While we’re on the subject, can we please kill off “kilowatt hours”. Fuckers, those are Joules, 3.6 Megajoules specifically.

The gram was originally defined as “the absolute weight of a volume of pure water equal to the cube of the hundredth part of a metre [1 cm3], and at the temperature of melting ice”.

Why? Who knows. If they’d already decided on the meter, then the most logical definition of a litre is pretty awkward. You’d think it would be 1 cubic meter of water, because then the unit of volume is 1 unit of length cubed. Then one gram would be one litre of water, which is today 1000 kg, or one metric ton.

But, then for everyday things people would be using microlitres and micrograms.

Really, the metre should have been the replacement for the inch (i.e. the centimeter). Then you could have 1 litre being 1 cubic meter (what is today 1 mL) and then the unit of mass would be the mass of 1 L of water, which would be 1 gram. A person’s height would be in hm, and long travel distances would be in Mm.

But they do know litres, because the big soft drink bottles are sold in litres. They know grams if they do drugs or bake. They know prefixes like Mega, Giga and Tera if they are into computers, and they know about “kilo” if they pay an electrical bill. If they’re into military stuff, they know about “clicks”, which are actually… kilometers.

I think it’s slowly creeping in.

Or just 380 ML (Megalitres).

The #1 response I see from people about leaving Spotify and other services in favor of directly buying from artists is… Discoverbility of new music.

I’m not sure if this sentence is saying people are leaving Spotify because it’s bad for discoverability, or not leaving because other methods are bad for discoverability, or what.

But, assuming you’re saying that people keep using Spotify because it helps them discover new music. Sure, I’m not saying it doesn’t happen, I’m just saying it’s not very good, and it’s not a solved problem. It’s better than nothing, and if you directly buy from artists what you’re left with is basically nothing.

Personally, I’ve never found any new artist I care about via a tool. I found Too Many Zoos through viral social media. I love their stuff. I don’t remember how I heard about Khruangbin, but it also have been social media. I don’t think a tool would have found either of those for me because I wasn’t listening to anything that sounded like them.

I think this is giving Google too much credit.

Sure, some of it was about getting people to search multiple times so they get more ad impressions.

But, a big part of it is that Google has been waging war against SEO and spam from day 1, and they’re losing ground / lost.

This explanation that it was just a way to make more money implies that if Google wanted they could show you the exact results you wanted, but I don’t think that’s true anymore.

Google’s initial algorithm was Page Rank. They looked at how many sites linked to your site. The more sites that linked to it, the more people who thought your site was high quality. You couldn’t game it by anything you put on your site because the ranking came from other people’s impressions of you. But with enough money, that was easily defeated by buying up / starting up thousands of sites and promoting another site by linking to it.

Since then it has been a cat and mouse game. Google tweaks their process to reduce inauthentic “recommendations”, the SEO people change their methods and get back into the top results. For a while Google was winning, but even before AI, they were already losing more and more often.

The rise of Facebook and other social media just made it worse. Google didn’t have deep visibility into these sites, so they couldn’t carefully look over metadata to help judge whether something was authentic. Meanwhile, it became cheaper and cheaper to put up inauthentic content that attempted to boost a site’s placement on Google.

I think the pivot to AI is as much Google admitting defeat in the war against SEOs as it is about them trying to make money from a new venture.

So can I, as long as the genre is “poorly played”.

Because paying a subscription for songs you do own wouldn’t make any sense.

That’s sad too, to be honest.

Automatic curation wasn’t very good. Pandora and other tools tried to do it, and they had some success, but it wasn’t like having a friend who would say “oh, you like X, you should really listen to Y”.

Pandora and similar tools would say “oh, you listened to X, you must like music with distortion on the guitar, a lead female singer…”. But, that kind of soul-less take on music often meant the suggestions were really bad. Yes, in n-dimensional space, it was able to find music that was a short distance from other music. But, that’s not why people like what they like.

So, discovery really was (and still is) a problem that needs solving. But, the bigger issue is money.

A lot of people want to make money from music. Even musicians sometimes care about money. But, honestly, a lot of them would be making music regardless of the money, as long as they could afford to feed themselves.

The real people who care about money are the companies that own the IP for the music. They want you to listen to their music because when you do that they get paid. That means their goal isn’t for you to find something you enjoy listening to, it’s for you to select something that they make money from. That also means they want to interfere with anything that might make good suggestions for you because it might direct you away from what makes them money.

So, now on one side you have companies that want to sell you a tool that exactly meets your music desires, by believably simulating emotion, playing ability, and everything else, but one that can only interpolate from what it has been trained on, and can never innovate. On the other side are the companies that want you to choose anything you want, as long as it’s from their catalogue and they make some money from it, and if it’s not from their catalogue they’d prefer you never hear it. Everyone sucks here.

open AI has 1 billion

I don’t really believe that. Have 1 billion people used it at least once? Maybe? Maybe part of that 1 billion is the same people on different accounts. I definitely don’t think there are 1 billion regular users.

IOW, the plan is proceeding as “intended”. See: Airbnb, Uber, etc.

That’s where I disagree. Uber maybe has a viable business as long as they keep finding gullible or desperate people to drive for them. Their OpEx costs are tiny, it’s nothing more than a system that receives a message that someone wants a ride and sends out a bid to take that customer. Then it’s a pricing algorithm that drops the price that they pay to drivers as low as possible, while driving the cost of the ride as high as possible. They get whatever slim margin is left.

AirBnB is similar, virtually no OpEx costs, a very simple set of algorithms to match places for rent with people wanting to rent them, then fiddle with the prices and extract a middle-man fee. As long as they keep staying ahead of government regulations, there’s a business there.

But, nobody had to go out and explain to drivers why they might want to drive for Uber. “You drive people, we pay you money”, it’s a simple sales pitch. Same with AirBnB. Hotels and taxis were annoying enough that it was easy for consumers too. But, nobody has yet found a long-term justification for chatbots.

Sure, some companies are trying to use them. But, not successfully. Air Canada replaced customer service reps (presumably in a call centre in India or something) with a chatbot. But, when that chatbot hallucinated a policy when talking to a customer, the courts found that Air Canada had to honour that policy. That’s a massive risk long term, especially once people get good at figuring out how to talk to them.

People are using them to generate vast amounts of code… but code is a liability, not an asset. The things code can do might be an asset, but you have to be able to understand and maintain the code for it to work. If people are using AI to generate the code, and other people are using AI to review it, nobody actually understands it. This is just a time bomb waiting to blow up.

If there really were some kind of guaranteed money making thing that chatbots could do (other than online scams), the AI companies wouldn’t be out there trying to sell AI in general, they’d be buying companies and/or starting divisions that did that one thing.

IMO, the AI companies intended to hit AGI, and hand everything over to the machine god. Now they realize that it really just is a clever word-guessing machine with some vague capabilities that niche users might want (but not at the price they can meet), and they’re scrambling to find a seat before the music stops.

To put that in context, Alphabet (Google’s parent) is the biggest software company in the world. It’s the 3rd biggest company in the world after nVidia and Apple, but they’re mostly hardware companies.

Google’s annual revenues are creeping up towards $500b/year. So, the AI companies (Google included I suppose) would need 12x Google’s annual revenue just to break even.

There are 2 companies that have over $1b in annual revenue, Amazon and Wal*Mart. But, those are companies that sell goods to consumers, they’re not just digital services businesses. But, even then, you’d need 3x the revenue of Amazon and Wal*Mart combined to hit $6t.

Google, Apple, nVidia, Wal*Mart and Amazon each took decades to grow enough to capture hundreds of millions of dollars in revenue. So, somehow OpenAI and Anthropic are going to need to grow faster than any other company on earth to hit these trillion dollar targets.

Also putting this number into scale, there are about 8.3 billion people on the Earth. But, most of them are poor goat herders, subsistence farmers, sweatshop employees, and AI model trainers / reviewers living in Asia and Africa. These aren’t the kinds of people who are going to be buying a subscription to ChatGPT. There are maybe 1 billion people globally who have disposable cash they can spend on AI. So, to hit 6 trillion annually, each person living in a developed country would need to personally hit $6000 in AI spending on a yearly basis. Note, that includes children and the elderly. If you limit it to working-age people, that’s more like $10k per person per year.

Now, it doesn’t have to be that this is down to consumers. Maybe this is so valuable that Wal*Mart, the US government, and every other large employer pays for a ChatGPT subscription for each of their millions of employees. But, really, despite the fact that there have been very few actual successful AI deployments that saved companies money, or generated more profits, they’re suddenly going to be spending a significant fraction of each worker’s wage on AI subscriptions?

I think it’s pretty obvious that the AI companies thought that they were about to invent AI jesus. They thought they were just a few datacentres away from the machine god coming to life. They thought that AGI would change the world to such a degree that it might make the entire economy obsolete, so there was no limit on how much they should spend so either they would get there first (and maybe be seen as the god’s parents?) or out of some kind of altruism-type-feeling so that they could make a machine god which would align with their beliefs. But, that didn’t happen.

I saw a comment the other day about AI firms and CapEx vs OpEx. CapEx is Capital Expenditures – basically one-time costs for setting up a business. OpEx is Operational Expenses, basically ongoing costs for running a business. A good way to generate money from a business is to have a high CapEx and low OpEx. It’s hard for someone to set up a competing business because of the high initial costs, but once it’s up and running you can provide services for a low ongoing cost, and you can make a high profit on each customer. If CapEx is low but OpEx is high, sometimes it can work as a luxury business. Think an artist who hand-carves furniture. If both CapEx and OpEx are low, it’s basically a commodity business, think something like a nail salon. Cheap to set up, cheap to run, hard to fight off competitors.

The AI companies are stuck in a world where they have incredibly high CapEx (building DCs and training new models), and incredibly high OpEx (inference queries – basically talking to the chatbot, is also very expensive). Right now to try to hook customers they’re subsidizing the OpEx side of things, so each new customer costs them more money. Their paid tiers are even worse because people actually start really using the chatbots once they pay for them, and the money lost per customer goes way up for the paid tiers.

The only thing that would work is if the companies can somehow manage to get customers to pay incredibly high costs per token, so that not only are the OpEx costs covered, but the companies can start to pay down the initial CapEx. But, there’s just no justification for that yet. And the ruthless competition between all the AI companies means that if one of them starts trying to jack up costs to cover their OpEx, the others will keep theirs low to grab market share.

This is why the AI companies are doing all the things they’re doing. They’re talking about their models “escaping containment” and “hacking” other companies. Why? Because that makes it sound like these things are so powerful that they can’t be contained, rather than it being that the AI companies are incompetent and building a locked-down test environment, and how they are just running dumb scripts that they’re not even looking at before the execute. They’re boosting stories about the AIs being so powerful that they’ll destroy the world so that CEOs think they need to buy subscriptions to this powerful machine god for their companies. They’re inviting regulations and laws so that no new AI companies can enter the business, and so that they’re all forbidden from building new models, which is a massive CapEx (sorta) cost.

They can’t admit that the attempt to build a machine god has failed, because that will cause the bubble to pop. What they want to do is survive long enough to go public so that the costs they’ve spent can be foisted on the dumb retail investors – and at this point on pension funds and anybody holding index funds, because these companies are so huge that index fund managers would be obligated to buy them. So, they need to keep scaring people about how incredibly powerful and unstoppable their word-guessing-machines are until their investors have safely grabbed their profits and everything can be allowed to explode.

Guy who controls the cops and has absolute immunity for any crime he commits according to the Supreme Court commits another crime.

Also, the plumbing isn’t suitable. Office buildings have all the lines routed to a few huge bathrooms per floor, plus maybe a standard break room that’s in the same place on every floor.

An apartment needs its own bathroom, its own kitchen, etc. So, instead of the pipes going to two areas on each floor, they now need to be re-routed to go to 50 small areas on each floor. So, the entire plumbing for the building needs to be redone.

Also, modern office buildings are very big, and there are a lot of cubicles / offices / open areas that are nowhere near a window. When the place doesn’t have many walls, maybe that’s not so bad, because you can still vaguely see a window off in the distance. But, when it’s an apartment building, you can’t now have an apartment that doesn’t have windows. So, the apartments have to end up being these narrow things with windows only on one side at the end.

Big windows, but only on one narrow bit of an apartment is bad for ventilation, and it also means the place gets really hot when the sun hits those windows. When the thing was an office building the entire floor was one “climate control area”, with a limited number of walls and lots of wide corridors. So, if it got hot on one side it would even out across the floor. That doesn’t work when that floor is subdivided into 50 individual apartments. So, now you need brand new solutions to the problems of ventilation and climate control.

It’s less work than building a whole new building from scratch, but there are a lot of trade-offs.

You know that the developer is going to be spending hundreds of thousands if not millions on lawyers to draft the deal. I hope the city’s lawyers are up to the challenge of spotting all their loopholes and neutralizing them.

It’s not even ideal for the people who have rent control.

Sure, they have rent that’s a lot cheaper than it could be. But, they’re stuck. Maybe their place doesn’t have enough bedrooms. Maybe it’s not near their friends or family. Maybe it’s too big for them. Maybe they hate their neighbours. But, they have rent control, and they lose it if they move.

Meanwhile, someone else who that apartment would be perfect for has to look elsewhere.

It’s a bad system, but rent in NYC is so outrageous people will fight for the bad system because they doubt anything better is really possible.

That won’t be their first instinct. Their first instinct will be to kill anybody who seems at all threatening.

The phrase “at no cost to the city” is a red flag.

It either means someone is lying, or that there’s some kind of “public-private partnership” which are frequently handouts to corporations or rich people.

Maybe this is the rare case when the government will properly supervise the private side, and hold them to account.

The main thing is that he’s really good at publicizing what he’s doing.