todo:
* thesis defense ✅
Grateful to the committee and reviewers Marius Kloft, @arkrause, Rupak Majumdar, and @tobigerstenberg for their time and support. No words are enough to thank my advisor @autreche for everything I’ve learned from him over the years. 🙏
📢 Are you a PhD student, postdoc, or early-career researcher working in AI?
✅ Reminder: the submission deadline for #GreeksInAI is in 5 days.
📍Eugenides Foundation, Athens
📅15–17 July 2026
Submit now, via: https://t.co/Ew8r8Ufmx9
This 👇 The best way to use AI today isn't to outsource your work to it, but to build (cheaper) tools that will simplify your work in a future where AI might be less accessible.
Something that people making AI economics predictions should take into account is that 1) current AI use is heavily subsidized, 2) demand for smart, compute-heavy models is exploding w/ few price signals to discipline use, 3) we will be hitting serious compute constraints soon.
Translation: We have been living through a period of abundant AI; get ready for the age of compute scarcity. There will be more rationing, multi-tiered markets, new token allocation infrastructure, etc.
Today NeurIPS is announcing our official satellite event in Paris.
After responding to the call from Ellis following the success of EurIPS in December, we are pleased to reach a new milestone by joining forces with the NeurIPS organizing committee for the 2026 edition!
🤖Thrilled to share the Call for Papers for Greeks in AI 2026
🇬🇷Greeks in AI 2026
📍Eugenides Foundation, Athens, Greece
📅15–17 July 2026
We invite submissions presenting recent advances, ongoing research, and emerging ideasi n Artificial Intelligence and Machine Learning
In my first year of my PhD, I measured progress by papers. The count was zero. I felt like an imposter with no sense of progress.
That was an illusion - I was measuring the wrong thing.
I see this happening frequently, so I wrote a post with what a learned.
URL below.
Come and join us at @AmsterdamNLP! We have a postdoc position with @KatiaShutova in #NLProc, focusing on multilingual NLP, continual learning and LLM alignment. Looking for recent PhD graduates with an NLP/ML background and an interest in AI alignment. https://t.co/s9hEKRXhaf
Check out our latest work https://t.co/ShjyfA32Ta, where we show the way @OpenAI, @AnthropicAI and others charge users for using their reasoning models incentivize them to increase test-time computations even if they contribute little to the quality of the outputs.
Done with #ICML2026? Setting up experiments for a new project can be boring... Here is a template that automates parameter configs, output filenames, SLURM etc. Works for me, hope it works for you!
https://t.co/mXl93SDWMW
P.S. Templates are greener than AI coding agents🌱
The serious aspect here is that having a paper at NeurIPS can significantly boost a young researcher's career. As a sign of respect to authors who actually cared about the content of their papers yet got rejected, @NeurIPSConf needs to withdraw papers with hallucinated content.
Okay so, we just found that over 50 papers published at @Neurips 2025 have AI hallucinations
I don't think people realize how bad the slop is right now
It's not just that researchers from @GoogleDeepMind, @Meta, @MIT, @Cambridge_Uni are using AI - they allowed LLMs to generate hallucinations in their papers and didn't notice at all.
It's insane that these made it through peer review👇
🎓 IMDEA Software is offering a PhD position in Safe and Verified Cyber-Physical Systems!
IMDEA Software invites applications for one #PhD position under the supervision of Kaushik Mallik.
🔹 Who should apply: Candidates with a Master’s degree in Computer Science (or related fields).
📍 Location: Madrid, Spain
🗓️ Application deadline: February 10th, 2026
🔗 Apply here: https://t.co/0OKV7UmrZZ
What if AI agents aren't here to replace us, but to facilitate our decisions? In a study with 1600 participants, we show that a human with action choices narrowed by an AI makes better sequential decisions than an AI or a human alone.
📜 https://t.co/9TsoppEFfZ
Our decision support system is underpinned by a single parameter (epsilon) that adaptively controls the level of human agency. The resulting cumulative reward varies smoothly with epsilon, which allows us to find the optimal level of human agency using a bandit algorithm.
While I’m sad to leave @mpi_sws_ and its people, it's time to move on. Starting October, I will be a postdoctoral researcher at @HPI_DE, working with @SandraWachter5. Super excited about this next chapter!
todo:
* thesis defense ✅
Grateful to the committee and reviewers Marius Kloft, @arkrause, Rupak Majumdar, and @tobigerstenberg for their time and support. No words are enough to thank my advisor @autreche for everything I’ve learned from him over the years. 🙏
Grateful to the tutorial chairs @ShalitUri and @sineadwilliamso for the opportunity, and big thanks to the PC chairs @saramagliacane and @csilviavr for being on the ground and making UAI 2025 such a unique experience!
Last week I had the pleasure of presenting a 2.5-hour tutorial on "Counterfactuals in Minds and Machines" at UAI 2025 in Rio 🇧🇷, prepared together with @autreche and @tobigerstenberg . We've made all materials and references available here: https://t.co/Gx1n5IVfc6
Heading to Rio de Janeiro 🇧🇷 for UAI 2025 (@UncertaintyInAI) to present our tutorial with @tobigerstenberg and @autreche on "Counterfactuals in Minds and Machines" on Monday. Looking forward to this! If you are in Rio, let's meet!
In Athens 🇬🇷 for the Greeks in AI symposium. Super excited to present our work on "Counterfactual Token Generation in LLMs" (https://t.co/F4wXZTlZcR) and see all the amazing work Greek people all over the world are doing on AI! If you are in Athens, let's meet! Next, heading to👇