I obviously won't doubt the genius of Terrence Tao, but reading about PhD level research from any other discipline without knowing a yota about it would leave exactly the same impression, even from a lesser scientist. Of course some people are incredibly smart, but there is no need to mythologize their abilities without having any understanding of what they have actually done. Thus I think such an article can only have value for people who at least vaguely understand the mathematics involved.
I think Terrence on his blog does a decent job as coming across as someone who is really smart but has got to that position in maths by spending a lot of time in some pretty narrow fields. He openly admits, like for example when P/NP stuff come up a bit maybe 6 months back, areas where he isn't an expert.
He could probably easily make his knowledge look even more impressive but is humble about what he does and doesn't know.
"How? Because people in the field of neural networks and AI would never claim that Minsky "pioneered neural networks". To the contrary (and as Minsky's wikipedia article – i'm sure the source of this claim – obliquely notes), Minsky's pessimism about the abilities of neural network computing lead to the abandonment of artificial neural networks as a major research topic."
That is a very confused description of what happened. Minsky is in fact a very important contributor to early neural networks theory and what you refer to as his "pessimism", is in fact his proof that a neural network can not be trained in any way to "learn" the exclusive-or logical function (among other things). This is one of the fundamental results in NN theory.
It is easy if the language your compiler is written in is itself running on the JVM and you can access the nice libraries for generating the bytecode, for example ASM (http://asm.ow2.org/) which Clojure uses. I built a toy compiler for a Pascal-like language targeting JVM and conceptually it is easy (as you maybe meant), but generating binary bytecode "by hand" would require a lot of extra work. I took a look at ABCL source, it dedicates at least 100kb of Lisp code to a bytecode-generating library.
By the way, ABCL source code seems very interesting for anyone interested in compilers/interpreters/Lisp.
Speaking as someone who applied for MIT a few years ago, something like this is no longer possible and the "rat race" description used for comparision is now in fact valid for MIT as well.
Nowhere in the recrutation process you have much possiblity to show your "software code" - everything is very formalized and you have to submit your grades, essays on specified topics, pass the SATs and go through a interview (but the interviewer doesn't have to know anything about the discipline you want to study). Yes, you can describe your most interesting projects as part of your application, but if you read the admission blogs and other MIT materials, it is quite clearly implied that unless you have near-perfect grades and/or near-perfect SAT scores, they won't even look at the project descriptions, essays etc. Also there is no way of knowing why you were accepted or rejected, because the whole proccess is 100% opaque to the outside world.
I still think the MIT is awesome and the admission process probably has to look more or less like it looks like because of the volume of applications they have to go through. But the post and some of the comments seem to leave the impression that the MIT addmission comitee will look at every person as a "unique snowflake" to find the really outstanding candidates. In reality, the admission process has to be quite mechanical so that they can at all manage it and only after the initial 90% of the applications gets rejected, they can be scrutinize the remaining 10% in more detail. So, if you want to get-in, you have to "optimize grades and SAT" and "speaking French and Chinese, playing piano and painting abstract art" won't hurt either.
They have the application process. You can ignore it. I did. I had shit grades in high school (not shit-by-MIT-standards, but actual shit). I took no time to do homework, and instead spent time reading math books and programming. By the time I applied, I had done research at a reputable university laboratory (although I did not get to the point of publishing), and had a good recommendation from a well-known professor there. I had also taken several advanced math classes at a state school.
I sent MIT a custom admission packet. I filled out their paperwork for biographical information, but mostly ignored it.
MIT rejected me early admission. I called to ask why. They told me that they liked my packet, but given my grades, they were concerned about my maturity and my ability to work hard. I got an extra recommendation from a professor teaching the math class I was taking, whom I had asked to explicitly let them know that I was mature and a hardworker. At that point, MIT took me.
I didn't even bother applying to Princeton, Harvard, Stanford, and the like, since I knew I had no chance.
Optimizing for grades is a bad and stupid strategy. If you're late in the game (high school), it's the only strategy with a chance of success. If you're circa elementary school, the best strategy is to ignore school. The time with no homework will let you learn math and science (which you can do much more quickly than school will teach you), and to have real accomplishments by the time you apply.
Real accomplishments ALWAYS trump grades, for anything, be it university admissions, jobs, or YCombinator. Grades are a proxy to see whether you are smart and can do useful things. Accomplishments are a direct measure.
Unfortunately if I encouraged my daughter to ignore school and pursue something she's passionate about, Child Protective Services would be here in short order.
That's exactly what I'm talking about too. I didn't even think about attendance. There are a lot of busybodies in this world who will think nothing of trying to ruin your life, or at least make it temporarily miserable, if they don't like what you're doing. ESPECIALLY if they don't like how you're raising your kid.
All well and good, but I'm not sure doing "research at a reputable university laboratory" and having "a good recommendation from a well-known professor" can really be classed as bypassing the academic admissions system.
It's not, but it's certainly outside of the "standard" sequence. The useful lesson to extract is if anyone uses one metric (say, grades) as a proxy for another (say, ability to work and flourish in a university environment), and it's possible to score points on the second metric, do so. Even if there's no official channel for it, that's what people really care about and they'll immediately recognize it as such.
The reason that this sort of thing is so rare, I think, is that we don't know what metrics matter until we're well past the point of applying to something. Even if we're told, it's hard to overcome the expectation and pressure of doing well on the proxy metrics. For example, in college, I had no concept of what mattered for grad school: research. And even if someone sat me down and told me, I'm not sure if I would really understand it. It's a rare person that realizes it's even possible to obtain points in the metric that really matters.
That said, it was MIT's experience back when the SAT meant something (pre-1994) that class rank (i.e. grades) were one of the two best predictors of subsequent success.
Woah, slow down here. Optimizing for grades is NOT a bad or stupid strategy, especially when you're young. By doing well academically, you put yourself in a position to do well in the future. Sure, pursue your interests and shoot for real accomplishments--so that when you're applying to college with those real accomplishments you don't need to explain why you didn't get good grades. It's just putting yourself in the best possible position to be rewarded for your work.
Don't short change yourself. Do your best in school, but be pragmatic and work hard outside of it too.
> I took no time to do homework, and instead spent time reading math books and programming
vs.
>I got an extra recommendation from a professor teaching the math class I was taking, whom I had asked to explicitly let them know that I was mature and a hardworker...
These statements are irreconcilable. A hard worker does work that he/she is assigned, and does not make pre-mature value judgements on the worthiness of said work. Being smart does not mean you're a hard worker. That was a pretty irresponsible recommendation by the professor, considering you didn't exhibit the qualities you claimed to.
> These statements are irreconcilable. A hard worker does work that he/she is assigned, and does not make pre-mature value judgements on the worthiness of said work.
Nonsense. Work should always be in furtherance of some goal.
I don't want to hire the person who does exactly as I say, even if it's a dumb idea in the larger context. What's important is reaching the goal (ethically, legally &c).
While there are people who use anti-authoritarianism to justify laziness, that doesn't seem to be the case here.
Perhaps I was unclear. I had crappy grades in high school. High school homework was a pointless waste of time, so I didn't do it. I had good grades in 3 of 4 of my university math classes. Homework there was interesting and not a waste of my time, so I did it. When I hit MIT, my grades skyrocketed, because with a small number of exceptions, the classes were fun, and the problem sets were interesting and useful.
I did poorly in my first class -- I didn't realize this immediately -- I don't believe I sent a university transcript to MIT admissions, but if I did, they would have seen one bad and one good early, and one bad and two good normal admissions.
Maybe. But a smart worker evaluates what's important and what's not. Formal education and learning are sometimes completely orthogonal to each other. You might be thinking "well, how do you know what's important?" and the answer is that it's a gamble and decisions like those are not for everyone. There is no virtue in simply doing what you're told and being passive in your own education.
The OP claimed that he procured a reference that attested to his "hard-workingness". It's perfectly ok to be a smart worker, and I don't disagree with your point. My point of contention is with the abuse of the reference system - what value does the reference system have if you simply ask for (and receive) character attributes you do not possess?
So basically he was admitted on traditional merit, by misleading application reviewers.
About one decade. I'm still an MIT affiliate. Admissions has not changed, in this regard, since. There have been a lot of other changes (the students now are better rounded and better looking than when I was a student), but this is not one of them.
MIT Admissions intentionally takes high risks. They admit a number of students at the extremes who may end up very bad or very good. The risks don't always pay off (we had a few real idiots), but the policy is actually quite sound. MIT's name comes from it's most famous graduates -- the Feynmans, Aldrins, Metcalfes, and Kurzweils -- the cost of having a number of bad or flaky graduates is rather low -- only the people who work with them ever hear about them.
Props to you for your determination, hard work and perseverance. I think it speaks very well of your then maturity when you look at the way you approached what is, for most young adults, a very intimidating process.
That's not entirely correct. You don't need near perfect SATs/grades, especially at a place like MIT. I know plenty of people (myself included) who do not fit that bill. And there is plenty of opportunity to discuss your projects. I mailed MIT a packet of all the interesting work I thought I had done and they even have an optional essay where you can talk about "something you built" in their application. I think they really care about those pet projects and they are the best differentiator they have.
If you look at percentage rates, the self-selection of the candidates might make it look like it is not so rigoristic. But take a look at the absolute numbers, for example for the "SAT Reasoning Test Scores (Math)":
1172 / (1172 + 269 + 108 + 2) =~ 0.76
So 76% of the students enrolled have scores in the best 750-800 range.
The math portion of the SAT doesn't even cover precalculus (stops at Algebra II). Would you really expect someone who scored below 700 to have otherwise impressive credentials?
I'll agree that the more stellar applicants do tend to have high SAT scores but that's also in part because they are a) stellar and b) the SAT is a trivial test. I still disagree that SAT is a cutoff or benchmark. I highly doubt the admissions committee sees a 600 and throws the student out before reading the rest of their application to look for things that stand out. That said, there are tons of students with 800s that don't stand out at all and get weeded out.
Exceptional candidates will be exceptional regardless of their standardized testing scores.
Edit: Not trying to say that they admit all the exceptional people either. I'm absolutely positive plenty of people get screwed by the limited number of slots despite being more than qualified.
"Of course you need good scores and good grades to get into MIT. But most people who apply to MIT have good grades and scores. Having bad grades or scores will certainly hurt you, but I'm sorry to say that having great grades and scores doesn't really help you - it just means that you're competitive with most of the rest of our applicants. MIT is very self-selecting in that regard."
> Would you really expect someone who scored below 700 to have otherwise impressive credentials?
Errrr. I think I got a 680 and I'm pretty sure I got a 98% in Calc BC and a 5 on the AP. I also know a friend that got a 750 on the Math SAT, dropped out of Calc because he didn't understand it, and I had to tutor him.
I don't think I said "No one who gets below a 700 has impressive credentials". I was just trying to generalize that higher scoring students are more likely to have other impressive things about them.
Look the SATs are such ridiculously easy tests that scoring perfect or near-perfect on them doesn't say anything about you, only that you might be competent. Scoring lowly almost certainly means you aren't. SATs are only useful for distinguishing varying grades of mediocrity. That's why MIT asks for AMC/AIME scores.
I call BS to some extent, to both the original post and this comment.
I went to a major Ivy. I know firsthand of someone who got early admission without doing senior year (from an elite foreign school); in honors calculus we had a high school kid who ran circles around us (not a full-time degree student IIRC); in our CS program we had a Doogie Howser type who was admitted in his teens and last I heard was some kind of security guru.
I'm sure any top school has a small number of brilliant students with oddball backgrounds. In other countries you have to get the piece of paper, but in the US no one cares about a HS diploma. If you can show you've got the goods, it doesn't matter how you show it.
(Although MIT is more likely to care purely about ability to be a great engineer as opposed to a well-rounded human being - to that extent I can believe other schools would not have been receptive to a candidate based on code LOL)
MIT doesn't care purely" "about ability to be a great engineer as opposed to a well-rounded human being", some* of the latter is required. And in my experience that isn't too hard to find, how many great engineers or scientists that you've known or heard of are only interested in their work and have no interest in, say, music (that was common at MIT in the '80s at least).
I just applied this year, and was incredibly pleased to get accepted. (woo!)
The thing is, my grades weren't stellar. Don't get me wrong, they were great, and I was ~15th in my class-- but people had better. Similarly, I only had a 2010 on the SATs after taking it twice. (To be fair though, I did have good AP scores)
Yet, with all that, I was the only one in my class to get in, out of 6. That group included the valedictorian, and saledictorian--heck, I was the only person not in the top 5%. And, I've been the first person to get in from my school in at least the last 7 years.
And, for what it's worth, my extracurriculars were relatively flat: only one language (French), no instruments, no athletics, no arts.
1) Got rejected from MIT
2) Admissions process must be shitty
3) MIT is great but must admit shitty students instead of me
4) Read an article about a strong non-traditional
5) Equate that guy with self
6) Try to forget that self was just a weak traditional
I agree. It is very difficult to figure out where and who to show your code too.
When I applied to undergrad 5 (or was it 6?) years ago I applied to CMU. I submitted with my application an "abstract" about some software I wrote. The software was really good. In fact it got me a job my first year in college (at a security consulting company). But, I don't think it got to the right people at CMU. They probably looked at a low score I got on a test and passed.
Maybe I should have tried MIT instead of CMU. I didn't know anything about colleges so I only applied to three Case Western, CMU, and Northwestern and they were all near where I lived. Only when I was in college did I learn about other good schools out there.
I think you're overgeneralizing, especially 100 various things can be understood under the term "meditation". There is a program for mindfulness-based stress reduction using meditation as one of the elements, with scientifically proven effects: http://www.mindfullivingprograms.com/whatMBSR.php
This is master level, "Captain Obvious"-style trolling, beyond me how this is the top comment in a place like HN.