Justin Ales
justinales.bsky.social
Justin Ales
@justinales.bsky.social
Whether it reduces income depends on if Unis will replace EU with other international
students.

Top universities would trade some UK spots for EU, and admit more overseas from other nations. So revenue the same but better students, and harder for home students to in.

1/4
Tuition fees are likely to be a dealbreaker for the UK-EU talks, if the EU insists on £9.5k. You can’t ask struggling universities to take the hit. Or if you do, they’ll probably try not to recruit EU students. Awkward.
May 15, 2025 at 10:58 AM
The interactive components in this tutorial are well done. This is a great aid for teaching data literacy. I’ll be adding it to my data science class.
This excellent interactive tutorial on misleading data visualizations explores the idea of a "counter chart" — the graph you draw in response to refute a misleading claims

flowingdata.com/projects/dis...
Defense Against Dishonest Charts
This is a guide to protect ourselves and to preserve what is good about turning data into visual things.
flowingdata.com
February 16, 2025 at 6:13 PM
Reposted by Justin Ales
1. Today the NIH director issued a new directive slashing overhead rates to 15%.

I want to provide some context on what that means and why it matters.

grants.nih.gov/grants/guide...
NOT-OD-25-068: Supplemental Guidance to the 2024 NIH Grants Policy Statement: Indirect Cost Rates
NIH Funding Opportunities and Notices in the NIH Guide for Grants and Contracts: Supplemental Guidance to the 2024 NIH Grants Policy Statement: Indirect Cost Rates NOT-OD-25-068. OD
grants.nih.gov
February 8, 2025 at 12:18 AM
New paper. Really interesting outcome. We used experts to modify journal abstracts so make the outcome “incorrect”. Then tested both experts and AI LLMs to find the real results. Turns out AI LLMs beat the neuroscientists!
"Large language models surpass human experts in predicting neuroscience results" w @ken-lxl.bsky.social
and braingpt.org. LLMs integrate a noisy yet interrelated scientific literature to forecast outcomes. nature.com/articles/s41... 1/8
December 4, 2024 at 7:42 PM
Reposted by Justin Ales
"Large language models surpass human experts in predicting neuroscience results" w @ken-lxl.bsky.social and braingpt.org. LLMs integrate a noisy yet interrelated scientific literature to forecast outcomes. arxiv.org/abs/2403.03230 1/6
Large language models surpass human experts in predicting...
Scientific discoveries often hinge on synthesizing decades of research, a task that potentially outstrips human information processing capacities. Large language models (LLMs) offer a solution....
arxiv.org
March 7, 2024 at 10:48 AM
Reposted by Justin Ales
Utrecht University did not submit data to THE, & thus doesn’t appear in the World University Ranking 2024, because it deems the metrics invalid and sees them as promoting unhealthy competition. UU instead emphasises cooperation & open science and supports DORA.

dub.uu.nl/nl/nieuws/un...
Universiteit Utrecht doet niet meer mee met THE-ranking | DUB
De Universiteit Utrecht heeft dit jaar geen gegevens willen aanleveren voor de wereldranglijst van Times Higher Education. Daarmee is de UU niet opgenomen in de World University Ranking 2024. Rankings...
dub.uu.nl
September 28, 2023 at 2:26 PM