textninja

I agree, I sounded more defeatist than I intended. No doubt the moat is expansive, but it’s not impossible to cross. At least until the regulators get involved. [I did it again, didn’t I.]

We’re on two very different wavelengths my friend.

I think what’s happening is that I’m arguing between the lines and you’re arguing on them. It’s like matter and antimatter, but no matter.

The only good ethicist is a fired ethicist.

Relax, I’m being flippant, but think about it.

> N.B. textninja quietly admits my most important point, #3

Your most important point is the distinction between an assumption and a presupposition? I honestly didn’t see the need to challenge that.

> No disagreement here. There should be an independent monitoring body, something like the SEC.

We may disagree even more than you think ;)

> Imagine having ChatGPT level AI running in an ASIC inside earphones.

Oh, it’s going in brains. And no, don’t sign me up.

The moat between OpenAI and projects like this is access to expensive computing resources, an organizational mandate to invest in aforementioned compute, and finally, access to high quality training data. They were first to market so they’re only getting further ahead thanks to the invaluable real world usage data.

I don’t think their moat ever had anything to do with innovative architecture. If anything, projects like this will widen the chasm since it’s easier for OpenAI to implement new architecture in their models than it is for an independent researcher to scale their ideas.

If projects like this are seeds then the problem is that OpenAI owns all the land.

For 1, no, having more poison in a well and therefore making it more murky, green, and scary looking is not worse than having just enough poison in there to kill you.

For 2, I disagree - you may want chemicals in your well if they’re designed to counteract the poison that was already in there. See 1.

In 4, 5, and 6 it seems you’ve lost sight of the fact that the analogy is being used as an argument against something that’s occurring in the real world. Whether my counter argument perfectly fits the original analogy is irrelevant because I’ve merely co-opted it as a rhetorical device. Are you suggesting it doesn’t fit with reality?

For 7 and 8 I’ll just emphasize that the ethicists were employed by the chemical manufacturing companies. A conflict of interest is not a great foundation to build on.

> no ethicists is probably worse than fallible ethicists

How fallible are we talking here? “Probably” is carrying an awful lot of weight in that argument.

> You want to deflate the wallop this analogy packs, but I don't think you succeed.

That’s just like your opinion, man.

Good rhetorical imagery, but ultimately an emotional argument with a subtext loaded with assumptions:

1) that there is an unpoisoned well to begin with

2) that the chemicals are poison or can only be used for poison

3) that increased access to chemicals will result in more poisoned wells

4) that the people who invented the chemicals are benevolent stewards and will not just use them to further their own interests

5) that it is better for aforementioned benevolent stewards to hoard the chemicals than permit access to those who would concoct poisons, antidotes, and preventative measures, not to mention industry innovations that aren’t related to poison at all

6) that the interests of the people on the ethics committees are aligned with the interests of humanity

7) that ethics committees employed by a chemical manufacturing firms is the only thing standing between ethical and responsible management of chemicals and complete chaos and environmental destruction

8) that the fired ethicists were good at their jobs in the first place

Coding is clearly all those things.

Barely tangential to the article, but it’s getting exhausting seeing lazy take after lazy take where an author flippantly refers to crypto as a dumb fad when they really mean specific currencies or maybe NFTs as a whole.

Biological reproduction for one. Copy/paste for another. Biological reproduction is only tenuously related to understanding and copy/paste isn’t even related at all. We can copy around weights and biases all day without understanding them.

> Largely because it's impossible to replicate something when we don't entirely know how it works.

On the contrary. These are mostly orthogonal.

> But, color me skeptical

That is not a creative color.

> We will never solve the problem of a token prediction engine being able to generate a sequence of tokens that the vast majority of humans interpret as not corresponding to a true statement.

I think we already solved that problem by making sure the vast majority of humans never agree about anything.

It is true that we probably won’t ever get a machine trained on human output to ever be completely accurate (GIGO) but with the right systems and sensors we can at least get probabilistic accuracy. Let’s not forget how human consensus gets shaken up every so many centuries.

Interesting. Creativity inspires confidence for me.

For the most part I agree, but will qualify that the caveats you listed apply to the out of the box version of ChatGPT. I expect these limitations will be overcome by using it as programming substrate and connecting it to other models and APIs.

I am impressed by what we’ve seen from ChatGPT so far, but am especially excited to see what industry does with LLMs as new type of building block.

The black swan event includes the unexpected emergent capabilities of LLMs, which can pass professional examinations and exhibit many attributes of general intelligence, combined with the broad availability to market and massive pressure on big industry players to adapt and innovate or die.

If things keep going this way then pretty soon the black swans are going to outnumber the white ones.

According to style guides and grammar rules, sure, but if you don’t give your em-dashes room to breathe people might mistake them for a hyphen — true connoisseurs of fine punctuation require one space to announce their dignified entrance, and another to signal a triumphant exit.

Trust but verify. We’ll get processes built around this for accuracy sensitive applications. I imagine it will look something like a GAN configuration with two or more LLMs trained to adversarially critique and fact check the outputs of the other(s). Might even mimic the relationship between the hemispheres of our brains.

This is a black swan event which, if anything, vindicates the hype from dreamers who were ahead of their time. They had a lot of the theory right but lacked the horsepower. This will complete the last mile for other AI tech and be the lacquer that gives it the all important finishing touch.

In summary, spam it everywhere — including in articles admonishing its overuse — because it looks cool and you know it.

Unless it’s specifically trained not to, I’m fairly certain AI will bypass this without even trying. Such is the way of the unsupervised hill climb on huge unstructured datasets. To put it another way, in order for an artist’s style to truly be invisible to AI, it would have to be invisible to humans.

It’s The Selfish Meme. Your artistic style has as much “drive” to reproduce as your DNA. There seems to be a tug of war happening between the interests of the artist and the interests of the art.

Also, just a side point but I don’t think billionaires value comfort as much as you think. If that was the case they probably would have stopped somewhere in the tens of millions.

Hey sweet, my existence is pointless!

> If I am to take your reply seriously, you mean colonize the moon with ultra rich people and then turn it into an arc ship to settle a new world

Yes, I said it in a jokey way but I’m actually seriously speculating about that as a distant future scenario, though I’d include their servants, clones, entertainers, staff, recruits, and progeny in that, along with probably a DNA bank of every person, animal, and strand of bacteria on earth.

> with limited access to air, water, food, sex, security, shelter until they finally perished cold and alone in the nothing

Who said anything about that? I mean, they’d be “limited”, yes, in the strictly mathematical sense that these things are finite, but they need not lack in abundance. I did say luxury space craft, didn’t I? I’m not talking about heated seats here, I’m talking nuclear powered underground cities…

Another benefit might be protection from asteroids and meteorites if built underground.

I wonder which endgame is more probable: satellite, or space ship? We have people perfectly willing to shroud earth in sulphur to “protect” us from the sun, so who’s to say those same types won’t repurpose the moon into a luxury spacecraft and take it on a joyride to distant stars?

ChatGPT had fun with this one. Pretty soon YouTube will be replaced with ITube and we’ll have blockbusters on demand.

> In the pulse-pounding sci-fi blockbuster, Lunar Armageddon, the British government enlists Rolls-Royce to construct a series of nuclear reactors on the moon as the last hope for saving Earth from an imminent energy crisis. When legendary pinstriping artist Mark "Marksman" Court is recruited for his unique skills, he joins forces with the fearless commander, Emily "Em" Walker, to assemble an elite team of astronauts and experts. As they race against time to complete the high-stakes mission, Court and Em must navigate political intrigue and life-threatening challenges on the treacherous lunar landscape.

Also, you can have high general intelligence and still be confidently wrong.