Fabio E. Tonti
@tontief.bsky.social
Absentee mathematician gone teaching. Aspiring statistician. Original AndOrNot_robot. Stats & maths education.
Reposted by Fabio E. Tonti
What intro stats textbook to use?
statmodeling.stat.columbia.edu/2025/11/09/w...
statmodeling.stat.columbia.edu/2025/11/09/w...
What intro stats textbook to use? | Statistical Modeling, Causal Inference, and Social Science
statmodeling.stat.columbia.edu
November 9, 2025 at 4:20 PM
What intro stats textbook to use?
statmodeling.stat.columbia.edu/2025/11/09/w...
statmodeling.stat.columbia.edu/2025/11/09/w...
Ordinals (set theory) and ordinal variables have always been my favourites, and here both show up:)
Approiximately my Belz Lecture in Melbourne from a couple of weeks ago (there may be a recording later?)
notstatschat.rbind.io/2025/10/30/l...
notstatschat.rbind.io/2025/10/30/l...
Laws and Orders - Biased and Inefficient
notstatschat.rbind.io
November 9, 2025 at 7:46 AM
Ordinals (set theory) and ordinal variables have always been my favourites, and here both show up:)
I think we could only make that work with a lot more casual inference.
The advances we've made in statistics, experimental study design, and causal inference over the past century are remarkably useful for understanding our world. But there is never been a push to make people use them like we are seeing with generative AI. Perhaps take a moment to consider why.
November 7, 2025 at 3:55 PM
I think we could only make that work with a lot more casual inference.
Reposted by Fabio E. Tonti
AI tells you what you want to hear, statisticians usually tell you what you don't want to hear
The advances we've made in statistics, experimental study design, and causal inference over the past century are remarkably useful for understanding our world. But there is never been a push to make people use them like we are seeing with generative AI. Perhaps take a moment to consider why.
November 7, 2025 at 9:13 AM
AI tells you what you want to hear, statisticians usually tell you what you don't want to hear
Reposted by Fabio E. Tonti
I think I understand how it can be that LLMs are both exceptionally good and quite terrible at programming. It's because there are two entirely different skillsets that we both call "good at programming." LLMs have only one of them.
blog.genesmindsmachines.com/p/llms-excel...
blog.genesmindsmachines.com/p/llms-excel...
LLMs excel at programming—how can they be so bad at it?
My explanation for the mystery of why LLMs can be both exceptionally good and quite terrible at programming.
blog.genesmindsmachines.com
November 6, 2025 at 3:43 PM
I think I understand how it can be that LLMs are both exceptionally good and quite terrible at programming. It's because there are two entirely different skillsets that we both call "good at programming." LLMs have only one of them.
blog.genesmindsmachines.com/p/llms-excel...
blog.genesmindsmachines.com/p/llms-excel...
Are there no useful projections? From the image it looks like there should be.
Does anyone have examples of published 3d figures in fairly well-known publications?
Does anyone have examples of published 3d figures in fairly well-known publications?
"Journal style does not allow 3-dimensional figures. Please rework Figure 5 into a figure with only 2 axes if possible"
a man in a suit and tie is smoking a cigarette while sitting in an airplane .
ALT: a man in a suit and tie is smoking a cigarette while sitting in an airplane .
media.tenor.com
November 5, 2025 at 4:45 PM
Are there no useful projections? From the image it looks like there should be.
Does anyone have examples of published 3d figures in fairly well-known publications?
Does anyone have examples of published 3d figures in fairly well-known publications?
Reposted by Fabio E. Tonti
Blatant authoritarianism and the mindless injection of LLMs into higher education are both driving more and more academics to speak up about our social mission. A silver lining of sorts. We have been lazy to assume every understands our value.
November 2, 2025 at 1:42 PM
Blatant authoritarianism and the mindless injection of LLMs into higher education are both driving more and more academics to speak up about our social mission. A silver lining of sorts. We have been lazy to assume every understands our value.
Wait who has an unfavourable opinion of arXiv? Did I miss anything? Genuinely confused.
The 8chan of Academia™️ is becoming more storm drain, less open cesspool.
Another casualty of the LLM revolution.
In "arXiv’s CS category, review articles and position papers must now be accepted at a journal or a conference and complete successful peer review."
In "arXiv’s CS category, review articles and position papers must now be accepted at a journal or a conference and complete successful peer review."
November 2, 2025 at 7:33 AM
Wait who has an unfavourable opinion of arXiv? Did I miss anything? Genuinely confused.
😅
We need a graph database to relate them to each other together with an API Gateway and Load Balancer plus scaling implemented and we put an AI chatbot on top so that you can always ask in natural language how many of each are left
November 1, 2025 at 10:45 PM
😅
I actually found those battles of Lord's paradox extremely motivating:)
Maybe @statsepi.bsky.social or @epiellie.bsky.social can remember better than me. But the most memorable to me were the #EconTwitter vs #EpiTwitter wars, the odds ratio shitstorms, and Pearl and Senn arguing to the end of time about Lords Paradox. We should write a piece on the 'good old days'.
October 30, 2025 at 11:41 AM
I actually found those battles of Lord's paradox extremely motivating:)
Reposted by Fabio E. Tonti
JD Vance claims that diversity weakens unions, as people end up distrusting each other and not organizing.
Let me tell you two menswear stories related to this claim. 🧵
Let me tell you two menswear stories related to this claim. 🧵
October 30, 2025 at 8:07 AM
JD Vance claims that diversity weakens unions, as people end up distrusting each other and not organizing.
Let me tell you two menswear stories related to this claim. 🧵
Let me tell you two menswear stories related to this claim. 🧵
Reposted by Fabio E. Tonti
A response to all the people who say "if your results depend on your random seed you have bigger problems."
blog.genesmindsmachines.com/p/random-see...
blog.genesmindsmachines.com/p/random-see...
Random seeds and brown M&Ms
Your first mistake was assuming people actually understand how random numbers work.
blog.genesmindsmachines.com
October 23, 2025 at 4:48 PM
A response to all the people who say "if your results depend on your random seed you have bigger problems."
blog.genesmindsmachines.com/p/random-see...
blog.genesmindsmachines.com/p/random-see...
I do feel the whole point should be stretched into a Ted talk :)
Also, both cultures suffering is perfect! As we like to say in my hood: everyone's equally unhappy....
Also, both cultures suffering is perfect! As we like to say in my hood: everyone's equally unhappy....
We tried to tell y'all to stop calling everything "AI" many years ago and you just wouldn't listen and now the poor machine learners must also suffer alongside the statisticians 😜
so I can explain this: it's not generative AI: it's usually deep learning models trained on meteorology tasks and it can be quite effective
October 24, 2025 at 5:13 PM
I do feel the whole point should be stretched into a Ted talk :)
Also, both cultures suffering is perfect! As we like to say in my hood: everyone's equally unhappy....
Also, both cultures suffering is perfect! As we like to say in my hood: everyone's equally unhappy....
Sure, but @clauswilke.com 's writeup will be useful to many people. Deep down in this thread many good points are made. I still think Wilke's post is a useful read.
I’m seeing some misinformation about pseudo-random number generator best practices going around the internets. Let’s talk about why the pseudo-random number generator seed you use shouldn’t actually have any impact on your results and, consequently, you can choose whatever seed you damn well please.
October 22, 2025 at 7:21 PM
Sure, but @clauswilke.com 's writeup will be useful to many people. Deep down in this thread many good points are made. I still think Wilke's post is a useful read.
A very nice read, also a good overview if anyone is new to the topic.
We need to have a conversation about random seeds. Don't use 42.
blog.genesmindsmachines.com/p/if-your-ra...
blog.genesmindsmachines.com/p/if-your-ra...
If your random seed is 42 I will come to your office and set your computer on fire🔥
Figuratively. More likely you'll get a stern talking to.
blog.genesmindsmachines.com
October 22, 2025 at 3:58 PM
A very nice read, also a good overview if anyone is new to the topic.
Yay I'm looking forward to watching it! (had to be in a not-so-fun meeting at that time yesterday)
The recording is now available so that you can confirm that I indeed have a German accent and color-match my outfits with my Zoom background.
youtu.be/YL0co26ng-g?...
youtu.be/YL0co26ng-g?...
October 21, 2025 at 3:27 PM
Yay I'm looking forward to watching it! (had to be in a not-so-fun meeting at that time yesterday)
Reposted by Fabio E. Tonti
Before coffee: I hate everybody
After coffee: I feel good about hating everybody
After coffee: I feel good about hating everybody
October 20, 2025 at 2:00 PM
Before coffee: I hate everybody
After coffee: I feel good about hating everybody
After coffee: I feel good about hating everybody
Reposted by Fabio E. Tonti
In the most tech-bro format imaginable, but an interesting “in the tent” critique of LLMs as “cognition”. Offcourse it’s not going to be to many ppls taste given its decidedly conveyed in the sf-tech tent dialect… youtu.be/lXUZvyajciY?...
Andrej Karpathy — “We’re summoning ghosts, not building animals”
YouTube video by Dwarkesh Patel
youtu.be
October 20, 2025 at 4:55 PM
In the most tech-bro format imaginable, but an interesting “in the tent” critique of LLMs as “cognition”. Offcourse it’s not going to be to many ppls taste given its decidedly conveyed in the sf-tech tent dialect… youtu.be/lXUZvyajciY?...
Reposted by Fabio E. Tonti
Today I experienced an accomplished person with legit field expertise say that if they are worried about the accuracy of the answer chatgpt gives, they just ask "him" to confirm "his" results. Sometimes "he" affirms, and other times "he" changes them. This is all fine.
October 17, 2025 at 6:48 PM
Today I experienced an accomplished person with legit field expertise say that if they are worried about the accuracy of the answer chatgpt gives, they just ask "him" to confirm "his" results. Sometimes "he" affirms, and other times "he" changes them. This is all fine.
Reposted by Fabio E. Tonti
It’s such a great feeling to re-visit an article/blog/book that *used* to feel confusing and overwhelming and feel like it makes total sense now😎
October 17, 2025 at 12:43 AM
It’s such a great feeling to re-visit an article/blog/book that *used* to feel confusing and overwhelming and feel like it makes total sense now😎
Reposted by Fabio E. Tonti
I read and write, I explore and I question, I design and script and analyse, I interpret and communicate. I do this to train my mind in the hopes of one day generating new knowledge. New knowledge that might even be useful, and that no algorithm can yet be trained on.
Y'all. I just got ChatGPT to do everything in R for this manuscript. I mean EVERYTHING. And it's all legit and reproducible. I'm shook.
How are we mentoring our trainees in statistics now? Who needs to learn coding in R line by line, and who doesn't?
scienceforeveryone.science/statistics-i...
How are we mentoring our trainees in statistics now? Who needs to learn coding in R line by line, and who doesn't?
scienceforeveryone.science/statistics-i...
Statistics in the era of AI
How do we mentor, teach, and do stats when AI can do so much of the work?
scienceforeveryone.science
October 12, 2025 at 8:07 AM
I read and write, I explore and I question, I design and script and analyse, I interpret and communicate. I do this to train my mind in the hopes of one day generating new knowledge. New knowledge that might even be useful, and that no algorithm can yet be trained on.
"This could just as easily be work for a student or early-career researcher. They’ll overcome the issue you’re having and will learn something about code, data management, analysis, and research in general."
He makes this point many times. We need to think about processes and medium-term outcomes.
He makes this point many times. We need to think about processes and medium-term outcomes.
I clearly had a lot of thoughts/feelings from reading Terry's blog post, so I wrote one of my own
It ended up being a bit of a stream-of-thought post but I think it captures some of my worries
benharrap.com/post/2025-10...
#statssky #episky #academicsky
It ended up being a bit of a stream-of-thought post but I think it captures some of my worries
benharrap.com/post/2025-10...
#statssky #episky #academicsky
October 12, 2025 at 8:05 AM
"This could just as easily be work for a student or early-career researcher. They’ll overcome the issue you’re having and will learn something about code, data management, analysis, and research in general."
He makes this point many times. We need to think about processes and medium-term outcomes.
He makes this point many times. We need to think about processes and medium-term outcomes.
Reposted by Fabio E. Tonti
Love this, it hits one of my biggest concerns with a lot of this stuff - i can appreciate how it can be useful for an expert to do things they may already know, sure, but what happens in 5/10/20 years after we cut off the talent development pipeline at the knees? The process is so so important
October 10, 2025 at 2:07 AM
Love this, it hits one of my biggest concerns with a lot of this stuff - i can appreciate how it can be useful for an expert to do things they may already know, sure, but what happens in 5/10/20 years after we cut off the talent development pipeline at the knees? The process is so so important
Reposted by Fabio E. Tonti
Got a 10 minute video for you. The dots really clicked for me after watching this.
youtu.be/oIMFZf5dUFA?...
youtu.be/oIMFZf5dUFA?...
Demystifying . . . (dots): R package dev fundamentals
YouTube video by Josiah Parry
youtu.be
October 10, 2025 at 5:46 PM
Got a 10 minute video for you. The dots really clicked for me after watching this.
youtu.be/oIMFZf5dUFA?...
youtu.be/oIMFZf5dUFA?...