Erik Poppleton

They can be re-purposed for research usage, they’re the same kind of GPUs used in things like weather prediction, drug design, and data analysis. But that requires the capital to buy up the hardware when the AI companies fall apart, which research institutes are currently struggling with.

This really is part of the joy of science! I love hunting down papers from the early days of biochemistry or molecular biology and seeing the methods, the thought processes, and feeling the continuity with great names from the past.

This is why we fund scientists to study things like oyster slobber even if you don’t think it sounds important

Oof, this is gnarly! Thanks for sharing.

When you put it that way, I suppose what I’m asking for isn’t so different than what attention does with the tokens themselves.

Has there been any research on ‘fuzzy’ tokenizers? The reason we can mentally combine ‘going to’ and ‘gonna’ is because of this map between words and sound which lets those co-exist, while ‘boing to’ would be different. Probably too much processing power in the tokenizer layer for that to work tho.

Also, I added you to the Molecular Programming feed posting list. If you include a 🧬 emoji in your posts, they should get picked up by the feed:

bsky.app/profile/did:...

This is great! Would love to see if this method could be used to get the stacking energies of the inverted A|A stack in the kissing loops used in RNA origami.

Some really nice fundamental biophysics of DNA base stacking 🧬!

Anyway, learning another language is good for you because you die a thousand deaths of embarrassment and then you come out more humble and even more aware of your various human failings

This is sad news… it was such a cool museum when everything was working properly (which as you can read in the piece, it often wasn’t). I particularly liked the “living art” in the atrium.

I get that the news cycle is packed right now, but I just heard from a colleague at the Smithsonian that this is fully a GIANT SQUID BEING EATEN BY A SPERM WHALE and it’s possibly the first ever confirmed video according to a friend at NOAA

10 YEAR OLD ME IS LOSING HER MIND (a thread 🧵)

Postdoc position available in the BioEmu team at @msftresearch.bsky.social AI for Science - Berlin DE or Cambridge UK. Looking for candidates with backgrounds in #MachineLearning #AI Biophysics or Bioinformatics

jobs.careers.microsoft.com/global/en/jo...

I have it in my big annual new releases playlist and sometimes find it super grating and have to skip, but other times that loudness is exactly what my brain wants and it’s energizing. Different mixes for different days, I guess.

Yeah, that production is really loud and flat. Could have used some time in the sauna.

So this is what getting a lobotomy feels like (laudatory)

Were they trying to loose?? Authenticity is what always wins the day in these contests.

This was something @englishhistpod.bsky.social brought up in episodes about the 1500s as well! By that point English was already splitting into three registers based on vocabulary (largely Anglo-Saxon, largely French, largely Latin).

Ahh, wait, I get it! “MJ” stands for Midjourney. That account used to post AI re-interpretations of art pieces and historical artifacts.

YouTube video by NationalMuseumsScotland The Orkney Hood in the National Museum of Scotland

Idk what the images in that post are, but they’re not the Orkney Hood. You can see the actual hood in its exhibit in this video. Alas it isn’t nearly as cool.

youtu.be/Y5KveU3OC5c?...

Screenshot of a Bluesky post containing a screenshot of a Reddit post saying “The Hatman does NOT respect gender pronouns past 450mg” in a quote post of Melania Trump looming in a hat.

Why does this keep happening??

These graphs make no sense…

Student migration should be relatively constant because there’s only a certain number of seats at universities. So how are there more student visas being issued in 2025 than all types of migration combined pre-2021? Am I missing something?

Cation-Controlled Assembly, Activity, and Organization of Biomimetic DNA Receptors in Synthetic Cell Membranes Biological cells use cations as signaling messengers to regulate a variety of responses. Linking cations to the functionality of synthetic membranes is thus crucial to engineering advanced biomimetic agents such as synthetic cells. Here, we introduce bioinspired DNA-based receptors that exploit noncanonical G-quadruplexes for cation-actuated structural and functional responses in synthetic lipid membranes. Membrane confinement grants cation-dependent control over receptor assembly and, when supplemented with hemin cofactors, their peroxidase DNAzyme activity. Cation-mediated control extends to receptor lateral distribution to localize DNA-based catalysis within phase-separated membrane domains of model synthetic cells, imitating the localization of multimeric membrane complexes to signaling hubs in living cells. Our modular strategy paves the way for engineering from the bottom-up cation-responsive pathways for sensing, signaling, and communication in synthetic cellular systems.

Really happy to see this published in its final form @jacs.acspublications.org! 🧬 In this paper, we introduce a new class of biomimetic receptors to host functionality in synthetic cell membranes

Elita Peters @dianatanase.bsky.social @dimichelelab1.bsky.social

pubs.acs.org/doi/10.1021/...

What are those numbers relative to? Shouldn’t the angle be 0-135 or so?

So excited to see our newest study out in print! Led by the brilliant @jgezelle.bsky.social and @sophiekorn.bsky.social, we discovered how viruses from highly divergent families converge on a shared RNA structure mechanism to inhibit cellular nucleases. #RNAsky #LoveVirology

Zotero has dark mode now?!?!

The last blinding window on my desktop has gone dark in a huge win for cave-gremlin science.

🧵How to write and manage your first research budgets

The point of funding is to convert it into quality research. A well-spent research budget should fund the idea it was raised on, plus revision experiments, plus preliminary data for the next grant. So you need to spend, while avoiding waste.

I was asked twice to review the same blatantly AI-generated review paper by what I thought was a reputable publisher in the sciences. Really broke my trust in the publishing system, at this point I’m leery of anything published in a non-prestige journal, especially reviews, after 2022.

Excellent work on dynamic droplet segregation by @wverstraeten.bsky.social!

Something about this paper stimulated even more discussion than usual in group meetings, very interesting dynamics, both socially and under the microscope. 🧬🧪

Super detailed look at charge distribution on DNA helices producing alignment between similar sequences! 🧬