BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//cfp.nix.vegas//2026//talk//MVYVWQ
BEGIN:VTIMEZONE
TZID:PST
BEGIN:STANDARD
DTSTART:20001029T030000
RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10;UNTIL=20061029T100000Z
TZNAME:PST
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
END:STANDARD
BEGIN:STANDARD
DTSTART:20071104T030000
RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=11
TZNAME:PST
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20000402T030000
RRULE:FREQ=YEARLY;BYDAY=1SU;BYMONTH=4;UNTIL=20060402T110000Z
TZNAME:PDT
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
END:DAYLIGHT
BEGIN:DAYLIGHT
DTSTART:20070311T030000
RRULE:FREQ=YEARLY;BYDAY=2SU;BYMONTH=3
TZNAME:PDT
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
UID:pretalx-2026-MVYVWQ@cfp.nix.vegas
DTSTART;TZID=PST:20260808T113000
DTEND;TZID=PST:20260808T120000
DESCRIPTION:Reverse engineering binaries is slow\, detailed work\, but it's
  one of the few ways we have to verify what some software actually does. M
 achine learning might help improve some of those challenges\, but only if 
 you have high-quality\, real-world ground truth to train and evaluate on. 
 At MIT Lincoln Laboratory\, we have built a large dataset for machine lear
 ning on x86_64 binaries using nixpkgs\, taking advantage of Nix's reproduc
 ibility\, package coverage\, and instrumentable build environments.\nWe co
 mpiled tens of thousands of C\, C++\, Rust\, and Go packages\, often multi
 ple versions of each\, with consistent compiler flags\, captured source fi
 les\, and debug information. From these\, we extracted roughly 50 million 
 functions and aligned with their source code and short descriptions. Possi
 ble applications include fine-tuning large language models for various bin
 ary tasks or training custom\, highly efficient models for binary code und
 erstanding.\nI'll walk through what it takes to wrangle nixpkgs in an HPC 
 environment at this scale and show how open build tooling like Nix can hel
 p us to better understand and analyze the software that we all depend on.
DTSTAMP:20260725T021607Z
LOCATION:DEF CON Community Stage
SUMMARY:Compiling the World: A Binary Dataset Built from nixpkgs - Chris Co
 nnelly
URL:https://cfp.nix.vegas/2026/talk/MVYVWQ/
END:VEVENT
END:VCALENDAR