10 years ago I started working on program synthesis, coz I was curious if I could build a system that would surprise me, on the tiniest of intellectual tasks.
We wrote a paper about it.
I’m glad the community found it useful. Humbled by the most influential paper award.
We are working on something new, release this Fall. Front-end engineers, UI/UX designers, production engineers, and product people. Come build with us.
Languages can avoid a few pitfalls, but eventually we need to be delegate code creation to synthesis. Human energy should be spent on specifying *what* we want the code to do.
Only way to build software in the future will be through automation.
We built a synthesizer + verifier for Move. Have one for Solidity too. Synthesis is the fastest way to build safe apps.
The financial infrastructure of the future will be built by automation not manual coding.
Language specifics matter little, as most programmers will agree. Even more so the case when its an automation that is writing the code or proving it correct.
Come to our conference on the "State-of-the-art in Program Synthesis", Sep 19-20 in SF. The entirely technical program has some of the most interesting results in the field. https://t.co/YLe65Jzp4e
Who says innovation in Silicon Valley is dead? 100 amazing, mostly deep tech companies in our @khoslaventures portfolio doing unusual things described in 25 minutes! Excited to share some of them in the hopes of motivating other entrepreneurs https://t.co/nRA46p6y7K
Alex @vidiborvc worked with us this winter. He learnt a few things about program synthesis in the process. Can the promise of program synthesis be realized to create real business value? His thoughts: https://t.co/YWsjYXCbFw
@rbhar90 2/4 Bearish on intellectually attractive, but developer unfriendly annotation-based *new* languages.
3/4 Bullish on languages that have easier on-ramp (e.g., JS derivatives) but infrastructure to permit heavy weight verification.