The Armstrong and Miller Show | Hitler Has Only Got One Ball YouTube video by Hat Trick Comedy

No. All this is well-documented.

www.youtube.com/watch?v=Y1DW...

Planning a Dioscuri application (≤15 yrs post-PhD)? https://ncn.gov.pl/dioscuri/

We’d love to host a PI in philosophy of mind / cog sci / AI / neuro at IFiS PAN (Warsaw).

€300k/year, 5 years, your own Centre, serious autonomy.

If you want to build something real — let’s talk.

How realistic is this? The current centres seem to be 100% natural science, with a very heavy emphasis on biomedical stuff.

Jethro Tull - Velvet Green (Sight And Sound In Concert: Jethro Tull Live, 19th Feb, 1977) YouTube video by Jethro Tull

Well, we'll see (hear) about that on 07 Aug. It'll be spectacular:

www.youtube.com/watch?v=nZ7F...

Kyllä. Alkoi ehkä n. vuosi sitten.

Is your job finished? There's still a lot of tinkering there..

And: it's the northernmost medieval castle (still standing).

Both Maiden & Mr Collins should have been here long, long time ago already (well, Collins was, with Genesis). Oasis: oh yes.

The former.

So, how do you separate "robustness" from "replication" (or "reproduction"), or do you? In any case, I think it's clear that "robustness analysis" originates from Levins (though he did not use that precise term), and then got economics, and then to social sciences. "Sensitivity" amounts to the same:

Bitten by snakes 200 times – on purpose: US man’s quest to help deliver new antivenom Tim Friede put his ‘ass on the line’ to help stop snakebite deaths – whose numbers appear to be rising amid the climate crisis

Sure, replication is not one thing. It's repetition, but it's not repeating tout court, as discussed. It's about "repeating the relevant". Which leaves us with defining that. And providing such a definition is what I'd like to see being done. Incidentally, your example reminded me of this:

That's why this topic fascinates me, and that's why I'm trying to pin down the relevant terms (nailing down definitions - doing good ol' conceptual analysis - is an awfully underrated undertaking, btw). We cannot be expected to repeat everything. So, "replicating" is repeating of something.. what?

That sounds reasonable - to the point of being obvious. That's why focusing on "repeat" is fruitful. We cannot repeat experiments tout court. We can only repeat aspects of them, or something; the data/results relationship must be "similar in a relevant sense". But a lot hinges on these words.

Yea, nothing can be repeated simpliciter - and I think realising this is the key to unraveling this brouhaha. The way "replication" is used here seems to suggest this is not so. We agree on this, I take it. Then the question is: what can/should be repeated? It must be some data/results relationship.

Yea, that was the term I was mostly thinking of. They write:

replicate = "to repeat entire experiments, gathering and analysing the data from scratch"

This is more flexible in relation to your view? (BTW, here we have a further term: "repeat". And do we need to separate all of these from "copy"?)

So yea, maybe we have four different terms:

reproduction
replication
robustness/sensitivity
reliability

And central here is the relationship between the results and their justification (methods, data, analysis). And I find your definitions intuitive. But do they match these Nature studies?

So, how exactly would you define these:

reproduction = ?
replication = ?

Is there some further term to take into account?

We had a little exchange about robustness vs sensitivity here:

bsky.app/profile/tuom...

Thanks! I'll definitely have a look at these (I've read at least bits of Uljana's paper). Do you have an opinion about the replication/reproduction distinction? I read that Nature piece only a while ago, and I see these studies separate three things: replication, reproduction, robustness.

Precisely! And semantics is only necessary for understanding, not sufficient. And assessing reports of research finds amounts to much more than simple understanding. Perhaps machine learning can assist us in making such assessments. But talking in terms of "automating" them is simply nonsense.

Photo of Resignation letter from Dr. Martin Peterson

First Plato—now Dr. Peterson—out the door at Texas A&M!

Who’s next?

Here is Dr. Peterson’s letter of resignation, which he just shared with the faculty.

They are removing buttons? Or the damn thing? Sounds like they are in the business of spreading AI-slop everywhere.

I managed to have a look at this Nature piece now. The bit below made me pause. I'm not sure what's going on here, but this sounds like precisely the sort of work that we should *not delegate to our magnificent AI-overlords.

bsky.app/profile/natu...

I managed to have a look at this Nature piece now. The bit below made me pause. I'm not sure what's going on here, but this sounds like precisely the sort of work that we should *not delegate to our magnificent AI-overlords.

bsky.app/profile/natu...

Hmm.. Is this (Samuel Johnson 1756) the first time in history that the "causation is not correlation" idea is voiced? This is an interesting era, given Newton's influence (on everything). (Also, I wonder if he, or somebody, cared to explain what needs to be added to correlation to get to causation.)

Okay. And of course you don't want to use an unreliable platform. But some of these essay websites look good to me. (In fact, I refer to these sorts of pieces a lot in my teaching material, as these are easy/fast to read, and you can follow the references for details.)

Your uni/department/project does not have a blog? I've published several pieces on my uni's blog. It's been a very useful forum for interaction (if/when you can discuss the stuff further on social media). Some of the online essay websites could be good too.

Well yea, for the purpose of my class, I have enough here already (and I often use social media (BS) in my classes). I have a sort of a portfolio of this topic, with a lot of material (that I should update). The (PhD) students are supposed to do independent research and present the stuff to others.

Well, I'm all ears. And I might do some independent thinking based on what we've discussed here now. It would indeed be cool to be able to say something new about this stuff next month, not just the same-old same-old. (And I have a feeling you might be onto something here.)

Cool. And I might be on board with you. This sounds like a new, interesting take on this brouhaha. Have you (somebody) published something (anything) about this? (I'm having this topic in my class next month, and it would be great to have actually something new to discuss.)

So.. replication becomes, in effect, superfluous?