Well we can't see what was in the Debate.txt you fed the reliably left leaning Grok. Given your ethics I suspect it had lots of ellipses which you'll deny hide anything pertinent to the issue.

Yes, you can, mhaze.

//Well we can't see what was in the Debate.txt you fed the reliably left leaning Grok.//

Simply click on the attachment at it appears in a pane to the left.

Grok leans to the right, by the way.

Oh, it appears it doesn't show for others. So, I've uploaded it here for all to see:

http://filebin.net/j83yj4eyl0oejtly

It appears that link will only be valid for six days, so here’s a permanent one:

http://drive.google.com/file/d/1UuiCgyLvCe7NufXgFNLzxr9ZUqA3Gy0U/view

I invite anyone to try it themselves - not just with Grok (which is widely regarded as the most right-leaning of the major AIs), but with any model.

They all land on the same conclusion.

In fact, future generations can test it. I'll keep that link open for all to enjoy.

The jig is up, mhaze.

And for the record, even if Grok were left-leaning (which, ironically, it’s not), it wouldn’t matter - because the analysis it performed wasn’t ideological.

It didn’t “agree” with me because it shares my politics. It analysed structure: who cited sources, who addressed rebuttals, who remained logically consistent, and who fell back on vagueness, misdirection, or personal jabs.

Those aren’t left/right criteria. They’re just the fundamentals of good argument.

If your position had been stronger on those measures, Grok would’ve reflected it - regardless of worldview. AI doesn’t “like” or “dislike” anyone. It weighs what’s on the page.

"But Grok is getting better and is now my go-to when I'm checking facts for the first time." - mhaze
http://forum.onlineopinion.com.au/thread.asp?article=23401#398459

This is quite amusing. Here is John Daysh, who was using AI to write some of his posts at one stage, relying on AI for his rebuttal. But it doesn't really settle anything substantial, and the AI is wrong on the most substantial point because it is misled by Daysh, demonstrating that AI isn't some omniscient and infallible deus ex machina that can adjudicate every debate.

The point it is misled on is that every technology requires subsidies in the first instance. This is completely untrue. Who subsidised wooden tools, flint tools, fire, steam, internal combustion engine, flight?

It is also wrong about the IMF figure for fossil fuel subsidies because it doesn't analyse it. If it did it would find the claimed subsidy is actually a claimed externality, which according to the best evidence (Nordhaus's work for which he won the Nobel) doesn't represent a cost, let alone a subsidy.

Then it relies on weight of citation to accept Daysh's position that RE is cheaper than coal, which is of course an argument from authority and therefore illogical, and in this case wrong. The fact is that even CSIRO now sees coal as being as cheap as RE, but without properly considering system costs which tip the balance much more in favour of coal.

Maybe Daysh could go back to writing his own stuff, and do a bit less of the Gish Gallop? And a tip in using AI - it is just a research assistant with a highly retentive memory, logical within parameters, and prone to flattery, but as the small print says - check its conclusions because it makes mistakes.

John

"Renewables do lower wholesale prices"

Yes, because of the effect of grid saturation. The example of that loss making solar farm might become more common.

The second part of your answer is a contradiction. It stands to reason that if something is cheaper then it should not be more expensive. You are familiar with the concept of amortisation? That a higher cost means that something is more expensive is not a matter of equivocation.

As for nuclear power, the cost calculations are done on the basis of individual facility constructions. Would you attempt to calculate historical r and d costs for solar? It seems a bit silly apart from the interest you might get from learning the history of the subject.