You’re right. They’re not LLMs and they’re not particularly new.

The main new part is that new techniques in AI and better hardware means that we can get better answers than we used to be able to get. Many people also realize that there’s a lot of potential to develop systems that are much better at answering those questions.

So when people ask, “Why are companies investing in AI when customers hate AI.” Part of the answer is that they’re investing in something different than what most people think of when they hear “AI”.

A lot of people have come to realize that LLMs and generative AI aren’t what they thought it was. They’re not electric brains that are reasonable replacements for humans. They get really annoyed at the idea of a company trying to do that.

Some companies are just dumb and want to do it anyway because they misread their customers.

Some companies know their customer hate it but their research shows that they’ll still make more money doing it.

Many people that are actually working with AI realize that AI is great for a much larger set of problems. Many of those problems are worth a ton of money; (eg. monitoring biometric data to predict health risks earlier, natural disaster prediction and fraud detection).

I’ve been called many names, including “tankie”, so I’ll take a stab at responding.

I’m not mad about the debate at all. I expected something fairly similar. I’m mad that Biden and the Democratic leadership seems to have put their own interests above the interest of the party.

If Biden had gracefully stepped aside and given just about any other Democrat his full support, we’d be in a much better position now. Instead we have a candidate with a ton of baggage and who presents an easy target for Trump’s style of argument. Many mainstream Democrats, including the NYT, are finally starting to realize this. Unfortunately it’s probably a year too late. At this point it would just make it look like Demoratic kingmakers forced him out.

If I went by the modern definition of “tankie” as, an ant-american authoritarian communist. I probably wouldn’t be mad at any of this. I’d be cackling with glee because either of the current nominees will be terrible for the US. Neither of them has a serious long term plan. Neither of them can articulate a policy position. Both of them will continue to erode the power and moral authority of the United States.

Like it or not. Trump is likely to be the next president https://projects.fivethirtyeight.com/polls/. At this point it’s probably wise to start thinking about how to limit his impact and how to start cleaning up the mess afterwards.

It’s one thing to refrain from commenting but supporting Israel makes it clear that Germany learned nothing.

What was the post?

It’s functionally close enough to a conglomerate though.

I’m not exactly sure what ‘“free market” cultist’ is or if you’re accusing me of being one. Modern economists don’t normally align themselves with simplistic ideologies like “free market”, “communist” or “capitalist”. They’re aware of the historical and modern usage of these terms but they tend to focus on areas that are far to specific for those terms to even make sense. You won’t find a lot of economists that argue for complete Laissez-faire capitalism any more than you’ll find real economists arguing in favor of classical Marxism.

There is general agreement that conglomeration benefits management more than shareholders. There’s general agreement that they are more likely to arise under some economic conditions and that those conditions usually aren’t associated with socially optimal economic policies.

It’s a bit complicated.

The CEO and the other shareholders aren’t the same.

For the CEO, it’s a good way to diversify since they can’t diversify the normal way.

For the regular share holders it’s a way to diversify but it’s not as good as being able to buy and sell the individual components.

I’ll skip a lot of the math but the upshot is that their Sharpe Ratio (expected return divided by risk) is higher if they do their own diversification than if they buy one company that tries to diversify within it.

These weird combinations look fun but they’re generally the result of having conglomerates, companies that have gobbled up a bunch of smaller, unrelated companies.

Conglomerates are tricky to pull off because managing a lot of disparate business lines. A CEO who knows all about how to market construction equipment is likely to miss that one of their other products became an iconic sex toy years ago. The big problem is that more focused companies can typically outmaneuver you in their area of focus.

Theoretically, there might be synergies that make your company more effective but normally, conglomeration is drag on the risk-adjusted rate of return on your company. It’s much easier to pull off when your government has strong protectionist policies or if there are officials you can bribe to keep out the competition.

Why would a company do something that’s generally bad for the company? It’s generally good for the CEO. A CEO often has a very concentrated investment portfolio. Changes in the value of the company they’re running can have a huge impact on their personal wealth. Conglomeration allows a single company to be a diversified asset. It does it in a way that’s objectively worse for shareholders but better for the CEO.

I think you’re sort of right but it will depend heavily on how radical a shift the new technology is. In order for there to be this kind of divide there needs to be a steep learning curve to the technology. People are only willing to put up with those learning curves if there’s a significant advantage. That means that manufacturers can only successfully market “difficult” technologies if they provide a big advantage.

I’m not aware of any old people having difficulty transitioning from quills to, fountain pens to ball point pens. They all basically did the same thing and you only had to make minor adjustments. Nobody bothered learning how to use the Writer since it didn’t actually let you do anything better. They were willing to go through the significant curve of learning how to use typewriters because, once they did, they could write significantly faster.

Computers and cell phones are a whole different way of interacting with people and information than “hardcopy” was. You didn’t just swap some objects that did the same thing with a different approach. It wasn’t even just a slightly different way of doing the same thing. Those technologies allowed us to interact with the world in a totally new way. It was worth learning a bunch of weird computer stuff that older generations had never heard of because we could do things they never dreamed of. (eg I used to get rushed when talking with my grandmother to save on long distance bills, now I don’t even think about long distance costs other than latency.

I’m sure that sort of thing will happen again but it would require a far more disruptive technology than AR. That’s a small iteration that we’ve already been primed for. When Terminator 1 came out, nobody was confused when it switched to “terminator vision” and you saw the AR display. That’s why I joke about neural interfaces. In theory, that could give a person significantly higher throughput rates to their computer. There are all kinds of potential benefits to. It would be worth it for people to put up with steep learning curves, unintuitive interfaces and lots of troubleshooting if it meant they could suddenly “read” at 10,000 words a minute or control complex robots. Not everyone would go through that effort and it would create the kinds of divides that we saw with computers.

When I look at current technologies as an old(ish) person, it’s a very different view than my parents and grandparents had. They didn’t understand the new technologies. I have no trouble understanding them, I just think a lot of them are a waste of my time (unlike screwing around on Lemmy, which is totally productive /s).

My wife and I regularly joke that one day we’ll harass our kids to help us with our neural interfaces but I don’t think that sort of thing will happen any time soon.

When I was a kid in the 80’s a lot of people could already afford computers. They weren’t so cheap that everyone had them but they were affordable to a fair number of people if they really wanted one. A C64 cost $595 at launch, that’s under $2,000 in today’s dollars.

The biggest barrier to computers were that they weren’t “user friendly”. If you wanted to play a simple video game you needed to know some basic command line instructions. When I wanted to set up my first mouse for my 8086 it involved installing drivers and editing config.sys and autoexec.bat. You couldn’t really do anything with a computer those days unless you were willing to nerd out.

At the same time, nerding out on a computer could easily get you deep into the guts of your computer in a functional way. I learned that the only way I could play video games at night was if I opened up the computer and disconnected the speaker wire so it wouldn’t alert my parents. I also learned that I could “hack” Bards Tale by opening up the main file with debug and editing it so that the store would sell an infinite number of “Crystal Swords”.

Today there are 2 cell phones for every human on earth. Kids walk around with supercomputers in their pockets. But they’ve become so “user friendly” that you barely even need to be literate to operate one. That’s generally a good thing but it removes an incentive to figuring out how the stuff works. Most people only bother with that if they’re having some trouble getting it working in the first place.

At the same time it’s gotten much harder to make changes to your computer. The first Apple was a pile of circuits you needed to solder together. You can’t even remove the battery on a modern one (without jumping through a lot of hoops). If you edit some of your games it’s more likely to trigger some piracy or cheat protection than to let you actually change it.

There are still large communities of computer nerds but your average person today basically treats computers like magic boxes.

I’d expect that kind of gap in other areas. I’d take 3d printing as an example. You can get one now for a few hundred bucks. They’re already used in industry but, at this point, they’re still very fiddly. The people who have them at home are comfortable doing stuff like troubleshooting, flashing ROMs, wading through bad documentation and even printing custom upgrades for their printer.