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

I taught two courses to undergrad CS students at Stanford. I think about half of students cheated on problem sets and tests. The answers were all way too uniform and done without correction.

In the first course, there was a student who definitely didn’t cheat and was objectively horrible at all tasks assigned, consuming a frustratingly large portion of my grading time. This student ended up with a massively impressive resume including all the best places, and became the founder of a billion dollar cryptocurrency.

In the second course, the worst student also clearly didn’t cheat, and sometimes I gave extra points for answers that were wrong but the only ones in the class that were invented rather than regurgitated. That student founded one of the most influential companies in modern computer science, and also a 10 billion-dollar cryptocurrency.

I’m not sure what the lesson is. I had honest and very good students that didn’t do much of note, but there were only two honest but awful students, and both had superlative success. Just for reference, I was honest and pushed hard to land somewhere in the low middle, and I’ve had my share of ups and downs, but definitely no competition for the lanterne rouges.




Interesting that they both founded cryptocurrencies. It’s not surprising that academic success is uncorrelated with being able to run a good Ponzi scheme. Bernie Madoff was extremely successful in his time, but he wasn’t a prodigy either.


So by and by you're saying something like Cardano is definitely not a ponzi because it's extremely focused on doing everything academically?

https://docs.cardano.org/explore-cardano/relevant-research-p...

If this is a ponzi then I have to give them credit, to be doing research at a university level in order to swipe me of a bundle of dollars is amazing. They don't have anything about making themselves richer at the expense of new users either, for some reason it's all focused on developing a great blockchain that is sustainable and is usable for dapps which ultimately provides utility for end users.


I know what Cardano is because I've heard of it through the Haskell community. I bet most people have not. How do you sift through the thousands of ridiculous coins out there and say to yourself, "Oh, they probably have integrity"?

This continues to be a mystery to me as it seems no one can put themselves in the common man shoes. The common man doesn't care about tech. They don't care about the integrity behind it. They care about whether it can make them money.

List 5 coins I've never heard of and I'd probably say, "ponzi scheme" without hesitation. Do I actually know if it is? No. Does it matter? No. If the majority of shitcoins come off as a ponzi scheme, then why would anyone think something like Cardano isn't?


You don't need to know or trust any underlying technology though. The only thing you need is a basic grasp of Darwinian natural selection.

The "shittier" those "shitcoins" are, the quicker they'll die out. The ones that heavily incentivize early adopters might hang around longer, but they too will inevitably fizzle out if they don't offer any practical innovation.

But if something has a steadily growing user base year after year, maybe it's worth some attention after all?

Or you can just keep screeching "Shitcoins! Shitcoins!". I guess we can all settle for ourselves what form of discourse we consider more constructive and worthwhile.


GP said uncorrelated. I don’t think that’s an effective counterpoint, although GP seems to be indicating anti-correlation. Either way, this is about undergraduate students, and I also don’t think you can claim superlative undergraduate performance for that particular founder. Maybe crypto is where motivated underperformers found a place at the table because the overperformers are settled in more civilized places. Also, these research papers in cryptocurrency are probably not comparable to serious contemporary work in mainstream journals of pure math and theoretical computer science, but I am personally fascinated by the way that old research is slowly finding it’s way into applications through the sometimes absurd financialization of these ideas, enough to get scammed and swindled a bit myself.


That wouldn't be a 'billion dollar'+ cryptocurrency though would it


It’s a 16 billion dollar market cap currently.


Guess we'll find out which it is in the next few months. Proof of stake is just as vulnerable to Ponzi, especially if promising any sort of outsized interest returns through middle-men services. The 'scientific, peer-reviewed philosophy' statements appear as marketing, with peer review appearing to be crypto tech conferences (not itself a poor measure), not the classical 'peer-review' format the statements would allude to.


We're finding out it seems...


We are still saying that the entirety of CryptoCurrencies are Ponzi schemes ?

Yes some of them are, but its a very legitimate technology that will change the world.


Becoming successful requires a stubborn conviction that you can make your own ideas a reality. Doing well on tests requires the ability to re-enact the ideas of others.

They're both valuable skills of course, but they should be recognized as being distinct from each other; and measuring "intelligence" as some one-dimensional modality is a spectacular failure to do so, IMO.

This idea is also to an extent analogous to the tradeoff between exploration and exploitation in reinforcement learning (AI).


I’ve known a fair number of very successful people, and the strongest theme among them is ignorance of the challenges they had committed themselves to, perhaps even a stubborn disability to heed warning that there is yet another layer of the onion after years of peeling. I really don’t think it’s some kind of dreamer attitude or deceptive nature; maybe just blissful ignorance. In most cases, it starts with a simple idea that everybody had, e.g. a CS50 class project that everybody else did, but they all saw the technical and legal and market and competitive challenges, and most importantly and predictably failed to answer the biggest question, “why me?”. Somehow the person with the worst answer to that question is always the one that fails to ask it. Despite this observation, none of this has changed the way I see things in my real life. I have read that the parasympathetic nervous system causes people to pay attention to details and become overwhelmed by risks. Chronic pain or fear or lack of sleep or overtraining or infection can do that. I recently had a lidocaine injection into a painful joint deformity, and the rest of that day I experienced astonishing positivity. So something tells me that it’s not really all that complicated, but simple things like happiness and good health and relationships are far more rare than we would expect, at least among those with curiosity for difficult subjects.


Perhaps that shows that these students figured out they're not so great at computer science, but ended up knowing a lot of people who were - and how to use their skills to make a small fortune.


those who know, code. those who don't, find someone who can


You can buy problem sets and answers online and I suspect that's one way people might cheat especially on homework. A lot of schools just reuse textbook problem sets or have a bank of questions they use for problem sets and exams, especially for larger classes, and people can buy all the answers online and find the ones they need. I think during the Covid pandemic when things moved online that really accelerated but previously it was a grey area since it could be just a form of studying.

It's sad because I think a lot of those classes are curved and people who don't cheat probably have a disadvantage. Despite putting more effort in and maybe even knowing more than other students, people who don't cheat could end up at the bottom of the curve but might have actually fared better under different conditions.


Tests always get leaked somehow. The students that were hit by the curve were, like me in school, struggling honestly toward the middle. That method is probably not as good at knowledge acquisition as the various methods of cheating, but hopefully better at flexible problem solving. I think these unique bottom cases were playing a different game where the rules are made up and the scores don’t matter; the curve probably did not affect them.


> Tests always get leaked somehow.

Tangentially, it seems incredibly unfair that self-plagarism is a thing because "if you're not coming up with new material then you're not engaging with the course" and yet professors aren't even taking the incredibly simple step of changing a few numbers between repeats of an exam. It's basically self-plagarism from the professor, the professor isn't engaging with the course in the same way that a self-plagarizing student is not.

The dead-simple answer to students who are looking at test banks is to produce semi-novel material for each course, in the same fashion that students are expected to produce semi-novel material for each course. In both cases, yeah, it's kinda trite, there's only so many novel takes you can have on the best way to factor an integer or the implications of shakespeare on russian 20th century literature, but that's the system we've agreed upon.

It should cut both ways and the fact that professors suffer consequences and difficulty from not producing novel material is not really the student's problem. The solutions of forcing syllabus-level contracts-of-adhesion above and beyond the school academic policies is not a fair solution either.

I realize there's usually a "and whatever the syllabus says goes" clause but that's shitty, a lazy professor should not be able to exempt themselves from academic diligence by adding a "and it's OK for this professor to be lazy" clause in a contract of adhesion.


> Tests always get leaked somehow.

This is the side effect of lazy teaching. Unique tests cannot be leaked. Problem solved.

Or maybe don't do stupid tests for computer science.

This is a space where whiteboarding is a good idea, in fact, interactive whiteboarding with a whole class is a great way to teach and work through problems.

Any type of interactive test is hard to fake.


Fun example about lazy teaching: Our undergrad's HKN chapter kept a test bank of past exams in Physics, CS, etc. When I was studying for my Physics final I went back to past midterms and realized the professor recycled questions from previous years with some numbers changed up. I ended up filling my entire cheat sheet with the questions that weren't already on midterms but were in the test bank.

The result? The final contained a majority of questions from that cheat sheet. I still pose this scenario to a couple friends nowadays asking if this would be considered cheating, and the responses are actually a mix of "no, you're just taking what's publicly available" and "yes, you're manipulating the system to get where you are and taking spots from someone with academic integrity".


Just ask yourself. Did you learn the course well enough that you could have done as well without those questions saved?

As others have said, we go to college these days for a fastpass to middle class. If you haven’t needed the knowledge since, maybe you got what you needed out of the class and you worked the system in a legal way. Or maybe you paid a lot of money for that class and yet cheated yourself out of the knowledge you were paying to be taught. On some level it comes down to how you feel about it.


Since that class I've reached the conclusion that school is not the best way to learn concepts and materials but to learn how to work a bureaucratic organization to reach your own end goal (whether it's a job, ticket to the middle class, or simply knowledge), which in itself is a valuable skill. I think museums, seminars, and some NPR shows ended up being the best ways for me to really master the material I would've otherwise learned in school.


I think anyone who is naive enough to think that school is about anything but conformance deserves to fail. I did some courses before college and got a high grade and when I went there for my bachelor's degree, they didn't count that course towards my credit points, I had to do it again. In other words, my knowledge means jackshit at a educational institution. The only thing that matters is whether I can navigate the bureaucracy. Get that in your head.


> This is the side effect of lazy teaching. Unique tests cannot be leaked. Problem solved.

I have never heard of lazy teachers ever getting any consequences, so why should they stop doing these things? You could make it a part of their job, but of course lazy principals wont do that either. There just aren't any incentives to run a school properly.


The fact is few teachers at the university level are interested in teaching the majority of classes that students take. Some professors may enjoy teaching the higher grad level coursework, but overall professors are generally terrible teachers.

Teachers go to learn how to teach, in the abstract sense, for 4 years, to go to teach high school kids. They learn the techniques and tools to effectively translate subjects and educate people.

Professors do not do this.


>I think these unique bottom cases were playing a different game where the rules are made up and the scores don’t matter

you may be on to something with this one


this was a huge deal back in the 40s-70s. They were called 'bibles'


they got into Stanford, which confers above average intelligence and also bestows connections, so that probably helped. Interesting stories.


What do you mean by "confers above average intelligence?"


They become more intelligent when they receive the admission letter.


become recognized as more intelligent


> billion-dollar cryptocurrency

> I’m not sure what the lesson is

Gambling is a profitable business.


That you should leave your cheating until you've left education because there are fewer checks, people are more credulous and the rewards are greater?


It sounds like they were occupied with something that isn't college.


Was it Do Kwon?


My own feelings aside, Do might be a good case study. I don’t think he’s an intentional scammer; he just had an idea that he was convinced was brilliant, without the knowledge or maybe imagination to see how spectacularly dangerous it was. I personally avoided Terra/UST from genesis because I immediately saw that is was exploitable with a big enough position on UST. But it took some searching to find the exact peg mechanism, and by the time it collapsed, it had gotten so big that I started to question my own judgement. Do might have caught whiffs of the already risky ideas surrounding algorithmic stablecoins, and pushed forward something that looked passably similar at his introductory level of knowledge. The apparent possibilities at this level are far greater than what is possible with more knowledge in the area. The investor class is also stuck at this level of knowledge, because they make their money by raising from and selling to the level down from there. I don’t see any distinction between crypto and traditional equities in this regard, except that the SEC is supposed to protect ignorant consumers, which they clearly can’t, and the government is there to backstop any disasters at the expense of everybody smart enough to avoid it in the first place, which makes them the dumb ones in the end. So yeah, it was a stupid idea, but look hard enough and you’ll find the same stupid ideas behind every investment with any hope of keeping up with the cost of feeding, housing, and educating your kids. If there is evidence that shows that Do is the one who pulled the rug, then my theory is flawed. Basically, it is that people who achieve great investment success are operating at a level of genuine enthusiastic ignorance that cannot be faked by a scammer or an expert.


From my understanding of cryptocurrency and the people in it, I don't have the premise that founding a billion dollar (or 10 billion dollar) cryptocurrency is laudable.

It strikes me that your comment would read by crypto fanbase as a success story, and yet as crypto detractors like myself as even further reason to regard them as terrible... you may have just invented the "completely unintentional if-by-whiskey argument[1]"

[1] https://en.wikipedia.org/wiki/If-by-whiskey


As a benefit to society? Not laudable IMO. But it does take hard work to swindle people into your crypto - that's the perspective I read OPs comment in.


If it were a 216B cryptocurrency then it would indeed be laudable. It’s Ethereum and it is innovating in every aspect.


Cryptocurrencies were innovating in one particular aspect, solving the double spending problem for truly anonymous transactions. Beyond that it's not as innovative as people think. Even Bitcoin supports smart contracts, and the only real noteworthy achievement in the real world Ethereum has achieved is becoming a massive platform for gamblers in the form of speculation, which makes up almost the entirety of Ethereum's activity. No name engineers at companies like Google deal with far more complex and far more practical problems as part of their job. I say this as someone who has written their own block chain from scratch (not a fork) and understands how incredibly niche and overblown cryptocurrencies are.


Just because you went through a “how to build a blockchain in 1 hour” YouTube tutorial doesn’t make you an expert on the topic anymore than someone who “built a Twitter clone in 1 hour” is an expert on distributed databases.

The idea that no name engineers at Google work on more interesting problems is laughable to me as someone who worked there for 3 years and saw the vast majority of them work on garbage Angular apps or borgcfg . Not exactly designing a zero knowledge roll up chain in complexity.

The “cryptocurrency=bad” bros on hacker news are getting so tiresome. Many people think things like adtech are scams yet if every single time someone brought up Google we derailed into that conversation it would rightfully get moderated but with crypto, dang is asleep at the wheel.

Actual cryptography experts like Dan Boneh and Silvio Macali are full time on cryptocurrency, but since your L4 Google buddy thinks it’s a scam, that settles it so we bring it up every damn time.


There’s cool cryptography going on in Ethereum, but it’s not helping economize anything useful. At some point you have to notice that all of the “cool new stuff” in Ethereum for the last 7 years all seems to disappear once greater fool liquidity has been exhausted.

Meanwhile adtech is making the internet awful, and thus stories increasingly show up about how bad Google has been getting as well.

I used to read all the white papers for all the new coins, and was definitely on the side of Ethereum as an interesting experiment when it released, but nothing has happened to make me think a premined securities play hasn’t turned out to be a scam.


I did not use any guides or tutorials, it was truly from scratch with some of the cryptography concepts borrowed from bitcoin, but with a completely different auth scheme for transactions designed specifically for business application. https://github.com/Salgat/FangChain

And not everyone at companies like Google are doing the same thing. Just because you didn't know them doesn't mean there wasn't brilliant work being done there, especially with regard to their scaling and machine learning. The folks at deep mind are doing far more brilliant and innovative work than Ethereum.


> it is innovating in every aspect

Could you expand on what you mean here? What are the aspects?


Well decentralisation is hard, it's particularly hard when it's dealing with real peoples money, it's hard to scale and hard to maintain, Ethereum is getting closer and closer day by day to moving to an alternative system called Proof of Stake, it's a different system from Proof of Work, you may be familiar with Bitcoin and its energy use.

You can read more about issues on Vitaliks website for example when Musk tried to become a blockchain expert overnight he looked a bit silly:

https://vitalik.ca/general/2021/05/23/scaling.html

Vitalik is probably the only Billionaire in the world I don't hate and ethereums innovations are incredible.


> Ethereum is getting closer and closer day by day to moving to an alternative system called Proof of Stake

At least this is honest and doesn't provide a date given how many times the planned date has slipped.


EVM based deterministic execution has proven to be a resilient model for smart contracts. The other major inventions are not Ethereum itself, but it led to, to name a few, PoS consensus and zero knowledge proofs which I particularly admired. It has grown into a huge ecosystem of practical cryptographic schemes. Ponzi schemes are possible with just Bitcoin, and yeah Ethereum is full of crypto ponzis at the moment, but all these useful cryptographic inventions may not have been made practical without Ethereum.


The first is not revolutionary nor a swindle. The second is on that level or more so.




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

Search: