Tobias Wolfram

857 posts

Tobias Wolfram banner
Tobias Wolfram

Tobias Wolfram

@_twolfram

Sociogenomics, Behavioral Genetics, Statistical Genetics @herasight

Katılım Eylül 2018
1.3K Takip Edilen1.2K Takipçiler
Tobias Wolfram retweetledi
george davey smith
george davey smith@mendel_random·
Failure to understand that the fundamental principle of Mendelian randomization (MR) is of gene-environment equivalence contributes to the flood of nonsense MR papers that are appearing; Shah Ebrahim, Gib Hemani and I explain why in this short commentary. journals.plos.org/plosmedicine/a…
English
2
34
103
18.9K
Tobias Wolfram retweetledi
Jake Eaton
Jake Eaton@jkeatn·
This is second time we've used Anthropic Interviewer and the first time we've deployed it at scale. Quite accidentally, we ended up conducting (what we believe is) the largest qualitative study ever I'm a mixed-methods social scientist by training. Traditionally, when it came to understanding what people think, that meant quantitative analysis of lower resolution data (polls, surveys, etc.) or hand-wavey analysis of in-depth qualitative data. Using Claude to conduct *and* analyze interviews bridges that tradeoff between breadth and depth AI also makes access much, much easier. Had we run this study in person, in the real world, it would have taken hundreds (if not several thousand) enumerators many 1000s of hours to conduct. It also affords us access to places we could otherwise never go. I once led a five-person team in Tanzania that reached a few hundred people. It took 3 weeks. In this study we heard from people 80,000 people in 159 countries, in cities and rural areas, in daily life and in war zones, and more, in just one I'm still, even after months, beginning to wrap my head around the scale of this work. Like, to a social scientist, it's quite unbelievable. This could produce dozens of dissertations! It is also, of course, imperfect—certainly speaking to an AI is different than speaking to a person—and as a team we're all still figuring out how to make this research as useful as possible: what questions to ask and how, what to analyze and why, and how that all feeds back into what we do as a company. This is, as we say in the blog, a brand new form of social science Hat tip to @saffronhuang for leading this for the past few months. Here's one of my favorite quotes
Jake Eaton tweet media
Anthropic@AnthropicAI

We invited Claude users to share how they use AI, what they dream it could make possible, and what they fear it might do. Nearly 81,000 people responded in one week—the largest qualitative study of its kind. Read more: anthropic.com/features/81k-i…

English
27
66
765
110K
Tobias Wolfram retweetledi
Ruud Koopmans
Ruud Koopmans@Ruud_Koop_mans·
The social sciences are dead. It does not suffice that there is also good and solid social science. Science depends on consistent standards. If you allow garbage to coexist and receive the same credentials and funding as good science you kill the credibilty of the whole field.
James L. Nuzzo, PhD@JamesLNuzzo

PhD student at Australian National University publishes, what amounts to be a personal diary entry, in an academic journal--and will presumably include it as a thesis chapter to gain the PhD credential. The paper starts out: "My first drag performance was in 2023 at a doof underneath a bridge hosted by a group of trans and queer young people (like me at the time!). Dressed in very little other than mesh, tassels, and a painted face, I wasn’t sure what I was doing, awkward in my movements, but feeling confident as I was cheered on. I have since then developed my performance art, christening myself as SISTAboy and finding out that more people do in fact want to see this fat trans blak body boogie the night away! This has also highlighted to me the ways in which spaces continue to hold expectations and privileges on body and desire. As a fat Wiradjuri and trans person, I previously struggled to find spaces in the community where I felt wholly welcomed or understood."

English
20
100
548
60K
Tobias Wolfram retweetledi
Luca Repetto
Luca Repetto@luca_repetto_c·
What are the effects of large human-capital shocks on innovation? In a new paper, we study how WWI military deaths across British communities affected local invention over the next decades. We find that places that lost more young men became persistently less innovative. (🧵1/11)
Luca Repetto tweet media
English
44
316
1.6K
294.6K
Tobias Wolfram retweetledi
Herasight
Herasight@herasight·
This week, the Herasight team presented three papers at the ACMG Annual Meeting in Baltimore. The work addresses three important contributions to the science of PGT-P: 1. Validation of polygenic predictors 2. Type 1 diabetes risk modeling 3. Embryo genome imputation 🧵
Herasight tweet media
English
1
9
40
2.6K
Tobias Wolfram
Tobias Wolfram@_twolfram·
@Scientific_Bird The accepted paper should hopefully be published before the end of the month, we added some additional imo quite neat analyses. And V2 is already in the works.
English
2
0
24
731
Inquisitive Bird
Inquisitive Bird@Scientific_Bird·
New genetic predictor of intelligence dropped (by @_twolfram et al.) Substantial increase in predictive validity. Only slight within-family attenuation (shared environmental influences on intelligence must be weak). Decent portability to non-European ancestry groups too.
Inquisitive Bird tweet media
English
3
40
271
16K
Tobias Wolfram retweetledi
Athul Paul Jacob
Athul Paul Jacob@apjacob03·
What does encoding an interpreter INTO the weights of an autoregressive transformer + O(log N) attention get you? A transformer that executes arbitrary programs, Sudoku solvers, optimization algorithms, and more, with 100% accuracy across millions of tokens, exponentially fast
English
6
37
361
26K
Tobias Wolfram retweetledi
Simon Grimm
Simon Grimm@Simon__Grimm·
New post. Within 20 minutes, I created an AI teaching tool that would trigger the EU's high-risk AI rules. Given AI progress, these rules are rapidly becoming unworkable. They are currently up for reform—those interested in Europe's success should push for their removal.
Simon Grimm tweet media
English
1
3
20
728
Tobias Wolfram retweetledi
Dr Jingmei Li
Dr Jingmei Li@J_i_n_g_m_e_i·
There's currently no agreed-upon standard platform for PRS implementation. This study did a head-to-head comparison across the most commonly used arrays and low-pass WGS. Only 7 of 26 "high-risk" individuals were consistently flagged across all platforms. nature.com/articles/s4385…
English
0
12
27
2.6K
Tobias Wolfram retweetledi
Marios Georgakis
Marios Georgakis@MariosGeorgakis·
New GWAS of age-related hearing loss in a massive sample of 456,613 cases & 1,053,834 controls detects 140 loci. Hearing loss is a huge unmet need with >25% of people >60y affected and no available pharmacotherapy. If you want to find promising drug targets, start here!
Marios Georgakis tweet media
English
2
12
41
4.3K
Tobias Wolfram retweetledi
Alex Strudwick Young
Alex Strudwick Young@AlexTISYoung·
Herasight's ImputePGTA algorithm has made it possible to access polygenic embryo testing for IQ and diseases from routine IVF data (PGT-A). Now we've developed an innovative new algorithm that dramatically increases accuracy, especially for non-European ancestry couples.
Alex Strudwick Young tweet media
Jeremy@jeremyli__

Today we’re announcing an algorithmic breakthrough. Herasight’s ImputePGTA algorithm has enabled couples around the world to access polygenic embryo testing from routine IVF data (PGT-A). Now it yields substantially higher accuracy, especially for underrepresented ancestries

English
2
10
51
4.8K
Tobias Wolfram retweetledi
Herasight
Herasight@herasight·
This is a very technical breakthrough, but it means that couples (especially non-European ancestry) can benefit even more from polygenic embryo screening and that data files for unused embryos can help improve accuracy of the remaining embryos. Great work Jeremy and team!!
Jeremy@jeremyli__

Today we’re announcing an algorithmic breakthrough. Herasight’s ImputePGTA algorithm has enabled couples around the world to access polygenic embryo testing from routine IVF data (PGT-A). Now it yields substantially higher accuracy, especially for underrepresented ancestries

English
1
13
35
3.4K
Tobias Wolfram retweetledi
Camille Michèle Williams
Camille Michèle Williams@Cognitive_Camz·
Our multi-ancestry GWAS of EXTernalizing conditions (ADHD, substance use...) in ~4M people reveals neurodevelopmental risk, drug-repurposing targets, and yields one of the strongest psychiatric polygenic indices yet! 🧬🎉 doi.org/10.64898/2026.…
English
1
23
81
14.2K

Keşfet

@saffronhuang @Theory_Society @arctotherium42 @Scientific_Bird @NatureGenet @ErikVanZwet @stephensenn @elonmusk