1-4 is only unreasonable because you've written a strawman version of 4. Here is a version that makes total sense:

1. You make a superficially compelling argument for invading Iraq

2. A similar argument, if you squint, can be used to support invading Vietnam

3. This argument for invading vietnam was wrong because it made mistakes X, Y, and Z

4. Your argument for invading Iraq also makes mistakes X, Y and Z

5. Therefore, your argument is also wrong. 

Steps 1-3 are not strictly necessary here, but they add supporting evidence to the claims. 

As far as I can tell from the article, they are saying that you can make a counting argument that argues that it's impossible to make a working SGD model. They are using this a jumping off point to explain the mistakes that would lead to flawed counting arguments, and then they spend the rest of the article trying to prove that the AI misalignment counting argument is making these same mistakes. 

You can disagree with whether or not they have actually proved that AI misalignment made a comparable mistake, but that's a different problem to the one you claim is going on here. 

This again seems like another "bubble" thing. The vast majority of conservatives do not draw a distinction between USAID and foreign aid in general. And I would guess they do associate foreign aid with "woke", because "woke" is a word that is usually assigned based on vibes alone, for the things perceived as taking away from the average american to give to some other minority. Foreign aid involves spending american money to help foreigners, it's absolutely perceieved as "woke". 

Look, I wish we lived in a world where people were rational and actually defined their terms and made their decisions accordingly, but that's not the world we live in. 

I don't think foreign aid is at risk of being viewed as woke. Even the conservative criticisms of USAID tend to focus on things that look very ideological and very not like traditional foreign aid.

This just isn't true. Yes, exaggerated claims of "wastefulness" are one of the reasons they are against it, but there are many more who are ideologically opposed to foreign aid altogether. 

I can link you to this exchange I had with a conservative, where they explictly stated that saving the lives of a billion foreigners would not be worth increasing the national deficit by 4%, because they are ideologically opposed to american taxpayer money saving foreign lives, no matter how efficiently they do it. Or see the insanely aggressive responses to this seemingly innocuous scott alexander tweet. Or here is a popular right wing meme specifically mocking liberals for having large moral circles. 

I suspect that you are in a bubble, where the conservatives you know are fine with foreign aid, so you extend that to the rest of conservatives. But in a broader context, 73% of republicans want to cut foreign aid, while only 33% of democrats do. 

I think the term "goodharting" is great. All you have to do is look up goodharts law to understand what is talked about: the AI is optimising for the metric you evaluated it on, rather than the thing you actually want it to do. 

Your suggestions would rob this term of the specific technical meaning, which makes thing much vaguer and harder to talk about. 

I think employers should be obligated to have at least a kitchen area where people can heat up or make their own food. People should not be forced to buy lunch. 

I don't mind you using LLMs for elucidating discussion, although I don't think asking it to rate arguments is very valuable. 

The additional details of having subfield specific auditors that are opt-in does lessen my objections significantly. Of course, the issue of what counts as a subfield is kinda thorny. It would make most sense for, as claude suggests, journals to have an "auditor verified" badge, but then maybe you're giving too much power over content to the journals, which usually stick to accept/reject decisions (and even that can get quite political). 

Coming back to your original statement, ultimately I just don't buy that any of this can lead to "incredibly low rates of fraud/bias". If someone wants to do fraud or bias, they will just game the tools, or submit to journals with weak/nonexistent auditors. Perhaps the black box nature of AI might even make it easier to hide this kind of thing. 

Next: there are large areas of science where a tool telling you the best techniques to use will never be particularly useful. On the one hand there is research like mine, where it's so frontier that the "best practices" to put into such an auditor don't exist yet. On the other, you have statistics stuff that is so well known that there already exist software tools that implement the best practices: you just have to load up a well documented R package. What does an AI auditor add to this?

If I was tasked with reducing bias and fraud, I would mainly push for data transparency requirements in journal publications, and in beefing up the incentives for careful peer review, which is currently unpaid and unrewarding labour. Perhaps AI tools could be useful in parts of that process, but I don't see it as anywhere near as important than those other two things. 

My obvious answer is that the auditors should be held up to higher standards than the things they are auditing. This means that these should be particularly open, and should be open to other auditing. For example, the auditing code could be open-source, highly tested, and evaluated by both humans and AI systems. 

Yeah, I just don't buy that we could ever establish such a code in a way that would make it viable. Science chases novel projects and experiments, what is "meant" to happen will be different for each miniscule subfield of each field. If you release an open source code that has been proven to work for subfields A,B,C,D,E,F, someone in subfield G will immediately object that it's not transferable, and they may very well be right. And the only people who can tell if it works on subfield G is people who are in subfield G. 

You cannot avoid social and political aspects to this: Imagine if the AI-auditor code starts declaring that a controversial and widely used technique in, say, evolutionary psychology, is bad science. Does the evo-psych community accept this and abandon the technique, or do they say that the auditor code is flawed due to the biases of the code creators, and fork/reject the code? Essentially you are allowing whoever is controlling the auditor code to suppress fields they don't agree with. It's a centralization of science that is at odds with what allows science to actually work. 

As a working scientist, I strongly doubt that any of this will happen. 

First, existing AI's are nowhere near being able to do any of the things with an accuracy that makes them particularly useful. AI's are equipped to do things similar to their training set, but all science is on the frontier: it is a much harder task to figure out the correct experimental setup for something that has never been done before in the history of humanity. 

Right now I'm finishing up an article about how my field acually uses AI, and it's nothing like anything you proposed here: LLMs are used for BS grant applications and low-level coding, almost exclusively. I don't find it very useful for anything else. 

The bigger issue here is with the "auditors" themselves: who's in charge of them? If a working scientist disagrees with what the "auditor" says, what happens? What happens if someone like Elon is in charge, and decides to use the auditors for a political crusade against "woke science", as is currently literally happening right now? 

Catching errors in science is not something that can be boiled down to a formula: a massive part of the process is socio-cultural. You push out AI auditors, people are just going to game them, like they have with p-values, etc. This is not a problem with a technological solution.