BEGIN:VCALENDAR
PRODID:-//Grails iCalendar plugin//NONSGML Grails iCalendar plugin//EN
VERSION:2.0
CALSCALE:GREGORIAN
METHOD:PUBLISH
BEGIN:VEVENT
DTSTAMP:20261121T185405Z
DTSTART;TZID=Europe/Paris:20261126T180000
DTEND;TZID=Europe/Paris:20261126T193000
SUMMARY:Is AI Already Improving Itself?
UID:20261121T185406Z-iCalPlugin-Grails@fe80:0:0:0:4c58:70ff:fefb:3038%3
TZID:Europe/Paris
LOCATION:6 rue Amyot\, Paris\, France
DESCRIPTION:<p><strong>Is AI Already Improving Itself?</strong><br><em>Agentic research\, recursive self-improvement\, and the limits of automation</em></p>\n<p>Can AI systems already contribute to improving their own capabilities? And how close are we to genuinely autonomous AI research?</p>\n<p>For this ai-phi Causerie\, <strong>Louis Laborelli (INA)</strong> will lead an open discussion on agentic AI\, recursive self-improvement\, and AI-assisted research. Drawing on recent developments at frontier AI labs\, empirical evaluations\, and concrete examples of automated ML optimisation\, we will examine what current systems can achieve\, where their limitations lie\, and what accelerating AI research might imply for safety and governance.</p>\n<p>Inspired by the tradition of Parisian intellectual salons\, this Causerie welcomes questions\, critical perspectives\, and contributions from all participants.</p>\n<p><strong>Thursday\, 26 November 2026\, 18:00</strong><br>Sony CSL Paris\, 6 rue Amyot\, 75005 Paris</p>\n<p>In person only\; registration required. The discussion will not be recorded.</p>\n<p><strong>Information and registration:</strong><a href="https://ai-phi.github.io/posts/session-43-laborelli-is-ai-already-improving-itself/">https://ai-phi.github.io/posts/session-43-laborelli-is-ai-already-improving-itself/</a></p>
ORGANIZER;CN="Aïda Elamrani":
METHOD:PUBLISH
END:VEVENT
END:VCALENDAR
