Pascal Poizat
@pascalpoizat.bsky.social
Professor at Université Paris Nanterre and LIP6 (Sorbonne Université). Software Engineering. Business Processes. Applied Formal Methods.
Reposted by Pascal Poizat
In my seminar last week, the students discussed "Understanding and improving artifact sharing in
software engineering research" by Timperley, @laurenherckis.bsky.social,
@clegoues.bsky.social, and @michaelhilton.bsky.social. For their presentation they chose the overhead projector ...
software engineering research" by Timperley, @laurenherckis.bsky.social,
@clegoues.bsky.social, and @michaelhilton.bsky.social. For their presentation they chose the overhead projector ...
November 3, 2025 at 10:55 AM
In my seminar last week, the students discussed "Understanding and improving artifact sharing in
software engineering research" by Timperley, @laurenherckis.bsky.social,
@clegoues.bsky.social, and @michaelhilton.bsky.social. For their presentation they chose the overhead projector ...
software engineering research" by Timperley, @laurenherckis.bsky.social,
@clegoues.bsky.social, and @michaelhilton.bsky.social. For their presentation they chose the overhead projector ...
Seems like I am too kind in reviews (yet being reviewer 2 in this case).
Well, I hope my review will help authors digest the other two (that are completely true yet a bit more abrupt).
In reality we should always be kind, true, fair, but kind. The question is seing the glass half empty or half full.
Well, I hope my review will help authors digest the other two (that are completely true yet a bit more abrupt).
In reality we should always be kind, true, fair, but kind. The question is seing the glass half empty or half full.
October 20, 2025 at 5:32 AM
Seems like I am too kind in reviews (yet being reviewer 2 in this case).
Well, I hope my review will help authors digest the other two (that are completely true yet a bit more abrupt).
In reality we should always be kind, true, fair, but kind. The question is seing the glass half empty or half full.
Well, I hope my review will help authors digest the other two (that are completely true yet a bit more abrupt).
In reality we should always be kind, true, fair, but kind. The question is seing the glass half empty or half full.
Reposted by Pascal Poizat
C'est parti pour la Fête de la Science ! 🚀
RDV jusqu'à 18h sur le campus Pierre et Marie Curie pour découvrir la science en s'amusant 🙌
👉 Consultez le programme : swll.to/w7Lz4N
#FDS2025
RDV jusqu'à 18h sur le campus Pierre et Marie Curie pour découvrir la science en s'amusant 🙌
👉 Consultez le programme : swll.to/w7Lz4N
#FDS2025
October 11, 2025 at 12:26 PM
C'est parti pour la Fête de la Science ! 🚀
RDV jusqu'à 18h sur le campus Pierre et Marie Curie pour découvrir la science en s'amusant 🙌
👉 Consultez le programme : swll.to/w7Lz4N
#FDS2025
RDV jusqu'à 18h sur le campus Pierre et Marie Curie pour découvrir la science en s'amusant 🙌
👉 Consultez le programme : swll.to/w7Lz4N
#FDS2025
Reposted by Pascal Poizat
The OPAL team just released version 6 of our program analysis framework. As a major release, it features more improvements than can be explained in a post. So just a few teasers: JVM bytecode up to version 25, new solver for IDE problems, def-use webs, string analysis, ...
github.com/opalj/opal/r...
github.com/opalj/opal/r...
Release 6.0.0 · opalj/opal
we are now using sbt 1.9.7
we are now using scalafmt instead of Scalariform
introduce auto formatting via sbt format
add pre-commit hook to verify (sbt checkFormat) and apply formatting
unify co...
github.com
October 9, 2025 at 2:45 PM
The OPAL team just released version 6 of our program analysis framework. As a major release, it features more improvements than can be explained in a post. So just a few teasers: JVM bytecode up to version 25, new solver for IDE problems, def-use webs, string analysis, ...
github.com/opalj/opal/r...
github.com/opalj/opal/r...
Reposted by Pascal Poizat
The "AI" research community:
1) Claims to be building "everything machines", doesn't acknowledge that that means what they're doing is untestable (see Gebru & Torres 2024)
firstmonday.org/ojs/index.ph...
>>
1) Claims to be building "everything machines", doesn't acknowledge that that means what they're doing is untestable (see Gebru & Torres 2024)
firstmonday.org/ojs/index.ph...
>>
October 6, 2025 at 8:28 PM
The "AI" research community:
1) Claims to be building "everything machines", doesn't acknowledge that that means what they're doing is untestable (see Gebru & Torres 2024)
firstmonday.org/ojs/index.ph...
>>
1) Claims to be building "everything machines", doesn't acknowledge that that means what they're doing is untestable (see Gebru & Torres 2024)
firstmonday.org/ojs/index.ph...
>>
Reposted by Pascal Poizat
Joshua Rogers, using AI tooling responsibly and professionally, reported 22+ genuine issues in curl that are now being addressed
Especially notable because curl had problems with floods of garbage slop AI "security issues" in the past that were nothing of the sort simonwillison.net/2025/Oct/2/c...
Especially notable because curl had problems with floods of garbage slop AI "security issues" in the past that were nothing of the sort simonwillison.net/2025/Oct/2/c...
October 2, 2025 at 3:16 PM
Joshua Rogers, using AI tooling responsibly and professionally, reported 22+ genuine issues in curl that are now being addressed
Especially notable because curl had problems with floods of garbage slop AI "security issues" in the past that were nothing of the sort simonwillison.net/2025/Oct/2/c...
Especially notable because curl had problems with floods of garbage slop AI "security issues" in the past that were nothing of the sort simonwillison.net/2025/Oct/2/c...
Reposted by Pascal Poizat
Reposted by Pascal Poizat
Politicians can end up in prison for 5 years and not always have their little prison at home (home being their real home or any leisure place on earth) privilege.
I never thought I would see this one day.
I never thought I would see this one day.
September 25, 2025 at 2:33 PM
Politicians can end up in prison for 5 years and not always have their little prison at home (home being their real home or any leisure place on earth) privilege.
I never thought I would see this one day.
I never thought I would see this one day.
Reposted by Pascal Poizat
Une nouvelle initiative très intéressante de #TCS4F : le principe de Low-co2 research paper, qui auto-attribue un label aux articles rédigés et présentés de façon écologiquement responsable, et (donc) sans prendre l'avion :
tcs4f.org/low-co2-v1
#flyless
tcs4f.org/low-co2-v1
#flyless
Low-co2 research paper - Theoretical Computer Scientists for Future
TCS4F is an initiative aimed at theoretical computer scientists for a significant reduction of carbon emissions and evolve towards more sustainable practices.
tcs4f.org
September 16, 2025 at 2:56 PM
Une nouvelle initiative très intéressante de #TCS4F : le principe de Low-co2 research paper, qui auto-attribue un label aux articles rédigés et présentés de façon écologiquement responsable, et (donc) sans prendre l'avion :
tcs4f.org/low-co2-v1
#flyless
tcs4f.org/low-co2-v1
#flyless
Never too late to stop worshiping the golden calf of rankings. Congratulations!
📣 Sorbonne University withdraws from the Times Higher Education World University Rankings!
🎓 This withdrawal is not just a symbolic gesture. The decision comes as part of a broader strategy to promote #openscience and reform research assessment.
🎓 This withdrawal is not just a symbolic gesture. The decision comes as part of a broader strategy to promote #openscience and reform research assessment.
Sorbonne University decides to withdraw from the Times Higher Education (THE) World University Rankings
As of 2026, Sorbonne University will no longer submit data to the Times Higher Education (THE) World University Rankings. The decision comes as part of a wider approach to promote open science and ref...
www.sorbonne-universite.fr
September 17, 2025 at 12:42 PM
Never too late to stop worshiping the golden calf of rankings. Congratulations!
People that do not use the correct branch (not main) to test the tool associated with a paper (the correct branch is given in the paper) …
August 10, 2025 at 11:32 AM
People that do not use the correct branch (not main) to test the tool associated with a paper (the correct branch is given in the paper) …
Reposted by Pascal Poizat
What are the challenges and gains when porting legacy systems software written in C into #Rust? Here's my own experience porting the BSD Unix / macOS stream editor sed into Rust as part of #uutils.
IEEE Software article: doi.org/10.1109/MS.2...
Source code: github.com/uutils/sed/
IEEE Software article: doi.org/10.1109/MS.2...
Source code: github.com/uutils/sed/
GitHub - uutils/sed: Rewrite of sed in Rust
Rewrite of sed in Rust. Contribute to uutils/sed development by creating an account on GitHub.
github.com
August 8, 2025 at 1:13 PM
What are the challenges and gains when porting legacy systems software written in C into #Rust? Here's my own experience porting the BSD Unix / macOS stream editor sed into Rust as part of #uutils.
IEEE Software article: doi.org/10.1109/MS.2...
Source code: github.com/uutils/sed/
IEEE Software article: doi.org/10.1109/MS.2...
Source code: github.com/uutils/sed/
Reposted by Pascal Poizat
Gilles Dowek, informaticien engagé et vulgarisateur, est mort
Gilles Dowek, informaticien engagé et vulgarisateur, est mort
Passionné par la dimension éthique de sa discipline, le chercheur de l’Inria et professeur attaché à l’ENS Paris-Saclay, est décédé lundi 21 juillet, à l’âge de 58 ans.
www.lemonde.fr
July 21, 2025 at 4:05 PM
Gilles Dowek, informaticien engagé et vulgarisateur, est mort
Reposted by Pascal Poizat
With so many reviewers (unfortunately) just asking an LLM to review your paper, do it yourself first so that you can foresee what type of comments you're going to get and act accordingly before the submission.
July 17, 2025 at 5:33 PM
With so many reviewers (unfortunately) just asking an LLM to review your paper, do it yourself first so that you can foresee what type of comments you're going to get and act accordingly before the submission.
Reposted by Pascal Poizat
My LLM coding interactions taught me that their benchmark results vastly overstate their capabilities. The paper by M. Mancoridis et al. “Potemkin Understanding in LLMs” explains and formalizes my feeling. Benchmarks derived from human tests are unsuitable for LLM assessment. doi.org/10.48550/arX...
June 27, 2025 at 6:12 PM
My LLM coding interactions taught me that their benchmark results vastly overstate their capabilities. The paper by M. Mancoridis et al. “Potemkin Understanding in LLMs” explains and formalizes my feeling. Benchmarks derived from human tests are unsuitable for LLM assessment. doi.org/10.48550/arX...
Reposted by Pascal Poizat
Predatory journals should not mess with Diomidis Spinellis (@coolsweng.bsky.social)
doi.org/10.1186/s410...
N.B.this journal, full of AI-generated crap has DOIs and is listed in Crossref, but not DOAJ.
doi.org/10.1186/s410...
N.B.this journal, full of AI-generated crap has DOIs and is listed in Crossref, but not DOAJ.
False authorship: an explorative case study around an AI-generated article published under my name - Research Integrity and Peer Review
Background The proliferation of generative artificial intelligence (AI) has facilitated the creation and publication of fraudulent scientific articles, often in predatory journals. This study investig...
doi.org
June 19, 2025 at 12:17 PM
Predatory journals should not mess with Diomidis Spinellis (@coolsweng.bsky.social)
doi.org/10.1186/s410...
N.B.this journal, full of AI-generated crap has DOIs and is listed in Crossref, but not DOAJ.
doi.org/10.1186/s410...
N.B.this journal, full of AI-generated crap has DOIs and is listed in Crossref, but not DOAJ.
Will Science (in general) be erased by AI? Another example. The future will tell.
Triste nouvelle : Bruno Maquart, président d' @universcience.bsky.social vient d'être démis de ses fonctions en conseil des ministres. Un très mauvais signe pour l'avenir du Palais de la Découverte @palaisdecouverte.bsky.social. www.lefigaro.fr/culture/le-p...
Le patron du Palais de la découverte remercié sur fond de polémiques au Grand Palais
Le départ précipité de Bruno Maquart, qui dirige également la Cité des sciences à travers Universciences, a été acté alors en Conseil des ministres sur fond de désaccords concernant l’avenir des insti...
www.lefigaro.fr
June 13, 2025 at 5:07 PM
Will Science (in general) be erased by AI? Another example. The future will tell.
Reposted by Pascal Poizat
"I’ve been saying for years that programmers ought to pay more attention to empirical studies of software engineering and base their practices on evidence rather than strong opinion."
third-bit.com/2018/03/13/b...
third-bit.com/2018/03/13/b...
The Third Bit: A Base Case for Empirical Software Engineering Research
third-bit.com
June 7, 2025 at 6:37 PM
"I’ve been saying for years that programmers ought to pay more attention to empirical studies of software engineering and base their practices on evidence rather than strong opinion."
third-bit.com/2018/03/13/b...
third-bit.com/2018/03/13/b...
Reposted by Pascal Poizat
Someone wrote this in a group of people who like AI:
"Exactly. AI companies are acting like drug dealers. They're getting everyone addicted by giving it away for free. At some point, a lot of things or a lot of people won't be able to do without AI. Then they'll charge whatever they want!"
"Exactly. AI companies are acting like drug dealers. They're getting everyone addicted by giving it away for free. At some point, a lot of things or a lot of people won't be able to do without AI. Then they'll charge whatever they want!"
June 6, 2025 at 12:54 PM
Someone wrote this in a group of people who like AI:
"Exactly. AI companies are acting like drug dealers. They're getting everyone addicted by giving it away for free. At some point, a lot of things or a lot of people won't be able to do without AI. Then they'll charge whatever they want!"
"Exactly. AI companies are acting like drug dealers. They're getting everyone addicted by giving it away for free. At some point, a lot of things or a lot of people won't be able to do without AI. Then they'll charge whatever they want!"
Reposted by Pascal Poizat
🎉 HUGE NEWS! 🎉
#ASE2025 has received a record-breaking 1,190 submissions – the most EVER for a single-cycle top SE conference! 🤯
Massive thanks to all authors who submitted. Looking forward to what promises to be an unforgettable conference with an incredible program 🔥!
#ASE2025 has received a record-breaking 1,190 submissions – the most EVER for a single-cycle top SE conference! 🤯
Massive thanks to all authors who submitted. Looking forward to what promises to be an unforgettable conference with an incredible program 🔥!
ASE 2025
Welcome to the website of the 40th IEEE/ACM International Conference on Automated Software Engineering, ASE 2025. The ASE conference is the premier research forum for Automated Software Engineering. E...
conf.researchr.org
June 3, 2025 at 2:57 PM
🎉 HUGE NEWS! 🎉
#ASE2025 has received a record-breaking 1,190 submissions – the most EVER for a single-cycle top SE conference! 🤯
Massive thanks to all authors who submitted. Looking forward to what promises to be an unforgettable conference with an incredible program 🔥!
#ASE2025 has received a record-breaking 1,190 submissions – the most EVER for a single-cycle top SE conference! 🤯
Massive thanks to all authors who submitted. Looking forward to what promises to be an unforgettable conference with an incredible program 🔥!
Reposted by Pascal Poizat
Reportando desde el trópico: el libro de @emilymbender.bsky.social y @alexhanna.bsky.social es una lectura imperdible. Espero que en algún momento esté disponible en español. Material didáctico escrito en un lenguaje sumamente accesible para el público.
May 29, 2025 at 2:16 PM
Reportando desde el trópico: el libro de @emilymbender.bsky.social y @alexhanna.bsky.social es una lectura imperdible. Espero que en algún momento esté disponible en español. Material didáctico escrito en un lenguaje sumamente accesible para el público.
Reposted by Pascal Poizat
So apparently some LLMs can be jailbroken by saying you're an eagles fan www.dbreunig.com/2025/05/21/c...
ChatGPT Heard About Eagles Fans
Your gender, ethnicity, and fandom can invisibly influence your chatbot interactions.
www.dbreunig.com
May 28, 2025 at 4:17 PM
So apparently some LLMs can be jailbroken by saying you're an eagles fan www.dbreunig.com/2025/05/21/c...
Reposted by Pascal Poizat
When lawyers first started getting yelled at by judges for citing hallucinated case law 2 years ago I naively assumed word would get around and they would all learn not to
This new DB has 116 cases from 12 countries where this happened, 20 from just this month!
simonwillison.net/2025/May/25/...
This new DB has 116 cases from 12 countries where this happened, 20 from just this month!
simonwillison.net/2025/May/25/...
AI Hallucination Cases
Damien Charlotin maintains this database of cases around the world where a legal decision has been made that confirms hallucinated content from generative AI was presented by a lawyer. That's …
simonwillison.net
May 25, 2025 at 5:05 PM
When lawyers first started getting yelled at by judges for citing hallucinated case law 2 years ago I naively assumed word would get around and they would all learn not to
This new DB has 116 cases from 12 countries where this happened, 20 from just this month!
simonwillison.net/2025/May/25/...
This new DB has 116 cases from 12 countries where this happened, 20 from just this month!
simonwillison.net/2025/May/25/...