Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Announcing Starfighter

kalzumeus.com

Post reply on HN
Print view View original

371–380 of 418 posts

Re: Announcing Starfighter

#371

Post by pbsd »

tom_b wrote: post #369
tptacek wrote: post #96

Earlier quoted context omitted.

That's definitely the ambition here, and it squares with our experience with the Crypto Challenges and Microcorruption: the "top tier" of people in both those events were, as a general rule, not very experienced in either of those problem domains.

Am I wrong to be completely stunned at this statement? Did experienced candidates just not participate in the Crypto Challenges and Microcorruption? I would have expected at least some experienced candidates to drop into that "top tier" of people. What's up with this?

I finished 4th in Microcorruption. While exploiting vulnerabilities is not my day job, I have been playing CTFs for a long time, and am familiar with the process. It is certainly something one can overspecialize in.

Looking at the first page of the Hall of Fame, I see big names like Alex Sotirov, Russ Cox, Ricky Zhou, Ludvig Strigeus, and many other familiar names/handles usually seen at CTF events. There are also many unknowns, which I suppose was Thomas's point.

Re: Announcing Starfighter

#372

Post by gsaines »

Very cool guys. As a long time HN reader, I have a lot of respect for what you guys are doing. I also spent the last two years of my life building CodeCombat, which started life with a very similar mission ("a game that finds developer hiring leads"). We were YC W14 and roundly failed at that business model. I just sent Patrick an email at his Kalzumeus email, but I'd be happy to chat with you guys about we learned and potentially connect you to others who can help you avoid the most common mistakes. My email is in my profile.

Re: Announcing Starfighter

#373

Post by yummyfajitas »

pron wrote: post #356
yummyfajitas wrote: post #319

Earlier quoted context omitted.

I don't deny group injustice. I don't believe injustice based on group membership is worse than any other kind - slavery is an injustice even if it's equal opportunity. Insofar as having bad schools or bad parents is not an injustice, it's also not an injustice if some group has more of them than another. And insofar as such a thing is an injustice, the fact that it's correlated with some group is irrelevant to me. I…

Well, group injustice is just individual injustice done to many individuals. And, I think, you would not consider it unfair to demand a robber to pay back the money he's stolen. The problem with underprivileged groups is not the reality of the offense against them, nor the identity of the victim, but the identity of the perpetrator, and this is where study and of power comes in. Because the reality is that human soci…

That an electron is measured through its effect does not mean that we don't need to describe the electron itself.

We certainly do. So look at what physicists did. First they came up with a clear definition - "a discrete and indivisible negatively charged particle". Second, they went out of their way to distinguish concepts like an electron from charge as a fluid and other models. They didn't simply declare "well, electric current proves electrons exist", they went to crazy lengths like Millikan's oil drops to distinguish these concepts.

When they can't actually distinguish these concepts intrinsically (as was the case with Maxwell's equations vs Aether), they tend to drop the more complex theory.

In contrast, you seem quite resistant to doing any of these things. I don't quite understand why.

Re: Announcing Starfighter

#374

Post by tptacek »

woah wrote: post #304
sfink wrote: post #237

Hey job seekers. I'm level 45 in starfighter with a power VI blaster in C++ and IX in Python. $100 to get you to level 30, x1.5 per level above that (though obviously I can't get you above 45, but I know a guy so I can subcontract up to 52.) You just have to set up a port forward from your machine and I'll take care of the rest. Contact info is in my profile. Don't get suckered into using a bot. They may be cheaper i…

I'm not sure that cheating on this would be unethical. Programming is about getting things done. As a hiring metric, nothing can beat open source. The open source portfolio really is kind of a golden standard. It is real work that you have done, that has (hopefully) advanced the commons, because you love programming and have something unique to add to the world. Some fake video game programming may indicate something…

That's awesome if you're the guy who wrote Expressjs, and less awesome if you're an extremely talented systems programmer nobody's ever heard of who hasn't hit on a project that happens to capture the industry's attention.

Re: Announcing Starfighter

#375

Post by tptacek »

pbsd wrote: post #371
tom_b wrote: post #369

Earlier quoted context omitted.

Am I wrong to be completely stunned at this statement? Did experienced candidates just not participate in the Crypto Challenges and Microcorruption? I would have expected at least some experienced candidates to drop into that "top tier" of people. What's up with this?

I finished 4th in Microcorruption. While exploiting vulnerabilities is not my day job, I have been playing CTFs for a long time, and am familiar with the process. It is certainly something one can overspecialize in. Looking at the first page of the Hall of Fame, I see big names like Alex Sotirov, Russ Cox, Ricky Zhou, Ludvig Strigeus, and many other familiar names/handles usually seen at CTF events. There are also ma…

There were "elite" participants, but they were numerically dominated by strong participants without the background. Which, if you're a hiring manager, is a very interesting and exploitable bit of data. Our odds of hiring Russ Cox or Alex Sotirov were not good. Our odds of being the first serious job for the next Alex Sotirov were better.

Re: Announcing Starfighter

#376

Post by tptacek »

djb_hackernews wrote: post #370

I'm not sure how this solves the topcoder problem. That problem being is it attracts only certain kinds of software developers and while it is used as a recruiting tool it basically changes nothing when it comes to the interview and hiring process. Even if you are recruited on topcoder you sometimes still need to submit your resume to some online resume eater "to get you in the system". And then you go on the same ti…

Topcoder challenges were not at all relevant to the kinds of programming problems I needed solved, and were not credible enough for me to consider whether they were predictive.

Re: Announcing Starfighter

#377

Post by tptacek »

fsniper wrote: post #367
kasey_junk wrote: post #365

Earlier quoted context omitted.

I suspect that the Starfighter crew's answer around this is twofold: 1) Of course there will always be other ways to get technical work. You needn't feel compelled to use their product unless you think the time input is worth the reward. 2) If they haven't made the process fun , that is something you want to spend your valuable free time on, they have failed (at least in acquiring you as a product).

Perhaps you are right about the answer. But It is still another knot on the rope. The next company expecting me do some obscure online automated coding test will need to pair me with an engineer inside.

That's not how this works. We don't host challenges for other companies. We're an outreach project. It costs companies money to find people through us; skipping us and going right to employers saves them money. I don't expect many employers to ask candidates to go to us to make them more expensive.

And, by the way, if you don't have a great resume and you're looking to work in a field directly related to the CTF we're running: making you more expensive is exactly what we're going to do. :)

Re: Announcing Starfighter

#378

Post by moriara »

Don't get me wrong, I would love for the current system to go away, since every interview I had got me one step closer to an imminent hear attack. But, I think the most we will get out of Starfighter is that it will become just another step in the process, MAYBE replacing the phone interview. Nothing is preventing a group of people from sitting in front of the computer, collaborating on a solution, there is absolutely no way to catch that. Also, none of the companies that are most in demand will abandon their dreaded on-site interview any time soon. It works for them, they need to weed out as many people as possible.

Re: Announcing Starfighter

#379

Post by jmagoon »

tptacek wrote: post #256
tokenrove wrote: post #203

Earlier quoted context omitted.

> Does it make application programming easier, less error prone, more accessible? If you want to write systems that improve the safety of application programming, it helps to have a lower-level understanding of what is happening. As for Mars, if articles like this ( http://spinroot.com/dcas/ ) are any indication, there will need to be a lot of systems programming and especially tool building that requires deep system…

It's so weird to me, because for me, it is absolutely the other way around. It's easy to break into computer systems. Put a single quote in the wrong place on enough web apps in a /19's worth of IP addresses and you will get in. What's interesting about security is the tour of functionality it takes you on. Security is an opportunity to crawl into the ventilation ducts, open the hatches, and directly tinker with the…

Wow, that's such an awesome point. I still remember the rush from reading 2600 and trying to break into systems in middle/high school.

That joy and excitement is totally gone now in Bob's CRUD Shop(TM), so it's phenomenal to have a curated lab to play around with (that won't get you arrested).

Re: Announcing Starfighter

#380

Post by sfink »

tlrobinson wrote: post #311
sfink wrote: post #237

Hey job seekers. I'm level 45 in starfighter with a power VI blaster in C++ and IX in Python. $100 to get you to level 30, x1.5 per level above that (though obviously I can't get you above 45, but I know a guy so I can subcontract up to 52.) You just have to set up a port forward from your machine and I'll take care of the rest. Contact info is in my profile. Don't get suckered into using a bot. They may be cheaper i…

I don't think you understand their business. Contingency recruiters' fees can reach 30% or more of an engineer's first year of salary, which could work out to $50k or more per engineer. You can bet they will be looking closely at each candidate's solutions, plus automated code plagiarism detection is pretty easy. It also depends on the type of problem. If you just need one weird trick to pass the level (e.x. use this…

Sorry, I don't see the connection. Yes, a successful candidate is worth a lot of money to Kalzumeus, and they are incented to make sure that the solutions are real and not plagiarized. But that salary is worth even more to that successful candidate, and hence they would be willing to pay handsomely for someone to write customized, non-plagiarized solutions for them if it gets them in. (The mythical level 45 mentioned would cost $44K. If I'd done the math, I probably would've made it more. I guess the level 52 would be nearly $750K...)

For now, I'm assuming they have good protection against automated solutions. If it were just a matter of detecting plagiarism, you're right that it would be trivial. (Heck, I even cofounded a company that did code plagiarism detection, so I have a very good sense for how easy or hard it is. tl;dr: it's straightforward to defeat a plagiarism detector if you are sufficiently motivated and knowledgeable and willing to put in some time to go through all the code, but people/companies who plagiarize are typically none of those.)

Detecting whether someone else is doing the candidate's work, on the other hand, is impossible. Unless you bend the rules. Fortunately, the rules seem pretty pliable to me -- when a candidate arrives at a company's interview, spot-check them on their understanding of a few of "their" trickier solutions. It's a logistical nuisance, to be sure.

Post reply on HN