Lisp in AI, Research and Industry

Classic AI (1960s–1990s): Lisp was *the* AI language: SHRDLU, Macsyma, expert systems (MYCIN-era tools), Cyc, planning systems, and DARPA projects. Lisp Machines (Symbolics, LMI, TI Explorer, Xerox) were hardware built to run Lisp.

Industry today: Common Lisp: ITA Software/Google Flights (QPX), Grammarly (historically), Rigetti (quantum compilers: Quil), Ravenpack, Siscog (railway planning), the Mirai/Izware 3D tools, PTC Creo Elements/Direct (CAD), DWave, NASA projects (Deep Space 1 Remote Agent), and the game studio Naughty Dog (GOAL/GOOL, historically).

Clojure: Nubank (one of the largest Clojure users), Walmart Labs, Apple, Netflix, CircleCI, Metabase, Pitch, Funding Circle, Cisco.

Racket: education (HtDP), research (programming languages), and Naughty Dog's scripting language for later games.

Scheme: GNU Guix, GnuCash, LilyPond (music engraving, via Guile), TeXmacs, and Chez powering Racket CS.

Emacs Lisp: the entire Emacs ecosystem (Org mode, Magit).

Symbolic + neural: Lisp remains popular for symbolic reasoning, program synthesis, theorem proving (ACL2 is written in Common Lisp), and building DSLs around ML systems.