Toby Murray
banner
tobycmurray.bsky.social
Toby Murray
@tobycmurray.bsky.social
Professor at University of Melbourne and School of Computing and Information Systems cyber lead; Director @dsi-vic.bsky.social; Oxford DPhil (@compscioxford.bsky.social; @hertfordcollege.bsky.social). Cyber, verification, etc. He/him
Reposted by Toby Murray
call me a thursday night pub special, cause I'm a well done angus
February 13, 2026 at 1:50 AM
I’ve never been a Gates fanboy but I note that the passion project of the tech bloke who was the world’s richest man when I was a teen was about saving millions of lives, not ending them.
This is on Musk and he should be held accountable.
HORRIFYING -
Dismantling USAID & cancelling >80% of aid caused ~750,000 deaths by mid-2025

🧸Most were children
⌛️ Many died within the first year

😰 Preventable deaths were mainly from:
- HIV, TB, malaria
- Maternal & fetal deaths
- Malnutrition

news.westernu.ca/2026/01/u-s-...
February 10, 2026 at 12:05 PM
Reposted by Toby Murray
The robot apocalypse hasn't happened yet, but still I can't escape the feeling that something has gone horribly wrong... Cartoon for Dutch newspaper @trouw.nl.

More of my work for Trouw: www.trouw.nl/cartoons/tje...

#ArtificialIntelligence #creativity #work #GenerativeAI
February 2, 2026 at 7:38 AM
Reposted by Toby Murray
Thanks for the thoughts! I read that study mostly through the angle of education, and to me the key takeaway (explicitly made by the authors) is that early AI use when coding prevents the acquisition of the very skills later needed for effective use of AI in coding. (Cf. Breakdown of participants...
January 31, 2026 at 8:02 AM
Lots of half-baked commentary about Anthropic’s Jan 28 paper about AI coding assistants arxiv.org/pdf/2601.20245 . Firstly it is not primarily about productivity. But how AI use affects on-the-job learning. A thread of observations 🧵
arxiv.org
January 31, 2026 at 7:39 AM
Me 5 years ago: “people believe wrong things because they lack education. if only people had access to the *facts*”

Me today: “facts mean nothing against tribal loyalties and storytelling—even Grok knows the score”

(screenshot from X)
January 26, 2026 at 9:50 PM
Code and data files (downloaded from the CORE conference portal) available here: gist.github.com/tobycmurray/...
January 23, 2026 at 6:53 AM
Forget Fortune’s paywall. Read the original report here gptzero.me/news/neurips/
January 21, 2026 at 11:00 PM
New ICORE conference rankings are out. Theory, logic and formal methods seem to have been heavily penalised. I vibe coded an analysis of, for each Field of Research (FOR), how many associated venues increased or decreased in rank. Here's the summary. FOR 4613 had 13 venues decrease their rank! 1/2
January 21, 2026 at 6:43 AM
An implication of this work seems to be that using these LLMs in a “therapy-like context” made them *more prone* to harmful behaviour. x.com/anthropicai/...
x.com
January 20, 2026 at 4:31 AM
Reposted by Toby Murray
Hi @bsky.app @pfrazee.com - is it possible that we could have Trending topics for countries other than the US? Australia would certainly thank you and likely the UK and Canada as well !
January 19, 2026 at 11:16 PM
It turns out Sydney’s, ahem, crap sewerage system is behind the spate of shark attacks there in the last couple days. Pumping untreated sewerage into the harbour not only makes the sea gross after heavy rain, but also has the, err, flow-on effect of attracting sharks www.abc.net.au/news/2026-01...
'There’ll be more': Shark warning after three attacks in Sydney in 26 hours
A public policy shark expert says sewage leaking into the harbour after heavy rainfall is attracting bait fish followed by sharks.
www.abc.net.au
January 19, 2026 at 9:31 PM
Reposted by Toby Murray
In January 1976 I started work on a Masters thesis in Machine Learning in the math dept at Flinders University in Adelaide, South Australia (that dept's most famous graduate, Terence Tao, was only six months old). I have officially been full time in AI for fifty years now.
January 18, 2026 at 11:29 PM
I’d argue that the rise of AI has revealed to those of us who teach that, just perhaps, it was us who were cheating ourselves all along, that we were testing for understanding and not merely task proficiency. As educators it is our job to distinguish those who understand from those who think they do
The risk of AI for education is not students cheating in exams, it is people in general cheating themselves into believing they understand things they don’t.
January 17, 2026 at 3:34 AM
Reposted by Toby Murray
AI and the Corporate Capture of Knowledge

More than a decade after Aaron Swartz's death, the United States is still living inside the contradiction that destroyed him. Swartz believed that knowledge, especially publicly funded knowledge, should be freely accessible. Acting on that, he downloaded…
AI and the Corporate Capture of Knowledge
More than a decade after Aaron Swartz's death, the United States is still living inside the contradiction that destroyed him. Swartz believed that knowledge, especially publicly funded knowledge, should be freely accessible. Acting on that, he downloaded thousands of academic articles from the JSTOR archive with the intention of making them publicly available. For this, the federal government charged him with a felony and threatened decades in prison.
www.schneier.com
January 16, 2026 at 2:46 PM
Can’t imagine a better dickhead detector than a generative AI model allowing nudification and invoked by (public) tweet, and whose results are also published tweets
January 15, 2026 at 3:01 AM
This is such a wicked problem but one that lawmakers absolutely need to grapple with
I have an op-ed in the NYT today about the Grok scandal, sharing my research from last year finding that legal risk hinders AI companies from making their models safer against CSAM - an echo of the years where white-hat hackers were chilled from good-faith research. www.nytimes.com/2026/01/12/o...
Opinion | There’s One Easy Solution to the A.I. Porn Problem
www.nytimes.com
January 13, 2026 at 12:24 AM
“It’s almost like you succeed because of your passion rather than because of your confidence.” - Prof Emma Johnston AO, FAAS.

Vale. A shocking loss of an inspiring leader.

www.unimelb.edu.au/newsroom/new...
Notice of the passing of the University of Melbourne Vice-Chancellor, Professor Emma Johnston AO
The University of Melbourne advises with deep sadness of the passing of Vice-Chancellor, Professor Emma Johnston AO, due to complications associated with cancer.
www.unimelb.edu.au
December 29, 2025 at 1:26 AM
Reposted by Toby Murray
📢 Calling Aussie #cybersecurity companies

Want to go to #RSAC2026 as part of the AusTrade delegation in March?

EoIs close 21 December so don't miss out

https://bit.ly/4iYVcbn
December 16, 2025 at 3:54 AM
Reposted by Toby Murray
To the beautiful people of my former home of North #Bondi, Australia holds you in its broken heart tonight. In Bondi we always felt so safe, so alive. It’s devastating that that was destroyed today. To the Jewish community, always so central to the heart of Bondi, we are so incredibly sorry.
December 14, 2025 at 1:28 PM
Reposted by Toby Murray
Thinking of all the colleagues and friends I have who are connected to Brown University. What a devastating day for Brown, and for all of us.
December 14, 2025 at 3:19 AM
Let’s be real. There’s only one thing that matters: harm reduction. Let’s see social media platforms implement proper mechanisms to reduce harm. Brittle age checks are blunt policy against a complex kaleidoscope of cause and effect
I am sceptical about this ban. But it would indeed great if there were a scientific effort to measure the postlitive and negative effects to prove me wrong (or right) so that other countries can learn from Australia.
Does the Australian government have a measure of success it's tracking? It would be good for proponents of this law to set those out now, if they haven't already. 10/
December 10, 2025 at 8:01 AM
Reposted by Toby Murray
New on my blog:
50 years of proof assistants
lawrencecpaulson.github.io/2025/12/05/H...
50 years of proof assistants
lawrencecpaulson.github.io
December 5, 2025 at 8:37 PM
In Orwell’s 1984 “Nothing was your own except the few cubic centimetres inside your skull.”

But in a future in which human thought happens *in collaboration with Generative AI*, we risk the sanctity of those precious few centimetres—-unless we innovate now to prevent oncoming surveillance.
Invited talk by Zeynep @zey.bsky.social Tufekci at @neuripsconf.bsky.social was really very good!
Title: Are We Having the Wrong Nightmares About AI?
Spoiler: yep! 😉
Very thoughtful! Thanks!🙏
December 4, 2025 at 11:44 PM
Reposted by Toby Murray
not true
Few self-respecting scholars choose to become university administrators.

It is populated by swimmers in the shallow end of the intellectual pool, would-be and real fascists, and people of limited imagination.
November 30, 2025 at 7:32 AM