between batch and timesharing.

Currently training is batched and centralised, access is timeshared and centralised.

But eventually a previous generation of computing turns into transparent networked infrastructure, and then you get another layer of new kinds of applications on top of it.

That's what happened with the Internet, and it will happen again with AI.

There's been a lot of innovation in music tech with synths like the MicroFreak.

Korg have made a series of original synths (and VSTs) that are nods to old tech with a lot of updated features.

The VST market in general is a bubbling mix of clonery and originality.

The reissues and clones are for people who want a specific vintage sound but can't afford the insane prices of vintage originals.

The original 1977 price was around £5800, which according to the BoE inflation calculator is £35,000 in modern money.

These are an unusual design, because they're an organ/synth hybrid. There's a tuneable oscillator for each pitch class, but they're divided down for the octaves. So it's more like a string synth/organ architecture than three full synth voices per key (which would be insanely expensive), but the dividers for each octave are driven by single VCOs which can be tuned manually and modulated together.

The sound is... interesting. Analog, but doesn't have the legendary reputation of other synths. Korg always made quirky little synths, and this is more of a quirky big synth. They didn't make many of the original.

The reissue is aimed at the usual Gen X/Boomer demographic who would have wanted one when they first appeared, couldn't afford one, but have the cash to burn now.

There's negative curiosity involved in trying to reinvent the mid-90s web, but worse.

It's tragic that "your own domain and your own hand-written HTML" has somehow become a glorious ideal, and not a catastrophic failure of the modern imagination.

No one who is truly content-first and not tech-first is remotely (ha...) interested in any of this.

What content-first people want is a drop-in one-click package that sets up a server, secures it and keeps it secure, and installs something like Wordpress, but federated.

And is generally simpler, easier to set up, easier to use, and better.

And prettier.

(Email would be nice too, but we can't have nice things since the oligopolists locked that down with "spam filtering" of itty bitty private domains.)

Pick from a few themes, post, read, interact, leave the customisation and under-the-hood features until later, if you get around to it all.

That's it. That's a what will create a true content-first indienet.

Anything else is just FOSS tinkering for the sake of tinkering, not a credible attempt to create an open platform.

If IndieWeb wants content, burying it under a ton of tech soup is the exact opposite of what they claim they want.

How do the developers not realise that this set of protocols is completely unworkable for 90% of users?

Please stop developing systems where the engineering is front and centre. If you want to put content UX first, you need to do exactly that, with a one-click solution that gets users up and running instantly.

If you users have to go anywhere near a command line, or Docker, or they have to hand-edit HTML or CSS, you're creating walls most users aren't going to scale.

At the moment this looks more like NerdNet than IndieWeb.

The article is a fun read. There's a certain amount of hyperbole and subtext which may perhaps be unfamiliar to some readers.

But the point is fair. Where are the huge AI wins? Why haven't Anthropic and OpenAI produced a regular series of case studies where Company X did Fun AI Thing and incredible success ensued?

Why do these conversations always end up mentioning chat bots?

Why do c-suites seem to be full of people chasing trends instead of leading?

Does anyone remember what customers are, or are corporations aiming for Platonic abstraction, where business activities become ends in themselves, with no useful output?

Manias go back a long way.

The history of the UK's railway mania in the 1840s is worth researching, because it's uncannily similar, with a stock bubble, fraudsters and chancers, huge hype, cutthroat competition, boardroom intrigue and drama, miles and miles of unnecessary infrastructure in the wrong places, and eventually a technology that became useful, but could have been developed more intelligently.

Classic "1 star review because UPS lost my delivery" metric.

I'd guess most people have had a situation where there's a corporate problem, the support person you talk to literally doesn't have the tools or the agency to fix it, but then you're asked to rate their performance on whether or not they solved the issue, with no option to say "Actually they did their best but this isn't their fault."

The ZX81's 1k chess - actually 672 bytes - was, and still is, legendary.

Just wait until the same system runs payroll and you're getting paid $1.94400 annually.

The only thing that capitalism brings to hierarchy is an abstraction layer.

You get the same dysfunctions in any organisation where the management levels are staffed by narcissists and sociopaths who need hierarchy to feel a sense of self, and need to enforce it to self-soothe.

There are political structures even more hierarchical than US corporations, and some of them punish status infractions + perceived failures with violence or death.

The problem is the dysfunctional psychology of hierarchy. The more upper levels tend to performative hierarchy and status plays, the worse everything gets.

Perhaps that's why they're attempting to build a starship out of helium.

They're not going to put their name on it if they do.

Until fairly recently "jingles" - ad music - were a huge source of income for many musicians with full time careers. Likewise library music, which was prerecorded with a specific mood so music editors could drop into their projects without having to commission it, and then the composers would be paid for performance royalties/residuals.

Writers - not such a thing now, but in the 60s and 70s a surprising number of "serious" writers in the UK began their careers writing ad copy.

Famous movie directors - many started in the ad slop trenches.

Music videos are literally ads, and it's not a surprise they use many of the same techniques and are directed by ad industry people.

And then in tech there's ad tech...

AI is just the newest arrival to the party.

They're often a collection of meaningless visual cliches spliced together in an attempt to look meaningful, and always have been.

https://www.youtube.com/watch?v=fsgWUq0fdKk

No, it's built-in. Some people just have it. Take a room full of people, give them the same training, a few might stand out and get better, most will hit a ceiling, some will clearly have no ability at all.

People who have native abilities and persistence are the ones who get labelled talented.

You may be right that AI gen doesn't count as persistence. It's too early to tell on that. But I'm fairly sure that there's a difference between people who click regen robotically and people who review, steer, criticise, think, and then hit regen, with a very deliberate aim in mind.

No planning restrictions. Throw up as many as you can afford to build and launch. No grid strain because they're self-powered.

And they're likely out of national jurisdictions, so you can generate all the porn you want.

Technologically? Maaaaaybe a slight ping time reduction in certain configurations compared to terrestrial fiber, because even with multiple hops lasers can still be faster than fiber.

Super-expensive, questionably economic way to do it though.

See also, surface.

Circular definition. Number go up = "successful".

A non-hoarding economy would look very different, be far more exciting and inventive, and wouldn't be staging an ecological suicide run.

Think of it as a giant corporate tax break, but for the little people.

Yes. People can't always afford super economical things when the initial cost is high and the pay-off takes a while, but is easily worth it in the end.

No, their biggest mistake was having no clue what they were doing or, why.

Nokia were a tech slop factory - one new model roughly every two weeks, with no obvious strategy or rationale, many only superficially different.

Every so often they'd produce a classic that was ahead of its time, like the Communicator series, then by the time the surrounding infra had caught up they'd moved on and allowed a competitor to eat that space.

iPhone and Android were both killing them, and they had no idea how to respond.

Elop was the undertaker, and much hated for reasonable reasons. But the brand was already a zombie by that point.

Very rich people do prefer cluttered homey interiors. The ideal is plush, showy, old, and ostentatious. This is the saloon at Eaton Hall, owned by the Duke of Westminster, one of the richest UK billionaires from one of the oldest families. This particular room isn't particularly cluttered, but the living spaces really aren't minimal at all.

https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcS0H4Sp...

Minimalism is more of an (upper) middle class thing - doctors, lawyers, architects, software developers. The abstraction and clean lines are cool, rational, aspirational, and controlled.

Middle class people on the artier end of the spectrum are more likely to be maximalist, with bright colours, busy wallpapers, huge shelves of books, plants, and such.

Poor people need a lot of stuff because they have to hoard things in case they break. Mass market products are usually low quality, mass-produced from cheap materials, and not very well made.

Working class spaces are often chaotic, claustrophobic (because literally small), full of mismatched items and general clutter.

There's a particular kind of upper working class chrome-and-glass look which is its own attempt at minimalism, but doesn't quite get there. It's still chaotic, the spaces are disorganised, but it's very, very clean and shiny. Instead of feeling controlled and expensive it feels sterile.

The startup office look is a mix of minimalism with some maximalist stylings. It's supposed to be informal and "fun", but the big spaces are disorienting and loud, and the bright colours and lines are distracting.

Even if we accept your premise, and not everyone does, the confusion is whether we're capable of creating an equivalent arrangement, even in principle, using alternative materials.

That's not "irrelevant", it's fundamental.

Did you actually read the text? OPs are calling that Plan D.

They're proposing an alternative, which is a global brake on frontier AI research to keep the basilisk in its jar until we work out what we're dealing with and how to handle it.

Some work is also less valuable than people - especially hackers - realise.

Hackers have an addiction to tractable problems that require effort and some skill, but have a well-defined solution.

They don't require true originality or cleverness. Barrelling through them with adequate but not outstanding skills is more than enough.

Hacker systems like Linux, Vim, and Emacs, offer exactly this. You can tinker with them to solve consecutive microproblems in a satisfying way. Likewise other standard projects like working with vintage hardware or repurposing a consumer product to do something interesting.

This kind of work generates dopamine, where spending four days trying to track down an incredibly subtle bug in a giant stack owned by a few tens of people generates frustration.

So it's not that employees don't care, it's because some work really is hard and frustrating, and solving tractable problems is far easier and more satisfying.

But is it productive? Even educationally? Not necessarily.