mindcandy

I get that AI anxiety is running high. But, this is doomerism running hard.

Needle in a Haystack Simulator is the new Threes! https://www.theverge.com/22914955/threes-2048-ketchapp-copyc... A very simple game to clone that gets 10,000 headlines because creators have always been anxious about cheap clones. And, anxiety drives the clicks that pay the bills.

Meanwhile, local models are already extremely capable and slowly shrinking their gap vs. frontier models over time. And, it has been long established that the censorship trained into those models can be disabled easily with open-source tools. De-censored models will happily dig into topics that push beyond the limits of sanity, let alone political pearl-clutching.

Besides that. If you want to write about some topic and AI doesn't want to help you, you still have your fingers. Type like a human and you haven't lost anything compared to when Celeste was made.

My POV is that indicates you need to up your expectations for yourself. You apparently have the knowledge and experience to cause 80 hours worth of work to be completed in a few sentences given today’s tools.

So then, given your knowledge and experience, what challenge would require 80 hours of real, focused work out of you using today’s tools?

I get not wanting to contribute to closed, commercial models. We need to hurry up and get Folding-at-Home style distributed community model training going. At that point, I’d be pretty happy if my projects contributed to improving community models. Even if commercial models benefited too.

Good for them?

Agreed.

I've been seeing too much "I've been coding for decades and what I've always done has be reduced to telling AI to do it for me."

and not nearly enough "I've been coding for decades and recently worked hard for months with AI's help to make something I would have never considered possible before."

Much easier to complain that you are bored than to step back up to high-effort with your new fancy tools.

My understanding is that "The best reasoning they give" is the entire point. It such a huge issue that it drowns out all of the valid but smaller issues you are pointing out. And, the benefit to the water is just a bonus.

Great! That's how knowledge is supposed to work.

> I don’t like them. They talk to me in this grating LLM-voice

I find that what makes people hate LLM-voice isn’t the voice itself. It’s that the voice pings their brain with a reminder of their AI anxiety. That ping makes people instantly upset.

> They confidently bullshit me - often giving me useful, helpful answers. But also just making stuff up with the same assurance … When we think of AI agents, we shouldn’t anthropomorphize... They are (software) machines

I believe these ideas contradictory. I find that people get angry at LLMs bullshitting precisely because they are not anthropomorphizing. People bullshit all the time. But, these are machines! Machines are not supposed to bullshit! And, when they do, there’s supposed to be a bug filed and fixed.

By holding LLMs up to the expectations of machines instead of expectations of (simulated) people, users are frustrating themselves.

If you take the same “trust but verify (everything)” approach with LLMs as you do with people, you’ll be a lot less frustrated.

Micropayments would solve so many problems for the internet. And, cryptocurrencies would solve so many problems for micropayments. But, it's a non-starter because any proposal gets flooded with people popping veins about how crypto can't solve anything.

Check it out: As millions of people have recognized, we have a new discoverability problem! How about, out of the 9 billion people on Earth, we figure out how to get a few dozen of them motivated to work on that instead of sitting and moping for the rest of our lives about the end of human creativity?

They are ridiculous because the reality of the situation is boring. Boring doesn’t drive engagement. So, all the clickbait headlines and ragebait comments imply scandal. If they didn’t, they wouldn’t get attention.

After being clickbaited and ragebaited, media consumers feel deeply anxious and angry. But, explaining that they are angry over a boring situation feels silly, not righteous. So, they give summaries, impressions, sometimes extrapolation of the bait they have been consuming. That feels righteous.

This observation applies to a wide variety of topics trending in the various media every day. Distinguishing injustice from ragebait unfortunately requires non-trivial effort from the reader.

I’m trying to imaging what giving credit would look like… “Credit: Everyone who has made anything freely observable on the internet.” Would folks be less upset if that was included?

Frankly, I should have been putting something similar attached to any drawings or paintings I make physically. My entire life I’ve been training my brain on countless works of others. And, I’ve never asked one of them for permission or given anyone any credit.

Lots of wild guesses about mixed human-AI copyright in here. Last time I read the copyright office’s statements in this, their stance was basically: The human owns exactly what they contribute. The rest is public domain.

Yes, that is vague. I think the examples were like:

If you paint a symbol and use an AI filter over that to stylize it, you own the symbol aspect of the image but not the stylized final result.

You can own a book of AI images as a curated collection. But, not the individual images.

You are reacting to the author reacting to Zuck reacting to Zuck's daughter mentioning she uses AI to help plan recipes. It's a one-sentence, third-hand anecdote in a large article that folks in here are fixating on because of bike shedding.

Maybe Zuck's daughter appreciates cooking differently than you do. If AI helps her enjoy the part she likes about cooking by getting her past whatever friction would make her less likely to cook otherwise, then good for her.

Sure, it's possible. But, rent control is a serious disincentive to housing development. It requires correspondingly serious investment from the politicians to overcome that. In practice, what is done is to disincentivize for short term gain, not invest in long term gain, and complain that it's going terribly in the long term.

Exactly. The problem I don't know how to fix is that voting for rent control is short term thinking that ends up causing rental pressure to become much worse in the long run. There are few things more difficult than convincing large numbers of people to think long term.

So, here we are. Living today in the long-term consequences of decades of voting for more rent control while opposing urban real estate development. And, the solution everyone is calling for is more rent control while opposing urban real estate development.

Economists don't say no one would ever need to move without rent control. They do say the situation is a whole lot worse for all of the other folks in your neighborhood because you have rent control. And, in the long run, rent control is likely to be detrimental even to you in ways that take more than a single sentence to understand.

I didn't make up 98%. I got it from the video who got it from https://kentclarkcenter.org/surveys/rent-control/ Reading more closely, I stated it wrong. 2% agree rent control has helped. 81% explicitly do not, then there are the "other" categories.

Does anyone have any numbers on what percentage of recent datacenters don't follow those rules? Those are completely reasonable requirements. But, the high-pitched online discussion tends to imply building a datacenter requires doing everything the worst possible way every time.

I know there are problematic data centers. I just want to know what the stats are vs. the rhetoric.

I don't think outside intervention is necessary.

When I see stories about how a hardware startup like SpinLaunch has had to move over and over because no matter where they go, locals flood the town hall meetings to declare their relatively boring technology is too scary, I'm not surprised town halls are being flooded over AI datacenters.

[1] https://www.youtube.com/watch?v=kGxmCvLb9bs

Are you talking about the Box Elder County Stratos project? I did some digging into that. AFAIK, they are mostly going to be powered by a natural gas pipeline. I haven't heard of them using nuclear. Got any links?

It is in a remote valley. And, it's going to use less groundwater than the previous use of the land. The only things to really complain about are huge amounts of gas consumption and AI bad. If the locals pinned them down to building out solar and mini-nuke power, the only thing left would be AI bad.

Maybe my feed is biased. But, I’ve been seeing good arguments lately that “The rent is too damned high!” is largely a problem created by voters. And, I don’t know how we’re going to fix that.

Ex: https://youtu.be/yoT76zITRx8 TLDR: 60% of both left and right wing voters like rent control. 98% of economists oppose it. Even liberal lefty economists. Even though they support price controls elsewhere like utility monopolies. Rent control feels good in the short term, but is detrimental to the vast majority in the long term in a large variety of ways.

The actual solution, to the surprise of no one, is to increase supply.

Unfortunately, there was a paper I don’t have a link to showing that not just old NIMBY homeowners, but largely everyone consistently vote against measures that increase supply. It’s part of a larger effect where people measurably bias their voting against policies that sound like they will benefit rich people. Very similar policies in similar areas will get different votes depending on how the wording makes it clear the rich will benefit.

Well guess who almost always benefits immediately from real estate construction? So… we keep voting to prevent construction and then complaining that the rent keeps going up.

I find what people want from the data center build-out is pretty reasonable:

1. Don't build loud machines across the street from someone's house. 2. Don't drive up water/electricity bills by pulling from the municipal grid. 3. Don't burn the earth with CO2. 4. Don't facilitate techno-fascist surveillance. 5. Pay a lot of taxes.

But, when I see people complain about data centers, I usually see a spiral of rage-bait based on mis-framing, conjecture and repeating stuff others simply made up to get everyone mad. The online discussion tone is reaching a shrieking pink-froth level.

What I'm finding in my lazy research on the subject is that recent and planned datacenters

1. Are exceedingly rarely built where anyone can hear them. But, occasionally are. Thus ragebait articles. 2. Pull from muni grids for small scale centers. But, the hyperscalers are independent. 3. Hyperscalers are burning way too much natural gas. That's bad. But, not using nearly as much water as people think. That's good. 4. Techno-fascist surveillance doesn't need 1000s of datacenters. Just a handful off in the corner is plenty. 5. Datacenters usually are paying a lot of taxes. I think this drives the majority of the push from politicians. And, there's a minority from corruption.

Frankly, the biggest issue folks should be pushing for is to require hyperscalers build out solar power instead of natural gas. That will be a hard fight because it requires more up-front time and investment. From there, ensuring small scale centers don't affect people's bills is a good fight. But, passing laws to ban datacenters seems performative compared to passing laws to regulate their power and water impact.

Good points all around.

Hey! Thanks for the fun, good-faith argument! Such a rarity these days that most folks seem unaware are that arguments even can be collaborative.

I must disagree.

One could argue that a human body with a mostly-functional brain that will never again be conscious is human but not a person. And, simultaneously argue that a few million human stem cells, none of which are neurons yet, is a personified human. But, that would be a clearly inconsistent stance.

And, I’d argue that consensus literally means that a significant population agrees.

Formally and objectively it could be argued that even shortly post-birth, barely-conscious infants lack the required characteristics of an individual person. But, obviously it would require an unacceptable degree of callousness to terminate an infant. Certainly not a practice we agree should be encoded into law in this modern era.

In general, I find the pro-choice consensus to be reasonably conservative regarding the rights of an embryo vs. the rights of a fully-realized mid-life person. Strongly weighted towards the mother before 24 weeks and strongly weighted towards the embryo afterwards. But, if it comes down to the life of the fully-realized person vs. the life of the potential person, the choice goes to the mother. Or, if the potential is clearly a life of suffering and early death, the ethical choice is to reduce suffering via difficult actions.

Generally, I find that the pro-life arguments are largely based on a lack of consideration. I.e. “Life begins at the instant of insemination.” with literally zero further consideration. Or, on knowingly false propaganda: “Evil baby-killers are celebrating killing babies minutes before birth.”

To my knowledge, neither of those stances were in the mainstream before American politicians accepted that the Civil Rights Movement was not going away.

Before that, The Church saw its role in the abortion issue as charitably offering grief counseling to women who had experienced that difficult choice. After the political situation changed, the role of The Church changed to the demonization of unfortunate women for the purpose of manufacturing single-issue voters for the benefit of specific politicians.

A human is born, lives a life, ingests the wrong chemical and becomes brain-dead. Now you have a human that is alive, but the person who was that human is permanently gone. To avoid murder, are we obliged to keep the body alive as long as possible by all available means? I don’t expect you would think so. I expect you’d agree terminating the life of the human body is the preferable and not a great ethical quandary.

I would agree that at conception we have a human life. But, not that we have a human person. There are no thoughts, emotions, memories, hopes, dreams. There is a very strong potential for a person. But, at conception what makes a person is not there yet.

At some point in time between the human existing a single living cell and the human voting in elections and joining the military, a person develops. But, it’s not a definite transition point in time. It’s a gradient.

Unfortunately, for practical reasons, the law must pick a definite point in time. The consensus seems to be “around the time that the fetus can be kept alive and continue developing independent from the mother using current technology.” A practical choice, for sure.

One might argue that the official transition point should be a bit earlier. But, that’s not the core of the disagreement.

The core seems to be that some living people do not appear to take into consideration the difference between being alive vs. being a person.

With datacenters there is a huge amount of ragebait in the news and social media. People are anxious about jobs and reporters and influencers are cashing in on that anxiety by running headlines that imply doom. Datacenters are hugely problematic. But, everywhere I go I see media and commenters exaggerating, implying, assuming, or just making up outrageous claims. But, it's the hot topic right now and outrage drives the clicks that pays the bills. So, ragebait it is.

We solar, it's clear to me that Exon and friends have been funding anti-green-energy politicians and propaganda for decades. They knew full well that greenhouse gases would wreck the world. And, they new they could have transitioned to actually being "Beyond Petroleum". But, the profit margins would not have been as high. So, instead they are spending billions to convince everyone to let them keep wrecking the world for a few more decades.

A moratorium seems excessive when the requirements to make a majority of people happy are pretty clear.

The easy ones are:

1. Use energy and water independent of the municipal grid. 2. Don't build so close to homes that they hear the fans.

A lot of people are rightly concerned about water. That was a big problem with older evaporative cooling systems. But, the newer closed-loops systems being built now are much less worrying.

Water usage concerns can be addressed by doing what https://www.boxelderstratos.com/ did. That's an extremely controversial site because it's 40,000 acres. How could it be so huge? Because that's how much land they had to buy to be a net reduction in water table use compared to the previous owners. It's not a warehouse the size of San Francisco!

A lot of people are rightly concerned about their electricity bill. But, what I'm seeing is that new large datacenters are opting to generate energy on-site using huge amounts of natural gas. That's obviously also very bad because of CO2. But, it doesn't affect your electricity bill nearly as much.

Using the Stratos project as the example again, there's a lot of outrage about how it will "Use 9GW! That's more than double Utah grid capacity all by itself!" But, it's not on the grid. It's starting with 1.5GW of natural gas from a pre-existing pipeline. That's still hugely problematic. Just not in the way the ragebait implies. They claim they are going to build up a mix of gas and solar after launch. But, we'll see how that goes...

If there was regulation requiring them to build up with green energy, regulation with teeth, Stratos would not be nearly so controversial. The water is covered, the energy is off the grid, it's far from any residents, the only concern left is the tremendous amounts of CO2 to be produced. To be clear, Utah already has had several non-datacenter sites producing that much CO2 for a long time now. But, it's not great to add another one.

My point is that if New York put in place a list of reasonable requirements for building and actually enforced them, then a supermajority of residents would be OK with datacenter build-outs. Of course, a minority of folks are so enraged that they want to see all datacenters burned down. But, you can't please everyone.

Lots of folks in here calling out the "billion scans a month" and skipping over the "capable of" part.

They're not claiming they'll perform a billion scans. They're trying to build enough machines that if absolutely all of them were run at 100.00% capacity it would be theoretically possible to do a billion scans a month.

AFAICT, the "Rising layoffs year after year" news lately has been an intentional misreporting of "Layoff rates slowly returning to normal after the hiring spree during COVID stimulation"

https://blog.glassdoor.com/site-us/wp-content/uploads/sites/...

https://www.glassdoor.com/blog/worklife-trends-2026/

But, CEOs figured out that if they blame layoffs on AI, stock go up a lot. Reporters know that anxiety about AI drives the clicks that write the checks.

I do think that AI anxiety is making HR around the world anxious about hiring. That's my best guess for why everyone is finding they need to apply to 500 jobs to get anywhere. So, AI is making it hard for you to find a job not because it took your job, but because HR is reading ragebait and turning it into a self-fulfilling prophecy.

> AI takes in all the input, whether the original authors have consented or not, and do some "learning"

What would it mean for authors who publish content publicly to the web, without access restrictions, to provide consent for learning from it?

"EULA: Most people are allowed to learn from this text. If you work in an AI-related field, even though you can clearly see this page because you are reading this text right now, you are not permitted to learn anything from it. Bob Stanton, you are an a-hole. I do not consent to you learning from this web page. Dave Simmons, you are annoying. But, I'll give you a pass. For now... Also: plumbers. I do not like plumbers for reasons I will not elaborate. No plumbers may learn from my writing in an way."