Sign in

Mirco Musolesi

@mircomusolesi.bsky.social
161 followers 159 following 96 posts

Professor of Computer Science. Machine Intelligence Lab, UCL AI Centre, Department of Computer Science, University College London. Artificial Intelligence | Machine Learning | Computing | Books www.mircomusolesi.org

PostsRepliesMedia
Mirco Musolesi @mircomusolesi.bsky.social · 29/09/2026
Probably one of the best tech stories of this month. The team behind an NPR podcast noticed a flood of baffling comments under their Spotify episodes, but they weren’t bots. techcrunch.com/2026/09/25/t...
techcrunch.com
The hottest new hangout for middle schoolers is NPR's comment section? | TechCrunch
When NPR staffers flagged strange comments under their podcasts on Spotify as bots, it took a Gen Z colleague to (immediately) figure out the mystery.
010
Mirco Musolesi @mircomusolesi.bsky.social · 28/09/2026
The closing sentences of Margaret Boden's "Creativity and Artificial Intelligence" (1998). "Only a few AI-models can critically judge their own original ideas. And hardly any can combine evaluation with transformation." www.sciencedirect.com/science/arti...
011
Mirco Musolesi @mircomusolesi.bsky.social · 01/09/2026
Professional ethics as a problem of AI decision-making: our new preprint with Omri Asscher examines how AI systems navigate conflicting ethical objectives in law, medicine, and translation. Link to the preprint: arxiv.org/abs/2608.29306
000
Mirco Musolesi @mircomusolesi.bsky.social · 26/08/2026
In particular, LLM-based judges exhibit a systematic preference for AI-generated stories, consistently favoring their stylistic characteristics over the unpredictability and other qualities of human authored texts. 4/4
010
Mirco Musolesi @mircomusolesi.bsky.social · 26/08/2026
We compare these judgments with automated objective metrics and LLM-as-a-Judge evaluations. Our experiments reveal substantial misalignment between automatic evaluations and human assessments. 3/4
110
Mirco Musolesi @mircomusolesi.bsky.social · 26/08/2026
In this paper, we nvestigate whether current automatic evaluation methods can reliably capture human judgments of creativity. We collect human evaluations of human- and AI-generated short stories from the WritingPrompts dataset across 11 dimensions of creativity. 2/4
110
Mirco Musolesi @mircomusolesi.bsky.social · 26/08/2026
New preprint: “The Limits of Automatic Evaluation of Creativity in Large Language Models” with Alessandro Tutone and Giorgio Franceschelli. Link to the paper: arxiv.org/abs/2608.23705 1/4
110
Mirco Musolesi @mircomusolesi.bsky.social · 26/08/2026
Not sure. The probability that someone will receive emails with "Dear Mirco Musolesi" is actually pretty low...
000
Mirco Musolesi @mircomusolesi.bsky.social · 26/08/2026
We proposed this point of view/explanation (presenting a simple mathematical model supported by some empirical evidence) in a 2015 paper led by Kai Zhao: www.nature.com/articles/sre...
nature.com
Explaining the power-law distribution of human mobility through transportationmodality decomposition - Scientific Reports
Scientific Reports - Explaining the power-law distribution of human mobility through transportationmodality decomposition
000
Mirco Musolesi @mircomusolesi.bsky.social · 25/08/2026
Please talk to me in French this morning.
120
Mirco Musolesi @mircomusolesi.bsky.social · 18/08/2026
If you are an Apple user, you probably received the message about updating to Tahoe 26.6.2 (just 10 days after the previous update). Looking at the description of the security fixes, I noticed that a large number of bugs are credited to the OpenAI Codex team. support.apple.com/en-gb/148282
support.apple.com
About the security content of iOS 26.6.1 and iPadOS 26.6.1 – Apple Support (UK)
This update delivers security fixes that were first made available in the iOS 27 and iPadOS 27 betas. This document describes the security content of iOS 26.6.1 and iPadOS 26.6.1.
000
Mirco Musolesi @mircomusolesi.bsky.social · 17/08/2026
And she also invented the term “software engineering” itself. en.wikipedia.org/wiki/Margare... 5/5
en.wikipedia.org
Margaret Hamilton (software engineer) - Wikipedia
011
Mirco Musolesi @mircomusolesi.bsky.social · 17/08/2026
It is my typical example to prove that essentially you don’t always need an Nvidia H100 GPU if you write proper code. In fact, among other things, she made a significant contribution to the software of the Apollo programme (which fit into 4KB RAM and 74KB ROM, but that is another story). 4/5
111
Mirco Musolesi @mircomusolesi.bsky.social · 17/08/2026
And another person I tend to mention often is Margaret Hamilton. I often point to her as the ultimate example of someone who wrote software to go to the Moon using the hardware equivalent of a Casio watch. 3/5
120
Mirco Musolesi @mircomusolesi.bsky.social · 17/08/2026
For example, I usually mention that the first serious AI engineer and possibly philosopher was Ada Lovelace (even though she was quite skeptical that machines could think creatively - see Turing’s comments in his 1950 Mind paper). 2/5
120
Mirco Musolesi @mircomusolesi.bsky.social · 17/08/2026
Just discovered that Margaret Hamilton is 90 today. I like the history of computing and in my lectures I tend to refer to the Greats of the past quite often. 1/5
121
Mirco Musolesi @mircomusolesi.bsky.social · 17/08/2026
Using dynamical systems theory, we analyse the stability, sensitivity, and inherent stochasticity of agent behaviour, which are key properties for safe and reliable deployment under strict safety requirements. 4/4
010
Mirco Musolesi @mircomusolesi.bsky.social · 17/08/2026
In this paper, Multi-Agent Reinforcement Learning (MARL) meets dynamical systems theory from an individual level perspective. We model MARL training as a coupled stochastic dynamical system, capturing both agent interactions and environmental dynamics. 3/4
120
Mirco Musolesi @mircomusolesi.bsky.social · 17/08/2026
Details of the talk: Tuesday 18 August 2026 Room B-0305 Track 4 11:40 AM – 12:30 PM Multi-agent RL session + Poster session at 3pm 2/4
120
Mirco Musolesi @mircomusolesi.bsky.social · 17/08/2026
Tomorrow, our paper "An Agent-Centric Dynamical Systems Perspective on Multi-Agent Reinforcement Learning" with James Rudd-Jones and Maria Perez-Ortiz will be presented at RLC 2026 (@rl-conference.bsky.social) in Montreal. Paper: rlj.cs.umass.edu/2026/papers/... 1/4 #RLC2026
2162
Mirco Musolesi @mircomusolesi.bsky.social · 14/08/2026
This is a very interesting collection of articles and posts about Python by Guido van Rossum himself, including lots of fascinating details about the origins and development of the language: gvanrossum.github.io/python-histo...
gvanrossum.github.io
The History of Python
011
Mirco Musolesi @mircomusolesi.bsky.social · 02/08/2026
The side effects of buying (too many) books. www.newyorker.com/humor/daily-...
newyorker.com
The Six Stages of Having Too Many Books
What happens when a mere shelf will not do.
000
Mirco Musolesi @mircomusolesi.bsky.social · 04/05/2026
Thanks for your question - I think this is context dependent (i.e., different applications will weight this type of novelty differently).
010
Mirco Musolesi @mircomusolesi.bsky.social · 27/04/2026
New preprint with Giorgio Franceschelli: "On the Creativity of AI Agents". Can AI agents be truly creative? In this paper, we introduce a conceptual framework for analysing agent creativity and outline key open questions in this fast-moving area. Link to the paper: arxiv.org/abs/2604.13242
162
Mirco Musolesi @mircomusolesi.bsky.social · 25/04/2026
Marta and her poster.
010
Mirco Musolesi @mircomusolesi.bsky.social · 25/04/2026
Our paper with Marta Emili Garcia Segura and Steve Hailes "Opponent Shaping in LLM Agents" is going to be presented this week at #ICLR2026. Saturday 25 April 2026 10:30-1:00pm BRT Pavillion 4 4618 Link to the paper: openreview.net/pdf?id=yJoHT...
162
Mirco Musolesi @mircomusolesi.bsky.social · 23/03/2026
These LLM-generated emails are getting good. This one summed up one of the strands of our current work perfectly and even included a very handy link for my own website.
010
Mirco Musolesi @mircomusolesi.bsky.social · 04/03/2026
The Art of Computer Prompting. www-cs-faculty.stanford.edu/~knuth/paper...
010
Mirco Musolesi @mircomusolesi.bsky.social · 03/02/2026
Teaching "Generative Artificial Intelligence" this term. Started today's class with an example celebrating our founder. #UCL200
010
Mirco Musolesi @mircomusolesi.bsky.social · 23/01/2026
New paper in Transactions on Machine Learning Research with Giorgio Franceschelli: "DiffSampling: Enhancing Diversity and Accuracy in Neural Text Generation". Paper+code+video here: tmlr.infinite-conf.org/paper_pages/...
000
Mirco Musolesi @mircomusolesi.bsky.social · 06/01/2026
More formally, the probability is uniformly distributed across all possible weather conditions.
000
Mirco Musolesi @mircomusolesi.bsky.social · 05/12/2025
The overall website about the OECD AI Capability Indicators project and policy work can be found here: aicapabilityindicators.oecd.org 6/6
aicapabilityindicators.oecd.org
OECD AI Capability Indicators
000
Mirco Musolesi @mircomusolesi.bsky.social · 05/12/2025
The OECD AI Capability Indicators Technical Report has now been published and is available here: www.oecd.org/en/publicati... 5/6
oecd.org
OECD AI Capability Indicators Technical Report
The AIFS’ team is publishing this technical volume to showcase the methodology used to develop the OECD AI Capability Indicators, released on June 3 2025. The volume also acknowledges and addresses in...
100
Mirco Musolesi @mircomusolesi.bsky.social · 05/12/2025
We focused on defining AI creative capabilities - not only in the artistic sense but, perhaps more importantly, in terms of problem-solving and original thinking. This work is based on Giorgio's PhD thesis and current projects. 4/6
100
Mirco Musolesi @mircomusolesi.bsky.social · 05/12/2025
In essence, the question is how to assess the extent to which an AI system can perform a given task with a level of proficiency comparable to that of humans. 3/6
100
Mirco Musolesi @mircomusolesi.bsky.social · 05/12/2025
The current goal of the Working Group is to devise a series of metrics to quantify AI capabilities on a scale from 1 to 5 (similar to the “autonomy” scale used for assessing the capabilities of self-driving cars) in order to derive a global standard for governments, businesses, and other orgs. 2/6
100
Mirco Musolesi @mircomusolesi.bsky.social · 05/12/2025
During the past year, our lab (in particular, Giorgio Franceschelli and myself) have been involved in the @oecd-ocde.bsky.social Working Group on AI and the Future of Skills. 1/6
100
Mirco Musolesi @mircomusolesi.bsky.social · 11/11/2025
Rightmove added the strapline “believe it” to its logo. The reference to what an estate agent tells you is clear (don’t).
000
Reposted by Mirco Musolesi
Randall Munroe @xkcd.com · 10/11/2025
Earthquake Prediction Flowchart xkcd.com/3165/
Comic. Someone is claiming to predict the exact date of a future earthquake. Should you listen? [flowchart: start] → NO (There are big earthquakes constantly, so if anyone ever *does* figure this out, it will be immediately obvious that their method works and then the world’s seismologists will not shut up about it. You won’t need this flowchart.)
263821514
Mirco Musolesi @mircomusolesi.bsky.social · 10/11/2025
This blogpost by Cal Newport is pretty good: "Forget Chatbots You Need a Notebook". calnewport.com/forget-chatb...
calnewport.com
Forget Chatbots. You Need a Notebook. - Cal Newport
Back in 2012, as a young assistant professor, I traveled to Berkeley to attend a wedding. On the first morning after we arrived, my wife ... Read more
000
Mirco Musolesi @mircomusolesi.bsky.social · 05/11/2025
In fact, these look like insects and maybe they look like that to attract predators that will be able to spread them elsewhere. I would be very interested in an opinion from an expert. 2/2
000
Mirco Musolesi @mircomusolesi.bsky.social · 05/11/2025
More wonders of the natural world. Now some flowers of the plant in my office appear as if suspended in the void. This looks like a “technique” to invite pollinators. Or maybe it is a self-pollinating plant and these flowers contain seeds (?). 1/2
100
Mirco Musolesi @mircomusolesi.bsky.social · 30/10/2025
The Singularity is Not Near.
000
Reposted by Mirco Musolesi
Randall Munroe @xkcd.com · 27/10/2025
Continents xkcd.com/3159/
Comic. [two connected spacecraft hovering in space above the Earth] Houston, the view is superb. We can see the continents spread out below us, right where they’ve been since the Earth formed. [caption] I still can’t believe we developed spaceflight before we figured out that the continents moved.
674595609
Mirco Musolesi @mircomusolesi.bsky.social · 21/10/2025
Link to the source: www.bbc.co.uk/news/live/c5...
bbc.co.uk
000
Mirco Musolesi @mircomusolesi.bsky.social · 21/10/2025
And, by the way, regarding the value of a Computer Science degree in the "vibe-coding" era: one thing you learn in a good CS programme is how to design and build systems that are resilient and fault-tolerant. 2/2
100
Mirco Musolesi @mircomusolesi.bsky.social · 21/10/2025
Re the AWS outage, it’s quite cool that the BBC interviewed Ken Birman, who knows a thing or two about distributed systems. 1/2
100
Mirco Musolesi @mircomusolesi.bsky.social · 16/10/2025
We received the ACM UbiComp 2025 10-Year Impact Award for our paper “Trajectories of Depression: Unobtrusive Monitoring of Depressive States by means of Smartphone Mobility Traces Analysis” co-authored with the great Luca Canzian. Paper here: www.mircomusolesi.org/papers/ubico... #ubicomp2025
010
Mirco Musolesi @mircomusolesi.bsky.social · 15/10/2025
New preprint with Charles Westphal and Stephen Hailes: "A Generalized Information Bottleneck Theory of Deep Learning". In this work, we introduce the Generalized Information Bottleneck framework, a synergy-based reformulation of the Information Bottleneck theory. arxiv.org/abs/2509.26327
010
Mirco Musolesi @mircomusolesi.bsky.social · 10/10/2025
Lots of headlines claim graduates are more likely to be unemployed than non-graduates. Graduate unemployment panic makes good headlines, but the data says otherwise. Great analysis by @jburnmurdoch.ft.com. www.ft.com/content/ea9e... .
ft.com
What the graduate unemployment story gets wrong
People with a degree are faring better, not worse than their non-graduate counterparts
020