Sign in

Jeffrey Shallit 🇺🇦

@shallit.bsky.social
1.6K followers 1K following 3K posts

Grumpy old American mathematician and computer scientist, now retired. My Erdös number equals 1.

PostsRepliesMedia
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 6h
Assholes get defended by assholes; that's called "asshole magnetism".
130
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 7h
But you're not "defending" women's sports, you're just being an asshole.
241
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 15h
The Atlantic is not a serious publication. They live for contrarian articles and don't care if they're accurate. Why do you think almost all of the good science writers there left?
040
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 04/10/2026
I was just asking myself why they publish drivel like this, though: www.nytimes.com/2026/10/01/o...
nytimes.com
Opinion | These Things Are Computers. Stop Talking About Them as if They’re Human.
If we want to figure out the proper way to control A.I. models, we need to recognize what they really are.
000
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 03/10/2026
As the editor of a journal that gets AI-generated submissions pretty much every day (and sometimes 5 or 6 in a day) I can agree about the waste of time. But these are almost all submission by "independent researchers", not professional mathematicians.
010
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 03/10/2026
There's a huge difference between "low quality" and "wrong". "Low quality" can mean genuine junk, or it can just mean it wouldn't be published in the very top journals. In my own area LLM's are solving problems that have been open for decades.
100
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 03/10/2026
I don't see how the limit imposed by arxiv has anything to do with whether the results are correct or not. Do you want to explain how you made that connection? There is a lot of AI slop in mathematics, and I've commented on it. But most of that is not being done by professional mathematicians.
220
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 03/10/2026
Perhaps, but neither is "just a stochastic parrot". The main reason why JASP is not helpful is that it smuggles in part of the syllogism, which is "stochastic parrots can't possibly think/reason/be original/creative, etc." It essentially assumes what it wants to prove.
100
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 03/10/2026
Typical Bender disciple response: flat assertion, no evidence provided.
010
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 03/10/2026
Based on some of the truly insane pushback I've gotten describing what's happening in mathematics with LLM's, I'd say "disciples" is pretty accurate.
110
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 03/10/2026
I think the "Stochastic Parrot" label is simplistic, dumb, and misleading, but I definitely agree that SSPH is false. Ask any mathematician who has been paying attention about LLM's solving problems that defeated the best mathematicians for decades, and they know it's false, too.
000
Reposted by Jeffrey Shallit 🇺🇦
Musa Okwonga @okwonga.bsky.social · 02/10/2026
This can't be talked about enough. "President Donald Trump's move to cut most of the US funding towards foreign humanitarian aid could cause more than 14 million additional deaths by 2030, according to research published in The Lancet medical journal." www.bbc.com/news/article...
bbc.com
Trump's aid cut risks causing 14 million deaths, report finds
The figure from the Lancet, a leading medical journal, includes 700,000 child deaths globally a year.
9229148
Reposted by Jeffrey Shallit 🇺🇦
Dave Levitan @davelevitan.bsky.social · 02/10/2026
Plenty going on of course but the admin deciding to accept essentially NO refugees except for thousands of white South Africans is among the most batshit racist things the US government has done in a century
518941
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 01/10/2026
The most unprincipled person ever to hold the office of VP. A garbage human.
030
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 30/09/2026
You're talking about law, whereas I was asking how the OP was so certain the systems didn't have or couldn't possibly have the attributes he claimed they didn't. If a 7-state busy beaver TM can have unpredictable behavior, imagine what thousands of smart collaborating agents let loose could do.
000
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 30/09/2026
Since I wasn't addressing the complex issue of responsibility, you've largely missed the point.
100
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 30/09/2026
What test did you or could you do to establish whether, in fact, it *does* have any of those things? You certainly know that complicated systems can produce behavior that their creators could not or did not anticipate.
100
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 30/09/2026
It certainly is, and you provided no reason to believe otherwise. You seem rather incompetent.
100
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 30/09/2026
Yet another person who completely misrepresents my position, and yet believes they somehow have the moral high ground.
100
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 29/09/2026
What a piece of Schmitt
0160
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 29/09/2026
This guy here is a very nice example of someone who has no counter-arguments, and embarrassed by his inability to support his particular bias, turns to insults instead. Muted.
100
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 29/09/2026
If they're simply watching, they're almost certainly not learning that much.
000
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 29/09/2026
This reflexive dismissal of any new application of LLMs is really embarrassing for you. That's the nicest thing I can say in response.
210
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 29/09/2026
You don't even understand the point I am making. It's not anthropomorphism at all. Such a waste of time. Muting.
200
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 29/09/2026
It's pretty appropriate that Trump wants to change the initialism for the names of the Great Lakes from "HOMES" to "SHAME".
020
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 29/09/2026
Nothing new about this at all. Remember Newton? "If I have seen further, it is by standing on the shoulders of giants". I don't understand why you think this is a criticism or insightful. Every achievement of AI will be dismissed like this. It's boring.
100
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 28/09/2026
You don't even understand the point! The point is not all chatbots are equal. Unless you say specifically which model you're talking about, you're missing the entire point.
000
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 28/09/2026
How would that help resolve whether the results are new or not? I'm an expert in the field; the results are new.
110
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 28/09/2026
For the math problems I am specifically talking about, there was no prior work to steal.
000
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 28/09/2026
Again, you have zero idea of what you're talking about. You're still in the world of Chatbots last year.
100
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 28/09/2026
You're simply wrong when it comes to mathematics. For example, the ideas in this paper weren't "brute-forced". arxiv.org/abs/2609.20366
arxiv.org
Brown-Gerver-Ramsey Theorems in Small Dimensions
We consider infinite walks in $\mathbb{N}^k$ with standard unit basis vector steps that avoid $t$ collinear points, and show that these walks exist for $(k,t) \in \{(6,3), (4,4), (3,7)\}$. In particul...
010
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 28/09/2026
It generates "new stuff" all the time, you blithering goofball. Here's just one example from my own work. arxiv.org/abs/2609.20366
arxiv.org
Brown-Gerver-Ramsey Theorems in Small Dimensions
We consider infinite walks in $\mathbb{N}^k$ with standard unit basis vector steps that avoid $t$ collinear points, and show that these walks exist for $(k,t) \in \{(6,3), (4,4), (3,7)\}$. In particul...
110
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 28/09/2026
You have zero idea of what you're talking about.
100
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 28/09/2026
It doesn't "steal" anyone's work, you silly goofball. If it had, the problems would have already been solved by others. And of course GPT Astra 6 cites relevant work, if you just ask it to. So many people commenting who have zero idea of what's going on.
100
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 27/09/2026
I think it would huge if you could show that a worm's brain has a model of that same worm in its environment. Have at it, no doubt a Nobel awaits.
100
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 27/09/2026
All things being equal, I definitely would prefer to vote for a candidate who didn't attend prayer breakfasts. But I can't always get what I want.
000
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 27/09/2026
That's ridiculously false, at least in the case of mathematics. GPT 6 Astra can find proofs for problems that defied the best human mathematicians for decades. It doesn't "pass" information "as their own" because you can ask it to document its sources and it tells you. So much misinformation!
551
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 27/09/2026
Again, it's a continuum. A worm probably has a much less detailed model of the world than a human has, not big enough to include themselves in the model. Mammals that pass the Gordon Gallup mirror test probably have a much more detailed model of the world that includes themselves.
100
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 27/09/2026
A pretty weak defense. It explains nothing. What specifically does it add to our understanding? Suppose you say take Nagel's definition and look at an embodied robot. I say, "There is something it is like to be that robot", and "you say, no, there's not". How could we possibly resolve it?
210
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 27/09/2026
No, that's not what it means. This is the typical response of people who simply have no understanding of the revolution in mathematics that's taking place, despite the fact that it's all mathematicians are talking about these days online. Educate yourself, you don't have to be ignorant forever.
400
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 27/09/2026
The physical environment of a physical object or organism that it inhabits and interacts with. For an LLM running on a compute, this might be, for example, understanding that it's on the Earth, that the Earth is inhabited by living beings, and so forth.
100
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 27/09/2026
You sound skeptical. But if so, you're wrong. (Love being corrected by people who have simply no idea what's happening.)
300
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 27/09/2026
Meanwhile, mathematicians: it's completely amazing, we're getting solutions to problems that have been open for decades.
210
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 27/09/2026
Lots of people, what the heck is your point?
000
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 27/09/2026
What's the precise difference between "thinking" and "simulated thinking"? I keep asking philosophers and other folks for a coherent explanation of the difference but they never provide one.
000
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 27/09/2026
Why does it have to be the "same mechanism"? If there's 1 lesson that we should draw from the work of Turing and others, it's that the underlying mechanism by which computation is carried out isn't essential. You can build an adder out of springs and plastic parts, or silicon chips. They both add.
000
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 27/09/2026
Not when philosophers can't agree on *anything*, no. survey2020.philpeople.org Better to ignore the entirely and study neuroscience, linguistics, and the theory of computation.
survey2020.philpeople.org
PhilPapers Survey 2020
This is a survey of professional philosophers in the English-speaking world and others concerning their views on some central philosophical questions, following up on the 2009 PhilPapers Survey with a...
210
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 27/09/2026
Yes, that's right, LLM's are solving hard math problems that defeated the world's brightest minds for decades, but they're doing that not by "thinking" and "reasoning". They are doing it by "floofing" instead, undefined. You're so smart!
100
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 27/09/2026
Do you have anything other than self-published screeds one could read; for example, published in refereed journals?
000
Jeffrey Shallit 🇺🇦 @shallit.bsky.social · 27/09/2026
Well, let's focus on one specific issue, your praise of Nagel. Do you think has definition of conscious experience as "there is something it is like to be that" thing is testable, useful, adds anything to our understanding, or is generally more than babble?
110