GODDAMNIT PROTON. But, now I really want "lovingly handcrafted code" to be a selling point for a rival company.
Kelly Sovacool
@kelly.sova.cool
Bioinformatics Software Engineer. PhD from Schloss Lab at UMich. https://sovacool.dev 🏃🏻♀️🚴🏻♀️🧗🏻♀️ #python #rstats #nextflow #snakemake she/her My views are my own.
The interface to AI is all wrong for the unsophisticated user who does not understand how ML/AI works. It is designed to give you a definitive answer. That is the wrong interface.
Your monthly reminder that if you're using ChatGPT to do your science, you're failing your colleagues, community, and all of us. Here, ChatGPT makes up data and presents it as the real thing — then apologizes when confronted.
Don’t have a spot to put your leaves? Put them right under the tree! I took some time yesterday to dig a live edge around this maple in my yard. I’ll rake its leaves to this spot (but not up against the trunk, think donut not volcano), then in the spring plant native perennials for a soft landing
your yearly PSA: want to do something positive wrt the ecosystem while saving yourself a bunch of time and effort? skip the bagging and just rake leaves into a pile wherever convenient (bonus points if you do it under a tree). www.nps.gov/articles/000...
Researching downballot candidates is so tedious. What do you mean the only information about this candidate for the community college board of trustees is buried in comment threads in an invite-only facebook group?
“AI” companies presenting their software as having “personalities” is what started this whole “is it conscious” bullshit. That was a choice. Remove that presentation layer and the software is no less capable, but still just software.
Artificial Intimacy: Who We Become When We Talk to Machines - MIT STS
From the author of the prescient best seller and classic Reclaiming Conversation, Sherry Turkle, comes an urgent warning about how our ever-increasing reliance on human-like AI chatbots is…
sts-program.mit.edu
Excellent piece about the “zombie” science of “race realism” - the claim that biological races exist. It reviews two books which should “finally nail down the coffin” of this claim, covering its history and a careful discussion of the responsibilities of scientists when researching related topics
The “zombie science” of biological race realism
Review of Adam Hochman’s The Race Illusion and Lorusso and Winther’s Remapping Race in a Global Context.
academic.oup.com
if only there was as much hysteria and panic around our kids' total loss of mobility freedom and youthful independence at the hands of our car first at all cost cities that there is for them on e-scooters. Our kids want freedom, we just keep saying no.
I wish this place had polls. Lots of debate over screen protector removal in the replies. My general feeling is that unless you can’t see what’s underneath then the film is offering protection and you get a little bump of excitement when you finally peel it off and it’s like a brand new screen.
Respectfully speaking, I think this whole " (Insert demographics) loves their cars" line is highly misleading. Most Americans don't have a choice. Car dependency is forced upon them. People don't love the stress and debt.
I hopped on the train from Chicago back home to Michigan. I break down why I’m focused on the idea that we should speed up trains between Chicago and Detroit, and Detroit to Toronto! In Michigan we love our cars and driving, but taking the train is a great way to travel, and we can make it better.
I have to say, teenage me really thought adults were more competent than they actually are.
Forgiving infrastructure assumes people make mistakes. A person on a bike who loses their balance and falls shouldn't pay for it with their life just because there was no buffer between them and a vehicle 100 times their weight.
There are many many reasons I want El-Sayed to win. One of the lesser reasons is when I make my @5calls.org, it would be nice to only have to yell at one of my senators instead of both of them.
Senator, please cut the xenophobic rhetoric and get to the real problem: lack of laws enforcing privacy protections for everyone on US soil. I don’t want American-made vehicles spying on their customers either!
This from @michellegoldberg.bsky.social articulates the danger effectively. www.nytimes.com/2026/10/16/o...
Have to be clear just how dangerous this is. If it continues, it effectively removes constitutional rights from an entire class of people. Imagine you had to fear that if you criticize the regime, any error on paperwork filled out by your family decades ago could be used to strip your citizenship.
You seem to have misspelled “if we let automotive manufacturers of any national origin spy on their customers.” Beyond military security, there is obviously a privacy risk for all Americans using modern vehicles from any auto manufacturer, including American companies
I say this with all love and light but if you genuinely believe lines of code that only operate by means of electrical currents pushing through chemically manipulated plates of rocks full of substrate from Ajinomoto (yes the MSG company) are consciosus, you need to see a therapist. That is delusion.
Chicago is good and there should be more of it. Upzone Chicago
Bird biking (v.) - going for a bike ride with the primary purpose of birding. Set out hoping to spot a green heron, but more than happy to have seen a wood duck instead!
i think a lot of folks who are tuned in to the many harms of ai/surveillance tech struggle with how to talk about it without getting stuck on the details of particular devices or platforms, and imo this is the core frame: the future is up for grabs. it doesn't have to be this way. let's change it
"It's already here, what do you mean you think you can refuse?" is just another variation of the agency-stealing inevitability narrative. The future is not yet written, and we can shape it, starting with rejecting inevitability and claiming our agency. /fin
1/ Why I sometimes ignore Nature Methods papers when picking bioinformatics tools. A practical guide to choosing tools that actually work.
There are a lot of tells for faked writing---the anodyne register of it all, the sentence to sentence confusion---but the biggest one that I see is the lack of an audience the thing is supposedly "written" for
So the next time someone who is not an expert in the field concerned loses their mind over some supposedly revolutionary finding…bear this in mind. Whether this has to do with AI or not, it’s just a good rule of thumb. Don’t be someone else’s hype man if you don’t understand what you’re hyping.
After leaving tanker, oil from Exxon Valdez devastates Alaskan coastline on its own, killing over 250,000 seabirds.
Writing my PhD dissertation in Quarto saved my butt when I had to re-run the whole analysis pipeline — two weeks before it was due — after discovering a bug in the code. Thankfully it didn’t meaningfully change the results!
in 2020 my lab quit Microsoft Word and began writing in R Markdown, an open-source document-publishing system that facilitates a clear, permanent link between statements in text (scientific prose) and the code producing analyses and visualizations supporting those statements (scientific analyses)
a real joy of all this is that the vibe is so different to commercial software. We all know Word sucks but there's nothing we can ever do about it. one just suffers in community-developed software, if something doesn't work or doesn't exist, you open an issue, discuss, and contribute. for example:
open source software is amazing bc the incentives of users and developers are so much more in alignment than in commercial settings We've been experimenting with a transition to Quarto/Typst for scientific writing, it didn't yet have a good way to name highlight styles, I asked, and now it does!
The thing we need--that we know we need!--and that is EXTREMELY POPULAR, whenever it is built! is a model where people can walk to grocery stores, restaurants, parks, and schools where they live, and get to work without using their own individual car.
PSA: Do not be part of the RLHF process. It's not just improving the experiment, it's consenting to the model incorporating your input and ideas and then feeding them back to others (including the employees who ask it to "do a breakthrough").
I'm very late to this, but:
Fascinating insights into the early history of Zotero, including the fact that the name is the Albanian word for "to learn well or master". Hats off to everyone involved in this project. Also if you're not using Zotero, why not? Dump that closed rubbish you're using now and switch.
"As programming increasingly becomes a caffeinated bender with magical bots, this measured pace and emphasis on communal thinking likely holds an important lesson." Love this piece on the development of Zotero. newsletter.dancohen.org/archive/the-...
Yeah OpenAI or Anthropic or whoever might go bankrupt, some specific use cases like ‘therapists’ might be litigated to oblivion. But things that involve interfacing with code are never, ever going back to the modal default being a human being typing things out, like it’s just full stop not happening
AI might be a bubble, but is it a tulip bubble or a railroad bubble