AI Destroys Institutions

It's like for some reason we thought that like some good percentage of us aren't just tribal worker drones who fundamentally just want fats, sugars, salts, dopamine and seratonin. People actively vote against things like UBI, higher corporate taxes, making utilities public. People actively choose to believe misinformation because it suits their own personal tribal narratives.

Regulations do more harm than learning process from mistakes.

I had a wake up call on this yesterday. After a recent HN thread about Zed editor, I decided to give it another try, so I loaded it up, disabled AI, and tried writing some code from scratch. No AI completion, no intellisence. Two things came to mind. First, my editor seems so much more peaceful without being told what to do. Second, it was a bit scary how lost I felt. It was obvious that my own ability to communicate through code had declined a bit since I began using AI coding assistants. It turns out that as expected, coding assistants really are competitive cognitive artifacts. After that experience, I've decided that I am going to do at least part of my coding with all completions turned off. Unfortunately at work you are paid to produce quickly, so I think my AI free editor will have to be reserved for personal projects.

Further related to your statement about thought, the hallucinations persist, and even last night I got a response about 80's pop culture that was over 50% bullshit. Just imagine what intentional persuasion through LLM models will do to society. Independent thought has never been more important.

No sir, there was nothing there to begin with - if you read recent history, you'll see that it's full of stupidity, and a few rabble rousers leaving entire nations by the nose.

With the mollification off the smartphone, we've merely taken off the edge of this killing machine.

SSRN is where most draft law review/journal articles are published, which may be the source of confusion.

For most other fields, it is a source of draft/published science papers, but for law, it's pretty much any kind of article that is going to show up in a law review/journal.

I guess i'll start with calling two well known law professors "$500 an hour nitpickers" when they don't earn 500 an hour and have been professors for 15+ years (20+ in Jessica's case), so aren't earning anything close to 500 an hour, is not a great start?

I don't know if they are nitpickers, i've never taken their classes :)

Also, this is an op-ed, not a science paper. Which you'd know if you had bothered to read it at all.

You say elsewhere you didn't bother to read anything other than the abstract, because "you didn't need to", so besides being a totally uninformed opinion, complaining about something else being speculation when you are literally speculating on the contents of the paper is pretty ironic.

I also find it amazingly humorous given that Jessica's previous papers on IP has been celebrated by HN, in part because she roughly believes copyright/patents as they currently exist are all glorified BS that doesn't help anything, and has written many papers as to why :)

Did you read the paper?

You can criticise the hourly cost of lawyers all you like, and it should be a beautiful demonstration to people like you that no, "high costs means more people go into the profession and lower the costs" is not and has never been a reality. But to think that any AI could ever be efficient in a system such common law, the most batshit insane, inefficient, "rethoric matters more than logic" system is delusional.

when one of the juniors makes a mistake, i can talk to them about it and help them understand where they went wrong, if they continue to make mistakes we can change their position to something more suited for them. we can always let them go if they have too much hubris to learn.

who do we hold to account when a model makes a mistake? we’re already beginning to see, after major fuckups, companies blackhole nullrouting accountability into “not our fault, don’t look at us, ai was wrong”

the other thing is, if you have done a good job selecting your team, you’ll have people who understand their limits, who understand when to ask for help, who understand when they don’t know something. a major problem with current models is that it will always just guess or stretch toward random rather than halt.

so yes, people will make mistakes, but at least you can count on being able to mitigate for those after.

I can't see how, given that potential is 99% shorcuts.

More like a war on the traditional, human-based knowledge, leveraged by people who believe that via coveting the world's supply of RAM, SSDs, GPUs, and what not, can achieve their own monopoly on knowledge under the pretense of liberating it. Note that running your own LLM becomes impossible if you can no longer afford the hardware to run it on.

So is the AI better?

No. It's quicker, easier, more seductive.

Martin Luther used it to spread his influence extremely quickly for example. Similarly, the clergy used new innovations in book layout and writing to spread Christianity across Europe a thousand years before that.

What is weird about LLMs though, is that it isn't a simple catalyst of human labor. The printing press or the internet can be used to spread information quickly that you have previously compiled or created. These technologies both have a democratizing effect and have objectively created new opportunities.

But LLMs are to some degree parasitical to human labor. I feel like their centralizing effect is stronger than their democratizing one.

Currently companies start to shift from enhancing productivity of their employees with giving them access to LLMs, they start to offshore to lower cost countries and give the cheap labor LLMs to bypass language and quality barriers. The position isn't lost, it's just moving somewhere else.

In the field of software development this won't be a an anxiety of an elite or threat to expertise or status, but rather a direct consequence to livelihood when people won't be hired and lose access to the economy until they retrain for a different field. So a layer on top of that you can argue with authority and control, but it rather has economic factors to it that produce the anxiety.

In that sense, doesn't any knowledge work have a monopoly on knowledge? It is the entire point to have experts in fields that know the details and have the experience, so that things can be done as expected, since not many have the time nor the capabilities to get into the critical details.

If you believe there is any good will when you can centralize that knowledge to the hands of even less people, you produce the same pattern you are complaining about, especially when it comes to how businesses are tweaking their margins. It really is a force multiplier and equalizer, but a tool, that can be used in good ways or bad ways depending on how you look at it.

It's not so simple that we can say "printing press good, nobody speak ill of the printing press."

Before the printing press, only the clergy could "identity" witches but the printing press "democratized knowledge" of witch identification at larger scale.

The algorithmic version of "It ain’t what you don’t know that gets you into trouble. It’s what you know for sure that just ain’t so" is going to cause huge trouble in the short and medium term.

the church did the thinking for the peasants. the church decided what the peasants heard, etc… this is moving absolutely in that direction.

the models now do the thinking for us, the ai companies decide what we get to see, these companies decide how much we pay to access it. this is the future.

Every large enough corporate wants to become the new Oracle.

(By the way, are you confusing affordance, the UX concept, with affordability?)

ZIRP, Covid, Anti-nuclear power, immigration crisis across the west, debt enslavement of future generations to buy votes, socializing losses and privatizing gains... Nancy is a better investor than Warren.

I am not defending billionaires, the vast majority of them are grifting scum. But to put this at their feet is not the right level of analysis when the institutions themselves are actively working to undermine the populace for the benefit of those that are supposed to be stewards of said institutions.

[0] https://en.wikipedia.org/wiki/Affordance