Live data from Hacker News

The Abundance Illusion

carlyle.com

31–40 of 46 posts

Re: The Abundance Illusion

#31
post #8

61% AI generated, according to Pangram https://www.pangram.com/history/e5a00ace-94cc-436e-b87b-a094... c'mon, you're Carlyle, a trusted institution for financial advice! How can I trust what you're saying if the AI-generated text is so blatantly obvious?

I do not understand why this matters. Judge the content on its merit. It makes no difference if “AI wrote it”.

Nonsense. Of course the tool matters.

That's why I only use code written in Vim. Emacs corrupts the othewise identical bits, just like AI. Gets 'em all greasy and then they smell funny.

Re: The Abundance Illusion

#32
post #8

61% AI generated, according to Pangram https://www.pangram.com/history/e5a00ace-94cc-436e-b87b-a094... c'mon, you're Carlyle, a trusted institution for financial advice! How can I trust what you're saying if the AI-generated text is so blatantly obvious?

I do not understand why this matters. Judge the content on its merit. It makes no difference if “AI wrote it”.

I wasted some time trying to get a sense of the content, to "judge it on its merit". It is a total slop. And if i knew beforehand that it is AI, i'd have spent much less time as the first short look would have confirmed that it is slop.

Re: The Abundance Illusion

#33

Earlier quoted context omitted.

I do not understand why this matters. Judge the content on its merit. It makes no difference if “AI wrote it”.

Nonsense. Of course the tool matters. That's why I only use code written in Vim. Emacs corrupts the othewise identical bits, just like AI. Gets 'em all greasy and then they smell funny.

Get me some holy water

Re: The Abundance Illusion

#34
post #20

Earlier quoted context omitted.

Yes there are studies, for example last year Pangram's false positives were measured to be under 0.5%. https://www.pangram.com/blog/third-party-pangram-evals Personally, at first I thought these sorts of tools were dumb and wouldn't really work, but I think it works because it just isn't designed to be "adversarial". If you want your AI to trick Pangram, you can make an AI to trick Pangram. It just catches people who…

Any binary classifier can have a FPR under 0.5% if you don't have any restriction on FNR...

While I am quite skeptical of the claims linked above, that link does indeed cover the FNR at the FPR of 0.005, and finds it broadly to be on the same order of magnitude, i.e. also below 0.005.

Re: The Abundance Illusion

#35
post #9

'Figure 2: US energy exports outpace production' We're producing 14,000,000 barrels of oil per day and exporting about 6,000,000 per day, but the author used a graph with two different y-axis scales to make the lines cross so it looks like more oil is being exported than produced.

And to top it off both Y axes have the exact same label. The graph has no indication as to which line belongs to which axis. Of course, we can easily infer which is which from the surrounding article, but still... terrible data visualization

Different scales between the left and right sides as well. The vertical distance for 500k bpd on the left is about half of the right.

Re: The Abundance Illusion

#36
post #24

Earlier quoted context omitted.

If you look at US share of energy consumption by source you can see that solar is 2.83% of energy used in the US. Natural gas is 34.2% . https://ourworldindata.org/profile/energy/united-states

That's a really interesting/informative link you have. Some highlights are: - Energy use per capita has been declining since 2000 - Overall energy use has been basically flat since 2000 - Essentially, coal is being replaced by gas, almost 1:1. Lump both together and their relative share of the market is pretty stable. - Solar went from 0.3 to 2.8% in the last 10 years, wind from 0.3% to 4.2% in the last 20 years. 7%…

[dead]

Re: The Abundance Illusion

#37
post #8

61% AI generated, according to Pangram https://www.pangram.com/history/e5a00ace-94cc-436e-b87b-a094... c'mon, you're Carlyle, a trusted institution for financial advice! How can I trust what you're saying if the AI-generated text is so blatantly obvious?

Is there a particular part you are claiming is incorrect? If so, point it out and name your sources. If not, what’s your point?

Re: The Abundance Illusion

#38
post #21

Some have touted the U.S. as having achieved "energy independence" because of its status as net exporter of oil. This is premature. The U.S. is marginally a net exporter. However, it still imports a lot of oil. It just exports slightly more than it imports. So, if the world markets gets too dicey, it'd be easy to just cut off the export taps and keep all that oil for Americans, right? Not so fast! First, it's the wro…

The US on net exports 2.5M barrels a day, or roughly 10% of its total consumption, and the majority of its imports are from countries like Canada that are themselves net exporters and would be unlikely to retaliate against a US export ban.

There are definitely technical issues with refinery capacity, but I don’t think they’re insurmountable if the US seriously wanted to attempt an export ban, even in the short term. The fallout from the rest of the world in the form of other trade retaliations would likely be very serious though.

Re: The Abundance Illusion

#39
post #8

61% AI generated, according to Pangram https://www.pangram.com/history/e5a00ace-94cc-436e-b87b-a094... c'mon, you're Carlyle, a trusted institution for financial advice! How can I trust what you're saying if the AI-generated text is so blatantly obvious?

I do not understand why this matters. Judge the content on its merit. It makes no difference if “AI wrote it”.

If AI wrote it, it is not worth effort engaging with.

Re: The Abundance Illusion

#40
post #8

61% AI generated, according to Pangram https://www.pangram.com/history/e5a00ace-94cc-436e-b87b-a094... c'mon, you're Carlyle, a trusted institution for financial advice! How can I trust what you're saying if the AI-generated text is so blatantly obvious?

I do not understand why this matters. Judge the content on its merit. It makes no difference if “AI wrote it”.

Article from a trusted organization is supposed to be grounded in real-life events and filtered through their specific area of expertise. They bet their reputation on it, at least in theory. Current agents have none of that, and you can't trust their output to the same extent in practice. Too much recognizable slop in the article suggests high degree of autonomy of the agent that has been used, and raises doubts about trusting the entire article. It might be the valid opinion of the author rephrased by the model, but you can't tell as it obscures actual intent and the amount of real life data.
Post reply on HN