Saloni

@scientificdiscovery.dev

44923 Followers 1172 Following

Co-founder & editor, Works in Progress. Writer, Scientific Discovery. Podcaster, Hard Drugs. Advisor, Coefficient Giving. // Previously at Our World in Data. Newsletter: https://scientificdiscovery.dev Podcast: https://harddrugs.worksinprogress.co 🏳️‍🌈

Our PETRUSHKA (Personalise antidEpressant TReatment for Unipolar depreSsion combining individual cHoices, risKs and big datA)
Google search: How many days of the week have poop in it Google: Technically exactly one day of the week contains the word poop-- Wednesday (Wednes-poop)
Naomi Wolf tweet: Terrifying. Also confirms/explains the conversation I overheard in a restaurant in Manhattan 2 yes ago in which an Apple employee was boasting about attending a top secret demo: they had a new tech to develop vaccines w nanopatticles (sic) that let you travel back in time. Not kidding
Much of the neglect by others of lysozyme and to an even greater extent of penicillin was due, I think, to Fleming's inability to express himself clearly and lucidly in either words or print. He was certainly one of the worst lecturers I have ever heard. It cannot have been due to shyness because, so far as I know, he never refused an invitation to speak in public and he must have given hundreds of lectures and speeches after he became famous. In a dull monotone, without humour or emphasis, they may have given the audience the facts but that is about the best that can be said for them. His papers, too, gave the facts but little more. Prolixity, at which Wright was an adept, should it is true be avoided, but Fleming went to the other extreme and was much too economical with words.
You do not have access to this file
A logo of a cat with its mouth open, above the text 'Clinical Trials'
'Dick' = fat or thick
bored pooh
Advantages of jointly tackling trial and market design What are the major advantages of my suggestion? Because everything is done in the same trial, it is cheaper, faster, and statistically more efficient (less bias, more power). A shared design allows the funder to make better head-to-head comparisons between these antivirals. In some cases it also avoids companies from competing for trial subjects. It can de-risk a sensitive part of the R&D process for firms. For producers many uncertainties revolve around whether their intervention can feasibly be tested. Under this proposal the funder gets control of the information that is generated, but assumes (some of) the burden of creating the testing infrastructure. This solution can also be thought about as an intermediate solution between push and pull funding—especially if the funder also covers some/all costs of the trial.
Mail rail
Giving What We Can's Fundraiser
Malaria prevalence over the twentieth century - map of regions where malaria was eliminated.
Card room. Some cards laid out on a table with Benjamin Franklin memorabilia behind them
Does the brain really change at four major turning points around ages nine, 32, 66 and 83?
A line chart showing the rate of aging according to 'biological clock' measures, except it looks like a roller coaster.
Peter Gray not realising that the survey methodology had changed between 2009 and 2013, as well as the question phrasing.
That's the joke
me: if siblings share 50% of their DNA with each other, and humans share 50% of their genes with bananas, is my brother as related to me as a banana? claude fable: *switched to Opus 4.8*
me: why did the chicken cross the road? claude fable: *switched to Opus 4.8*
To End A Plague. America's fight to defeat AIDS in Africa. By Emily Bass.
A cute pic of Gila monster (a lizard)
“The risk was in not doing it.” Dana Lewis was blessed with a problem doctors recognized and knew how to treat, more or less. Her body didn’t produce enough insulin. The treatment for this condition is straightforward — more insulin — but it requires exquisitely precise dosing. If a diabetic injects too much insulin, their blood sugar will crash, which, left unchecked, can be fatal. Administering too little risks high blood sugar, which slowly damages organs.3 Today, diabetics can monitor their blood sugar and choose their insulin dose with wearable continuous glucose monitors, which alert them when it falls or spikes to dangerous levels. But back in 2014, when Dana’s story started, continuous glucose monitors (CGMs) were new to the consumer market and not yet considered reliable. Their function was to trigger a blood stick check or serve as a last-minute alert to catastrophically low blood sugar. But these monitors aren’t perfectly reliable, and they were even less so in 2014. The sensor in Dana’s could be thrown into a panic just by lying on the arm it was attached to, so false alarms were common. After numerous false alarms, she learned to sleep through the alerts. She complained to the manufacturer, who told her, “It works for most people” — as if that solved her problem. Dana’s biggest fear was having a hypoglycemic episode in her sleep and dying before she had a chance to do something about it. She did not find the fact that the app woke up other people in time very comforting. So she and her new boyfriend (now husband) Scott Leibrand set to work on a solution. First, they needed real-time access to the data from the CGM, the only way to identify problems as they happened. Luckily John Costik, a software engineer and parent of a diabetic preschooler, had already created Nightscout, a program to extract data from CGMs and upload it to cloud storage. Scott, also a software engineer, wrote a program to monitor these uploads, and if they reported low bloo…
Pokémon card for a potion
In most cases, the more covariates we have in the data, the higher the chances we catch all confounders. (Overfitting can be a risk in prediction problems, but in public health there are usually too few rather than too many covariates in the model.) This is why, in a world with a growing wealth of data, observational evidence treated with causal methods ends up winning.
Wegener's puzzle: Glaciers once traversed continents that are now tropical. Map of continents and glacier movement across them.
That sign can't stop me because I can't read!
You wouldn't download a dataset without reading the data quality notes
Toy Story 5 poster
My result on In The Weights dot Com. I am in the weights for AI models, in the top 10%.
RDD of cumulative risk of shingles over time, with a sudden drop when shingles vaccines are introduced.
I have no Purpose and No in a world destination I'm just Wondering arou World where Can't possess anything
'What is this? A school for ants?!?'
Nintendogs
Club Penguin poster
ACS estimates of annual cancer cases and deaths by site in the US.
My baby boom ridgeline plot
My prompt in Claude Fable: "What's the role of microtubules in cell biology?" Claude Fable: *flagged for messages on most cybersecurity or biology topics* *switched to Opus 4.8*
Ten respiratory pathogens antigen test kit
Me standing in front of a poster for the 10% Pledge (to donate 10% of your lifetime income to effective charities) with my face on it. The text quote from me says 'You don't have to work on the front lines to make a difference.'
A thumbnail of me, Jacob Trefethen and Katharine Collins's faces on top of the text '2Malaria 2Vaccine' with a mosquito as the cherry on top.
Infographic about fish and aquaculture
Selfie of me and Joey
Bubble chart of divorce rate by occupation.
A man dancing on stairs on a stage, with the word Rationale in bright lights behind him
Me at the podcast studio wearing a jumper that says Baroque
Me at the podcast studio wearing a jumper that says Baroque
A woman turning and looking at the viewer. Pick the word that fits best. 1. aghast, 2. baffled, 3. distrustful, 4. terrified.
A shopfront that says SALONI
The billion dollar molecule, by Barry Werth
Me on the red circle on the TED stage, with graphs of survival curves in childhood leukemia on slides behind me
Me and Jacob in front of giant letters spelling TED
me on the TED stage
Topology is informally about studying which shapes can be continuously deformed into each other — the classic example being that a coffee mug and a donut are "the same" topologically because one can be smoothly morphed into the other (they both have exactly one hole). So "girl topology" would be about which types of girls can be continuously deformed into each other. The joke is that Saloni is interested in knowing which girl types are topologically equivalent — i.e., which ones are secretly the same girl just stretched or squished a bit. Can a horse girl be smoothly transformed into a wine mom? Is a girl boss just a deformed tiger mom?
My result: German. German 42%, autistic 18%.
My girl typology radar chart