Brendan O'Loughlin
LightNews LightNews
banner
brendanoloughlin.bsky.social
Brendan O'Loughlin
@brendanoloughlin.bsky.social
1K followers 570 following 730 posts
He/him PhD student in the Carstens Lab at The Ohio State University
Posts Media Videos Starter Packs
Pinned
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · Mar 2
I would like to announce that I'm starting my PhD this fall in the Carstens Lab at Ohio State University!

I haven't nailed down a specific topic yet, but it will most likely be in computational evolutionary biology using insects as a study system.

I'm very excited to take this next step!
4 40
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 21h
If I'm understanding this right, this would imply a lot (most?) of published lineage introgression claims could be false positives?

I'm just getting into this sort of math so forgive me if this is a dumb question
1
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 21h
I also miss east African tea a lot, its different
1
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 6d
That allows them to mimic human variability (if i ask you the exact same question twice you will almost certainly word the answer differently each time)

If someone knows better than me please correct me!
2
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 6d
If i had to hazard a guess, it would be because LLM outputs are based on slight randomness. Unlike a traditional computer program where the same inputs get the same outputs every time, LLMs have a degree of randomness that allows them to give different responses to the same prompt.
1 1
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 9d
Ahhh, traditional cheating. Humanity is healing
3
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 9d
One time as a child was enough haha
1
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 10d
These are not fun to get in your pants
1 2
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 11d
Awesome stuff!
1
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 12d
Is this a new genus or a revision?
1 2
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 17d
True but that would be so much effort 🫠

I talked to someone who knows more than me about computation and they said it might be possible to account for biases using some ML method but that it would be a huge pain no matter what
1
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 17d
Not really helpful to teaching your students, just what I came up with when I self-reflected on why it's not a problem for me lol
1
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 17d
I feel like it's easier to understand if you know a bit of calculus. Not really the math, but just the conceptual idea that weird things happen over infinite timescales (e.g. asymptotes approach a number bet never reach it etc.)
1 1
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 18d
My life goal before I die is to figure out how to unbias inat data to extract actual trend estimates
1 1
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 19d
I collected and dissected this male Pterostichus (Abacidus). Is there any way that the species can be identified by the genitalia?
#Coleoptera
Carabid beetle top down held by tweezers
1
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 22d
Or not, looks like i cant
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 22d
It would be very helpful! I'll send you a DM
1
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 22d
The coloration is essentially identical, but the genitalia and barcodes are very different.
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 22d
Anyone want to describe a new moth species with me lol?

I just discovered that a super common holartic species are two different species between North America and Europe, but I don't have any experience with Lepidoptera taxonomy
3 6 8
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 24d
I will!
1
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 24d
I will be! They are not my advisors but I will be working with the evolution wasps and some other small arthropods
1 2
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 24d
Thanks I'll remember this!
1
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 24d
I'm very curious if an accurate 3D model can be made of a ~1 mm wasp
1
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 24d
Its possible! Speaking of which, I've been looking into using microCT for some of my small wasps. I believe we have a machine at my institution, but would you be willing to answer questions I have when I get around to trying to do it?
1
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 24d
The "spiral" is a little more scrunched inwards than I would expect. Doesn't change the ID though!
1
brendanoloughlin.bsky.social
Brendan O'Loughlin @brendanoloughlin.bsky.social · 24d
This is Agelenopsis pennsylvanica if you want a species ID!

Looks a little weird though, maybe due to drying?
1 2