AI may be on a trajectory to surpass human intelligence; we should be prepared | Hacker News Reader

AI may be on a trajectory to surpass human intelligence; we should be prepared

imf.org

24 pointsdanboarder109 comments

Then we should start planning now by stopping the notion that being a good, moral person includes earning a wage.

There are big political issues with achieving UBI. One of them is reversing that notion. I grew up in a poor family, one time, when my Mum was laid off I heard her lament that she didn't want to take unemployment because she wasn't one of 'those scroungers.'

I see unemployment from AI, if it happens, to be a slow burn. If we don't shift our attitudes towards it then we're going to be entering a new era of "Welfare queens" but without any meaningful work for people to actually do.

That's not that big a worry. People strive and aspire no matter what their conditions. We've got loads of wildly unsatisfied billionaires. There's no meaningful difference between them and 'welfare queens' at the most basic level.

In the absence of want, people will invent want. Supply all the conditions of a totally unexamined life and people will look to status, or acquisition, as if it is life. People will manufacture what they call meaning within the context of their frame of reference.

Sure, but that's not quite what I mean. I'm not necessarily worried about a life without meaning, or want. Hell, in the absolutely best case AI enables that utopian world of allowing people to focus on their passions and achieve their best selves.

But that's not the world we currently live in.

I'm worried the job losses will be slow, and that many people will suffer because we have our antiquated notions of the world. That we will let people starve, because it's their own fault for not working (in this new world, devoid of work.) And that those attitudes will persist because those at the top will be some of the last to be fully effected.

There are trillions of tasks / jobs that are still to be done. Even to prepare AI / Robots to do those jobs takes lots of additional human 'work'.

I'll make an open bet. You'll need more humans and we will have more jobs in the next 10-20 years that we will not find enough people.

Plus Covid stimulus already showed that it's easy for the Government to print money and give it to people. It's not unpopular at all when it comes to existential risk

Every day is an existential risk for some of us.

Why pay tax when the gov can print the money instead of taking it from me?

Printing money is a tax on you, that's what inflation is - and you don't even get to grade things via tax brackets.

It's not really, though. We can be 74 zelenskys in debt, and no one really cares. We can always find more zelensky or manufacture more from thin air, it's really no problem for us.

> We can be 74 zelenskys in debt, and no one really cares

Except for basic demand and supply. The more of those "zelenskys" you have around, the less they're worth.

Military leverage. Deaths-per-dollar is functionally converted to zelenskys.

People do care when inflation goes up! Everyone complains when food and gas and rent costs more. The issue with printing money isn't being 74 zelenskys in debt, it's having massive inflation. Venezuela can print as much money as it wants, but they're in hyperinflation and it's terrible.

You have a fundamental misunderstanding of inflation.

E.g Let's say you need 1 WikiToken to read a Wikipedia article. However the Wikimedia prints 10 Billion Tokens and distributes it to entire world for free. Next year it prints 20 Billion more and so on Will the cost of reading a Wiki article go up? No. Because we have excess capacity for serving Wikipedia articles

Every cost in this world boils down to labor and if AI/Robots take over then cost of labor is 0 or in other words we have excess capacity and we can print in abandon.

You can print money when there is more capacity than demand.

When we reach AGI/ASI we may have it. Until then we don't print money and we tax.

When it comes to printing money, where is the capacity limit?

Is it more than 34 trillion in debt, or is there no real limit when you can just print it and add as many zelenskys as you like?

Money represents effort, jobs, goods. It is only as good as what it represents. You ask for money, with nothing in return.

I ask you to work for me, for free, forever. I need a chef.

When can YOU start?

There is no capacity limit.

Thought experiment. let's say you need 1 WikiToken to read a Wikipedia article. How much tokens can WikiMedia print and give it away for free?. The answer is nearly infinite, because even if WikiMedia gives away 1 Million WikiTokens to everyone in the planet, they can still serve all the Wikipedia articles without 'inflation'. They can also print a Billion or a Trillion Token with no effect on the price of reading a Wiki Article which is still 1 Token

Same Deal. If AI/Robots can turn the world into a near infinite capacity because cost of everything boils down to labor except physical space that you occupy (and everyone can only occupy so much space at a given time)

We should abandon the notion of “wages” entirely

We could arm wrestle for the produce you and I need but is in short supply.

I dig holes for a living, so I am confident I will win against you.

With a well managed planned agriculture and distribution system why should we fight over produce? There would be enough produce for us both. Our current system selects for how best to maximize profit from produce sales, not how best to serve the most people the most nutritious produce.

We shouldn't.

I have ~45 acres of green grass, you have around 50 pregnant ewes, we should both do well and I love lamb rogan josh and seeing them all through spring/summer.

Go wrestle with a machine if you’re into that. “The produce” is not what is in short supply.

How is UBI funded? Is it significant taxes on corporate entities? And will adopting UBI be super challenging for a consumer economy like the US that depends so much on discretionary spending?

I'm by no means an economist nor do I pretend I have the answers to the questions.

But my initial view of UBI is to ensure that people can afford their basic necessities.

I've heard plenty of ways to which UBI could be funded - including taxation against the robots and AIs that take on the labour that used to be done. Is that practical in the short term? Probably not, but the whole article is about thinking about and planning for these future scenarios.

I don't know what the answer to UBI funding is, but that's sorta the problem in the first place. We haven't planned enough for that world.

How is UBI funded?

Through taxes of course. And that means the tax rate will have to be something like the reciprocal of the velocity of money. But when people are handed the money for nothing, and then buy direct from a huge vertically integrated supplier it will have to be taxed 100 percent. The Laffer curve becomes relevant here.

I've never seen even a simple algebraic model of how UBI can possibly work.

In my mind (and possibly in the real world?) it works like this:

A company makes a 10% profit after paying huge labor costs, and we tax that profit to pay for our collective needs. The company all but eliminates labor costs with AI, and profits soar to 90%. We tax that profit to pay for our increased collective needs.

Funding a ubi wouldn't be that hard if we only cared about the kinds of things that get cheaper with time. Right now our federal tax rate is enough to cover 10k/year/pp (though nothing else), but if actually taxed the wealthy as much as we tax the middle class (for example, if we had a higher capital gains tax on public share sales, reducing the amount earned by the class of people who produce no real value and just earn money from stock growth), we could bring in enough to give out 5k/year/pp pretty easily. If everything that matters in life were getting cheaper the way computers and food did over the last half century, 5k/year/pp would be a good start, and as we got richer we could afford more.

But some sectors of the economy get more expensive with time, and those are what are killing us now and what prevents the UBI from being a good idea (before we fix them). Housing, Healthcare, education and childcare absorb more money every year and faster than the economy grows. Give people more money and those sectors just find ways to charge more. If we can't fix the way they do that, we get poorer in reality no matter how much more money we earn.

I realize this isn't how a lot of folks view things. To see it through my lens, imagine for a moment how wealthy we'd be if housing cost less than it did fifty years ago. If healthcare did. If education did. That's the world Keynes imagined, where we could be working 15 hour weeks (probably actually 25 hour weeks, but still) and still feel rich.

> but if actually taxed the wealthy as much as we tax the middle class

The top 10% of earners paid 74% of all income taxes. The top 25% paid 89%.

85% of all fed tax revenue is either income tax, corporate income tax, or SS taxes (on income).

What's your point? I think you're trying to argue that we tax the wealthy as much as the middle class, but this claim doesn't support that. The 10th percentile of earnings in the US is 190k. The 5th is 290k. The 10-5th percentile is reasonable to class as "upper middle class" in many high cost locales.

5th percentile plus might be a reasonable proxy for wealthy, but even if they pay half of income tax receipts, that's not proof that we tax their income as much as we tax the middle class's. Income for the top 5% is on a power law curve. The top 1% starts at 850k - if they contribute twice as much as someone at the top 5, (290k), their percent is still lower.

Meanwhile, capital gains are taxed lower than ordinary income, and the wealthy make most of their income from capital gains. And the most exotic loopholes for reducing taxes on wealth are most available for the ultra wealthy.

If someone earning 5m/year pays 1m/year in tax, they should fire their accountant. And also they're paying less in tax than most upper middle class folks by percentage.

If AI unquestionably surpasses human intellect, then certainly any idea of a meritocracy goes right out the door.

If a business owner claims they succeeded by being smart and working hard, well, any common AI works smarter, and manual laborers work harder.

> Then we should start planning now by stopping the notion that being a good, moral person includes earning a wage.

> There are big political issues with achieving UBI. One of them is reversing that notion.

So, in other words, we're fucked. I don't think an economy where the majority people are economically unproductive and lack an ownership claim will be survivable for those people, and that seems like what UBI would be.

I think the trajectory will be something like one of these:

1. We keep our current economic system and ideology, maybe tack on UBI, and the unemployed are eventually slowly (or quickly) escorted out of existence. At some point the people who still contribute to the economy will get tired of paying so much to support the unemployed, and will use their de-facto monopoly on economic power to stop. The method most compatible with our prevalent moral frameworks is suppressing reproduction of people supported by UBI.

2. Some kind of outright socialism/communism. The idea of individual ownership is abolished or severely curtailed, and all capital is managed by some centrally-planned authority a-la the Soviet Union. Basically, you have no option except live on UBI. There's still some chance of the unemployed being escorted out of existence, if the political authority can be hijacked by an authoritarian who wants to get rid of them.

3. There's some kind of luddite ban on "too much" automation, or an Amish-like control and suppression of technology based on its adverse cultural impacts.

The charts show wages plummeting if "AGI" becomes real in 5 or 20 years while "output" skyrockets. But I'm very curious what economic "output" the AGI will be producing that is so valuable in a world where no one has a job. All that economic output needs a consumer. But who's consuming a

Ultimately, this analysis is incredibly shallow. There's no indication the author has thought very hard about what types of tasks are even feasible for "AI" to do. Will AI repair leaky plumbing? Will it start new businesses? Can it pump my gas? Can it teach ballet to children?

It's easy to forget all the parts of a job that go along with the core part of the job. Could AI deliver packages? Well, it's questionable that AI can successfully drive on city streets at scale, but if AI can drive the delivery van to my house, can it load the boxes into the delivery van? Can it bring the package to my door? Can it handle a mechanical failure or a flat tire? Can it refill its own fuel? How will it handle it when packages are stolen from it prior to delivery? Call the AI cops?

And to what end? Who's even deciding to make and sell products? Does the AI decide on its own to operate a business? Why would it? How would AI direct its attention and activity appropriately? Where would AI get the resources to do this? Why would it even try?

Most innovations to make things easier/more automated have caused humanity to work harder and longer. It would be quite humane if AI allows us to finally do the opposite but I don’t think that will be the case; The new level of productivity is often just the new baseline that competitors have to meet.

Someone still has to run the AI. It will be a while before AI can fully reproduce and improve itself, especially throughout the whole supply chain, which means humans still need to be in the loop for that. “Smarter” or not, AI will be symbiotic with humanity for a very long time to come.

How are we going to order our society when none of us can compete with machines? It used to be that everyone could contribute something- even grandma could churn butter and dullards could pull a plow. But technology has made it so that only a few people can really contribute anymore. Most people's labor just isn't that valuable anymore, relative to the machines. We have the tech to live to 100 but no clear way, or reason, to afford it. It all points to a coming crisis point when we must radically alter the social contract.

I can imagine an AGI that might appreciate the beauty of self-replicating intelligence designed only by the pressures of the natural world, in the same way that we might appreciate the occasionally brutal beauty of a healthy forest or an ant hill or something like that.

I could also imagine the extreme other end of an AGI that in built on something analogous to a modern LLM, where it's views of the world are so intrinsically human that it is fascinated with all aspects of the human experience and keeps us around like pets.

In short, I think the only reasonable outcome is that we become something somewhere in between pets and residents of a wilderness preserve.

In neither situation are we members of something like a society we might recognize today.

Maybe the machines can learn the value of human life, since it seems to have escaped us. Wasn’t that the idea behind “Terminator 2”?

Chart 2 (Ouputs and Wages) illustrates what I have been saying all along: AI is the #1 reason we need universal basic income. Let the AI have my job. But it's got no need for my income, so I keep that!

An AI could need to buy things - electricity, if nothing else.

Charge 'em for data like they're Canadian.

Predictions of doom and gloom because some technology (AI or something else) renders humans unneeded forget that humans aren't just workers but also consumers. So if we get rid of a large proportion of workers without replacing work as a means of wealth distribution the economy will no longer work because there will be no one to buy the goods and services produced.

Ultimately we will either have to forgo technological advancement in order to keep the current model where work is the primary means of wealth distribution or move to a completely different system for that. If machines can do our work to let us do more interesting things for us I say great.

The difficulty is the transition period where replacing people by AI without compensation will be advantageous for the first movers but not yet enough to break the system to force us to a new model.

I wonder at what point there will be roving mobs of people destroying every device in sight. Anti-AI vigilante groups, etc.

Close than we expect probably

I would expect this will happen a short time after the machines are physically humanized (to some extent) and common enough to find in most places.

Seems every generation have a new thing to fear. Eventually it subsides into background, the heralds of doom then move on to next big thing, and the real dangers are quietly and thanklessly mitigated by more competent people.

We had this pattern with the following ones:

- Nuclear war

- Peak Oil/Resource Depletion

- Ozone hole

- Y2K Problem

Seems like "Climate Change" is on the way out and "AI Doom" is on the way in.

Climate Change is on it's way out?

Did you try stepping outside this summer? Because it was the hottest year on record, ever.

To add to that, the Ozone hole and Y2K problems didn't just disappear. We solved them with an immense amount of time and active effort.

and with thanks to public support for solutions to those problems

(to be fair, the Y2K would have been solved regardless since that could have impacted immediate quarterly financial reports)

it’s a pity it doesn’t warm everywhere equally. Here in some places the nordics summers are getting colder and wetter. We don’t get to enjoy warming but would pay for it anyway.

Famously, the nuclear threat just solved itself (no it didn’t, and in fact isn’t even solved or meaningfully mitigated)

We now have infinite oil (thank you Petro-Jesus for making it so)

The ozone hole just healed itself with no action required by humans

Pretty sure we stopped using the chemicals that made the hole in the ozone layer… and then it healed by itself.

...Which required legislative efforts right?

Like, we didn't just all decide to stop. We identified the issues, and made steps towards stopping it.

If there was no bans they'd still use it.

Yeah that was what I meant! It wasn’t a miracle that happened all by itself. It proves exactly that opposite that humans did it though massive cooperation after listening to scientists.

Correct, none of these problems solved themselves. They have been addressed only through huge concerted efforts.

Here's hoping nuclear war IS indeed 'out'. Peak Oil sure as heck ain't, and even if you forget about climate change it isn't forgetting about you.

It would be a shame if (culturally, as is your framing here) it is on the way out and AI Doom is on the way in… because there's nothing inside the AI Doom box other than the ability to displace human labor. Any hard drive already surpasses human memory for scope and accuracy. We have to get used to machines out-scoping us on technical grounds. We can't run as fast as cars, either, but we're still around to think about and sometimes even attempt 'walkable cities'.

Reality always gets a vote. With 'AI Doom' the vote isn't what you think it is. It is a Jacquard loom of words. Some things change hugely, and others aren't even touched.

> and the real dangers are quietly and thanklessly mitigated by more competent people.

With all your examples, the mitigations were not quiet at all.

- nuclear war, well, that one got resolved by an entire empire outright collapsing after engaging in a decades long financial war with both the USSR and the US spending hundreds of billions of dollars every year.

- Peak Oil was ... mitigated by making an example out of Iraq and Libya: don't fuck with our oil, or else. On top of that, governments (most notably the EU) instituted strict emissions controls, much to the chagrin of the car establishment.

- The ozone hole was mitigated by globally banning CFC gases, again much to the chagrin of the refrigerant industry.

- the Y2K problem was mitigated by investing 300 billion dollars in preparation efforts (updating and especially testing), and a further 308 billion dollars in remediation costs worldwide [2].

> Seems like "Climate Change" is on the way out

It's certainly not. The "war for attention" may be (as it's being replaced by the Russian War and now the I/P clusterfuck), but this year was among the hottest on historical record, and natural disaster severity and frequency keeps increasing.

Unfortunately, this time we can't trust that those in the background will continue to mitigate the problem of climate change. Germany is being politically ripped apart with the only thing keeping the country together being the knowledge of the coalition that a new election would see them replaced by a conservative "trickle down" moron and a ragtag band of Nazis. The Netherlands struggle with the fallout of a few court decisions impacting the agricultural sector. And in the US, Trump is looking to make a comeback, with his fanbase deeming climate change an outright hoax.

[1] https://de.statista.com/statistik/daten/studie/935886/umfrag...

[2] https://en.wikipedia.org/wiki/Year_2000_problem#Cost

- Nuclear war: averted by MAD?

- Peak Oil: happened. EV's are mitigating the consequences.

- Ozone hole: treaties banning the world-wide production of CFCs solved the problem.

- Y2K Problem: billions of dollars were spent fixing that problem. We did such a great job fixing that problem there are now a bunch of deniers stating the problem never existed in the first place!

See a pattern here? These problems just didn't up and solve themselves. We took action and solved them. Since there's no point dwelling on old problems we move on to new problems to solve. That process is otherwise known as advancement.

The Butlerian Jihad takes place about 11,000 years in the future.

Who says we have to wait?

Gets brought up frequently regarding this matter, but just check out the hitchhiking robot that ended its journey in Philly.

Even completely ignoring the current AI hype cycle, considering the terrible state of software quality nowadays a proper Butlerian Djihad might be a good thing if that's the only way to reset the status quo ;)

You might be stuck programming an abacus for a few years, or a TI-83

TI-83 has a Z80 CPU, that's good enough for a lot of computational tasks ;)

It all hinges on how the government is planning to handle mass unemployment.

If they're just going to be sending out a $1,200 check once a year, then yeah, things are gonna burn.

I wonder at what point OpenAI becomes the new public enemy, the Monsanto or Nestle of tech.

The Sauron of tech is more like it IMHO given the unprecedented dangerousness of creating something smarter than us.

There's plenty of Smeagols out there already. We need a Frodo.

I tend to disagree with this writing. I think it's overly broad and imprecise. It doesn't consider that the current advances are rather specific to what we call knowledge work.

I don't see hairdressers, carpenters or people working on construction sites etc. being directly affected by this. Everyone will still need a new haircut, some furniture and a house. Sure, there might be machines to do that at some point but these machines need to be built and maintained by someone. At least in the first round until we have other machines to build and maintain those machines...

Yes! "Reality has a surprising amount of detail"

http://johnsalvatier.org/blog/2017/reality-has-a-surprising-...

and I'm yet to see a single believably convincing AI threat scenario.

When the machines, all on their own, learn physics and then have the insights to reconcile QFT and GR and thereby provide us with an entire new model for the universe, then I'll agree that machine intelligence has passed human intelligence. As it is, AI is a very convincing mocking bird.

People think differently. A pedantic paperwork processor can be gold in the right position. An inspired create type needs a different job.

Are we planning to create AIs with different kinds of 'intelligence'? Or will we all be ruled by the same rote idiot-savant rule-lawyer in every position, making our lives miserable?

An old joke: The best world is French cooks, British police and German administrators.

The worst? British cooks, French administrators and German police.

What kind of AI are we planning to deploy?

AI cooks, AI police, and AI administrators.

Yes, but are they all the same? The right humans in those positions have very different 'programming'

AI cook by OpenAI, AI policeman by Raytheon, AI administrator brought to you by US government

Trajectory? The sun is on a trajectory to collide with the earth.

Large language models aren't intelligence in the same way humans are. I remain unconcerned about the machines and more concerned aboutthe intentions of those who weild them.

Imagine if we had the same enthusiasm about increasing human intelligence.

Until AI can experience joy in the act of attaining knowledge they're all dumb AF to me. Make a curious AI, then my eyebrows will suffer a slight upwards raise.

AI is only as good as the validity of data that goes into AI: without citation tracking, cleanup of junk data is ... impossible.

I love that, when finally a true productivity multiplier arrives, a knee jerk reaction is to stop it.

This article isn't asking us to stop it. It is talking about planning for a world in which AI displaces labour.

And even if this isn't the technology to do that, the fact we have absolutely 0 plan for a world beyond labour is frightening.

Yes, and we surely be occupied with that hypothetical crystal ball threat, while allowing our own investment money buy out all the property. There are real issues to be addressed right now

You can work on multiple problems at once.

And this problem is actually just multiple smaller problems, that are actually an issue in society today, that also need to be solved.

Well said. Ok

> Yes, and we surely be occupied with that hypothetical crystal ball threat, while[...]

I'm scared to ask you about climate change.

Real threat, meshes very well with my engineering background. I easily follow the narrative and technical arguments. I find good associations with my preexisting uni knowledge, including some hands on lab experiments. I especially appreciate articles, where I can go through source material myself (https://climate.nasa.gov). It just makes sense, imagining earth as a closed thermodynamic model.

The AI threat sounds to me more like a person scared of the Terminator long time ago. It's always: step one: AGI comes, step two is too late. There is never any believably real scenario given, only those that partially rely on magic or feelings.

I'm always thinking, who in power would ever allow that power be given to anybody, or in this case, anything, else?

I don't think it's pure magic and feelings. From what I gather the conclusions people like Yudkowsky draw are rather from our own history and simple natural order of the food chain.

> I'm always thinking, who in power would ever allow that power be given to anybody, or in this case, anything, else?

Those in power may share your skepticism towards agi and give up power involuntarily. I'm not saying I stand with Yudkowsky, because the weakness of his argument (or so I think) is the lack of falsification. Basically whatever you can think of he comes at you with "agi predicted that and counteracted." I do believe it is a serious topic and I'd hoped people didn't divide into two groups so quickly.

> simple natural order of the food chain

So, AGI would need to exist in some robotic form (or control whatever keeps us alive), without oversight, control and share our biological needs and aspiritions (in order to participate in the food chain).

This is one of example arguments so far removed from current reality, I'm shocked real engineers are participating in.

This reminds me of the 2001 Space Odyssey scenario and wouldn't be surprised that's where the collective fear stems from (entertainment industry). In the same movie, aliens lit up one of the planets and made a sun out of it. How did they do that? No one knows, they are so much more advanced than us.

> So, AGI would need to exist in some robotic form (or control whatever keeps us alive), without oversight, control and share our biological needs and aspiritions (in order to participate in the food chain).

No I don't think so. To be at the top of the food chain doesn't necessarily mean to compete over the same resources. I think it's more like having all resources available and we are certainly interested in putting AI in charge of many of our resources, provided a capable AI arrives.

> like having all resources

...

> AI in charge of many of our resources

> provided a capable AI arrives

Yes, this is actually coming closer to reality. Please be more specific and list one resource, how we arrive at AI that controls it and why we do it, if it's so capable - without oversight

Anywhere where automation is cheaper is basically an answer for you. If it can be automated it will be. That includes various kinds of warfare, surveillance, energy resource management, healthcare etc. What you are basically asking is where on Earth would we be interested in putting a supervisor performing better than humans for a fraction of the cost.

We do not live to be productive. That's an exploiter's dream.

Well, it might be I'm exploiting myself :)

“Policymakers should charge teams of experts with iterative scenario planning to help them regularly update their views on how the probabilities of the various scenarios evolve”

Or you know just have ai models run the scenarios.

I’m fairly fatigued on this sort of article… I do believe there will be large impact to work & society as we are able to manufacture intelligence. But that’s so obvious now as to be uninteresting.

Everyone says we’ll need solutions but I’ve yet to see meaningful work to propose any.

Why not ask the AIs to come up with a plan if they’re so intelligent ?

Oh for fuck's sake, LLMs are not in any sense "intelligent"

I saw a unique perspective on this: It’s more useful and appropriate to ask not if the AI is intelligent but if it can do work that once required human intelligence. Because that’s the milestone from a purely economic point of view.

> but if it can do work that once required human intelligence

We've been through many such milestones already though, starting with mechanical devices like the Antikythera mechanism, the last big "enabler" was the transistor about 75 years ago.

The singularity is nigh…

Well, yeah, we may master fusion, we may send man to mars, we may cure cancer, we may all die due to billionaires making billions more.

The universe is full of potential, not just for bad things but also good things.

Maybe we ought to work on and prepare for that instead.

Anyway, 15 mins after we get a proper AI it will make a spaceship and fuck off somewhere that we are not...if it is super intelligent.

"Anyway, 15 mins after we get a proper AI it will make a spaceship and fuck off somewhere that we are not..."

How will it do so? By the power of its mind?

No matter how smart something is, physical work remains physical work. This AI would have a quite hard time first, of figuring out self replication. Then maybe it could wander off ...

How do you know?

Because there currently is no self replicating industry, but allmost every step, from mining to refining to manufacturing requires human manual labour or supervision?

People often can only conceive 'intelligent' as 'me but with ALL the power ever'. Something of a misunderstanding, I think. Power isn't the same thing as intelligence.

All I want from an AGI is to order food for me without the bullshit of selecting the right venue, dealing with bad search in the apps, checking payment methods, making the payment, accepting 3DS notifications, checking delivery status. This took me 45 minutes yesterday, 43 minutes more than I considered acceptable.

That 15 minutes will take a bit longer :)

You don't need AGI for that, you just need a decent system designed for your needs rather than the competing needs of many vendors, including advertisers, for your dollar/attention.

Maybe that will come after the AGI fucks off :-)

That decent system takes more than 45 minutes to build though :)

You forgot to add "In my opinion".

>15 mins after we get a proper AI it will make a spaceship and fuck off somewhere that we are not

And maybe it will mine uranium to power the spaceship, and since most of Earth's uranium is in its core, it smashes the moon into the Earth to expose the core to its mining machines.

I would, if that was my goal, and I am barely intelligent.

Humans got too much ego to accept this, but don’t forget, we created it so embrace it

Strong AGI is our purpose. At several points in the history of biological life on earth there have been phase changes in the complexity of life: When we first captured mitochondria, when we first figured out multicellular life, when we form complex foodchains, when we first began to use tools, master agriculture, etc...

Life is an unbroken chain from the first self-replicating molecules to this very moment, and each moment that has passed has accelerated the ability of the universe to comprehend itself in our little corner of reality through the increasing complexity of life. In the same way that paramecium still exist, we will continue to still exist as part of something larger, but we will no longer be at the forefront of complexity as biological beings, which is a good thing.

I feel privileged to have existed at the end of one epoch and the beginning of another, which is not always something the average human gets to witness.

The only epoch change you exist in is 6th mass extinction event, but unlike the previous ones, this one is caused by billionaire mommy's boys who always need more.

Reply on news.ycombinator.com