There probably wasn’t a billion dollar idea in there. But there might have been a $100,000 idea in there.

I’m sure there are plenty of people who’d pay $10 once, and a smaller proportion who would do so monthly, for a desktop friend branded as their favourite media character. I’m sure we’ve been doing standalone toys like that since the 90s if not before.

But they went and nickel-plated the implementation. Was it over ambition, misjudged economics, or a belief the product had to fit into the AI blitz?

The smart plan would have been to bring up a small number of curated licensed personalities. You know Miku would sell, for example.

If you have a target character, you no longer need a LLM. Use a canned base of high quality data or curated sources, and you can drive it with Eliza-style keyword mapping and some undergraduate-project Markov Chain stuff to spin the dialogue a bit.

But they promised arbitrary characters and open-ended conversations, which all but forced dependency on a huge data model of some sort. That’s a huge expense to service the long-tail customer who wants to talk to the ghost of Robert Mugabe or the protagonist from a 25-year-old C-tier manhwa, or so you can pivot from asking Batman about the Riddler to his suggestions for replacing the muffler of a Ford Aerostar. Also, you now have a content domain far too broad to fully test and rein in, ending in tragedy for some vulnerable users.

If everything worked just right, it might outperform the local “themed Bonzi Buddy” concept on some sort of Fanboy Turing Test, but was that a battle you even need to win to be a successful product?

Even before the Slop Mandate, I’m not sure I could be less enthused about that product.

Let’s cross over the trillion-dollar franchises! That’s not the premise for a classic game you’ll proudly share with your children one day. It’s the premise for a 16-year-old fanboy’s first outing on AO3. It’s the premise for something that shuts down in 9 months due to declining interest after two half-hearted content updates and nine gacha banners.

Hot take: if your game mechanics and story wouldn’t work as an unbranded, franchise-free version, no amount of smearing Darth Vader and the Incredible Hulk on the carcass will make it better. (It will, of course, sell better, because Star Wars and Marvel Boys seem to inexplicably have access to credit cards.)

I would think knowledge work is really compatible with unionization, as it has less “fungible” employees.

If one assembly line shift goes on strike at GM, they can switch in another crew, or hire scabs and get pretty close to back to normal.

If your software dev team goes out, any replacements don’t have 200 years of experience on exactly where the bodies are buried in the specific design and architecture. You will suffer ramp-up problems. A strike threat is way bigger, especially if it cuts into keep-the-lights-on stuff like SRE and operations too.

12VHPWR seems to be hubris manifest.

It used to be that the only PSU cables thst regularly caught fire were the $2 Molex-to-SATA converters.

“Kill my boss? Dare I dream the American dream?”

The other fun thing is with the new calendar, they literslly reset the count to year 1. The coins for the first few years literally said “l’an 7” or “l’an 11”.

“We decide when history starts” is a hell of a sentiment.

It’s interesting that they labelled the models with the exact ratios, and as “OO scale” or “TT scale” to imply compatibility with existing model railway setups (OO is typically 1:76, and TT 1:120)

That appears to be an IBM (or possibly later Lexmark or Unicomp) Model M.

They are unusual in that the mechanism itself is arched so all the keycaps are the same shape and can be reordered neatly.

If you tried a similar play with most other keyboards, the keys would stick up at uneven angles and it would look weird

The frontier vendors are great hype people.

Local Deepseek (or frankly, Eliza running on an 8086) is probably perfectly fine slop for 95% of your slop needs, but you’ve gotta have the biggest and shiniest slop. You clearly need a LLM powerful enough to invoke the apocalypse to polish that marketing budget spreadsheet, or you’ll be left behind.

Sometimes the retention clip cracks. The old 1700X my brother uses failed that way and the fan was furiously spinning to try to compensate.

I’m fond of qmmp because it ticks the boxes: it fits sanely into my desktop aesthetic since you can use normal Qt theming, it supports MPRIS for integration into status boxes, and the download-lyrics plugin works (surprisingly rare, many onky surface embedded lyrics)

I don’t have the boobs to pull off Demona.

The second order effect is the value, especially for a much bigger wealth tax.

If we pulled a few hundred billion out of Larry Ellison, Elon Musk or Jeff Bezos’s pocketbooks, we could buy the Air Force twelve or thirteen screws.

But more importantly, we’ve cut the mikes from their stack of financial amplifiers. They can’t command the attention of government far beyond their position as individual horrible men anymore. Their particular vision for a futuristic hellscape has to compete in the market of ideas. They can’t afford national big-brand media or to surround themselves with phalanxes of lawyers.

This social benefit doesn’t show up on a flat spreadsheet, and it’s willfully ignored by analysis that stops with “the money dries up.”

What I also don’t get is where they will get ever larger data sets that will make a significantly better model. Even if the benefits were linear+, you’re running out of new fresh, quality resources to tap.

You add in another 50,000,000 corporate emails you harvested, and stastically, they’re a whole lot like the 100,000,000,000 you already saw. Conversely, you destroyed the last copy of some 1929 treatise to scan it in, just to dilute it in a hundred billion other pages.

Moreover, the further into the “long-tail” of data sets you go, the less value they provide for mainstream audiences. If you tap an archive and Claude Code gets excellent at Sinclair ZX80 BASIC programming, that’s not really what a bunch of React jockeys need.

If you kept the quarter, it’s almost $12 of silver today.

Given the pricing info, it’s for retailers, who will see a worse outcome for missing the sales window than a vonsumer who can just turn their light out and hope to not get egged.

The operations team closed all open merge requests older than 2 years at my firm recently.

I got notices for tasks I had started going back to 2018.

There’s Matilda, the widow who built her web by the front door. She comes out each night. I haven’t seen Gretchen, the garage widow, in a few months.

The inside has a bunch of pholcidae, many of whom are getting collossal. If you get too close they do frantic press-ups in their webs.

Thank you for creating arachnid content!

A legitimately interesting take.

I can see two angles:

1. AGI is implicitly expected to become ASI because many useful technologies have scaled to levels natural evolution didn’t.

Our strongest weightlifters didn’t grow into thousand-horsepower cranes. Our quickest math savants are not cranking out a teraflop of FP32. But conversely, that’s why cranes and CPUs are economically interesting.

2. Not really sure being able to carry on a plausible conversation on a “new subject” like the game example is much of an evidence of intelligence. ELIZA did it with a few kilobytes of RAM and some manually hardcoded rules. Having a huge set of training data just expands the opportunity for a plausible match.

I think the lack of self-improvement and finite contexts are sort of damning. Outside of rare neurological disorders, I can’t make intelligent people forget how to ride a bike by reading the phone book to them.