Hacker News new | past | comments | ask | show | jobs | submit login

This is very cool. Congratulations!

How did you beat Peffers' team? Did you simply start the computation sooner, or did you reach some insight that eluded them?

No Simon Peffers' team did something more impressive: they're using a FPGA. I simply started way earlier using consumer-grade hardware (a Core i7-6700 CPU).

But you made the correct decision on whether to start the computation immediately or to start by building a "more impressive" system. Maybe it was luck that you beat them, or maybe it wasn't, but either way several years out you made the correct engineering decision on how to solve a problem with finite resources. That's impressive.

For those who didn't read the Wired article:

"In mid-March, the group began to run an algorithm" ... "FPGA was about 10 times faster than a high-end commercial CPU running non-optimized software. Based on the chip’s computing efficiency" they "calculated that they would have the correct solution" "on the evening of May 10, just two months after they started the calculation. Yet when they reached out to MIT to let them know a solution was imminent, Rivest informed them that Fabrot had beaten them to the punch."

We read here that he started end of 2015! Rivest believed at the time he constructed the puzzle, almost 20 years ago, that he has constructed the puzzle that would keep the vault locked until 2034, based on his expectations then of the increase of the computing speed. Obviously, predicting the developments of 35 years is not so easy, and the single-CPU computer speeds did improve faster. Anyway...


It's actually the algorithmic improvements that were faster, if I read everything here and in the article right. Not the raw single-CPU computer speed per se.

> It's actually the algorithmic improvements that were faster

I’d really like to know about those! Any links?

Disagree. I'd choose to work with you first on a project, but until facing such a nice problem congrats!

No disrespect intended to that team. It's an impressive accomplishment, and they're the ones who can publish in a journal for all due academic credit.

However, it also depends on what you value. Sometimes for some problems crossing the finish line first is winning, or seems able to simply allow the worlds technical advances to compound faster than more novel approaches whose future utility is unknown.

During all those years one person stopped to think about the goal in the most optimal way and ask, what's really the simplest solution here? Is custom hardware the only practical way? Everyone was allowed to think about implementation costs.

In other words, the most impressive part is what was not done. Great work.

A real-life Tortoise and the Hare story! (In the sense that slow and steady progress won the day. I do not wish to imply that Peffers team has any of the negative qualities of the hare.)

Kinda: a friend told me exactly that the other day. But we then discussed it: technically the hare started first and then rested, which is not what Simon and his team did : )

Registration is open for Startup School 2019. Classes start July 22nd.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact