Is MIT losing it?

Colleges & Universities A-Z Massachusetts Institute of Technology
141

<p>I read that MIT outpaces all other universities in terms of # of patents. It is very good, probably the best, at converting research to startup companies. The problem is that start-ups are often invisible if they aren't software companies. For instance, controlled drug delivery is probably the most important discovery in chemical engineering in the past 20 years and it was pioneered by Robert Langer at MIT. But no one has ever heard of it. </p>

<p>(I don't believe that normalizing for size really is that relevant; MIT has a huge number of researchers of high quality. CalTech has great faculty too but is not as big and so I would guess has less impact overall.) </p>

<p>I have noticed the same thing as "scientist" about the big-name software companies, though. We still have huge names in comp sci., but a lot of what the most recent contributions have been theoretical. (For example, Chafi Goldwasser has recently made a huge contribution to the NP-completeness problem. Also, Ron Rivest is one of the huge names in artificial intelligence.)</p>

142

<p>



I read that MIT outpaces all other universities in terms of # of patents. It is very good, probably the best, at converting research to startup companies.
(I don't believe that normalizing for size really is that relevant; MIT has a huge number of researchers of high quality. CalTech has great faculty too but is not as big and so I would guess has less impact overall.)

MIT has most patents because it is the biggest engineering school. It also has the history advantage. This is something that separates MIT from its competitors such as Stanford, Berkeley, and Caltech. The huge number of engineers trained from MIT helps building its enormous reputation in academia and industry. Everybody knows that. And this has a huge direct impact on its US NEWS ranking in engineering. For example, the 3 categories US NEWS used for engineering school ranking: the # of ph.d produced, the total graduate students enrolled, and the total research expenditures in engineering obviously line up with the size, and helps MIT secure the #1 ranking. If you normalize all these by size, Caltech, Stanford, Berkeley, Princeton, and Harvard will be on par with or exceed MIT. For example, percentage wise, these schools are about the same in faculty membership at the national academy of engineering and some are a little better than MIT. So is another category: reasearch expenditures per faculty. 



I have noticed the same thing as "scientist" about the big-name software companies, though. We still have huge names in comp sci., but a lot of what the most recent contributions have been theoretical. (For example, Chafi Goldwasser has recently made a huge contribution to the NP-completeness problem. Also, Ron Rivest is one of the huge names in artificial intelligence.)

You are right that MIT is extremly strong and probably the stongest in theory of computer science now. Stanford used to be the strongest in theory. But after professor Floyd's death and Professor Knuth's retirement, MIT has surpassed Stanford in theory in my opinion.</p>

<p>I know Rivest is great in algorithm. But is he really great in AI?</p>

<p>Lots of Turing award winners on theory side have some sort of connections to Stanford: Knuth, Floyd, Wirth, Tarjan, Hopcroft, Yao, Rivest (now at MIT), and Pnueli. But that is history. I don't want to use Stanford's glorious history to deny MIT's current leading role in theory.</p>

143

<p>You raise some good points...</p>

<p>However, I don't agree that normalizing by size is fair. I think it is more impressive to have such a large faculty with an equal concentration of faculty worthy of the National Academy of Engineering.</p>

<p>Plus, CalTech has a number of associated labs (JPL, for instance) whose directors may not be formal faculty members, yet they prouduce research and patents which I think CalTech claims credit for. So perhaps CalTech is not quite as small as its faculty number might indicate.</p>

<p>Also, when I said MIT put out more patents, I meant patents per year.</p>

<p>Concerning Ron Rivest, this is not really my field but I though he was the biggest name in the field in a certain subfield of A.I. called "embedded intelligence." I could be mixing him up with someone else at MIT.</p>

144

<p>agree with you. For engineering, sometimes, the bigger the better.</p>

<p>



Concerning Ron Rivest, this is not really my field but I though he was the biggest name in the field in a certain subfield of A.I. called "embedded intelligence." I could be mixing him up with someone else at MIT.

Maybe you are talking about Rodney Brooks? He is a huge name in AI, the director of MIT CS and AI lab. By the way, both Rivest anf Brooks are Stanford trained Ph.d.</p>

145

<p>Here is a piece of information about number of patents from various schools which may be somewhat relevant.</p>

<p><a href="http://www.learn4good.com/top10/universities_usa.htm%5B/url%5D"&gt;http://www.learn4good.com/top10/universities_usa.htm&lt;/a&gt;&lt;/p>

146

<p>
[quote]

This is a crappy argument. You keep bringing up MIT's story in old history. Mind telling us when the 'galactic network' happened? The basic buiding blocks for today's internet are 'packet switching' and TCP/IP. Did Licklider contribute anything to this? Nothing!

[/quote]
</p>

<p>What are you talking about? After coming up with the idea of the Galactic Network, Licklider than became a main mover at DARPA, where he convinced a number of people, including Roberts, Ivan Sutherland, and Bob Taylor, that the concept of an all-encompassing worldwide network was important work to do. Hence, he was the original inspiration for the Internet in the first place! How can you call that 'nothing'? </p>

<p>
[quote]
Again, the building blocks of internet are packet switching and TCP/IP. Kleinrock shares 1/3 of the contribution to packet switching with Paul Barren and Donald Davies. But Donald Davies might argue that Kleinrock's contribution was less important. Cerf shares at least 1/2 of the contribution to TCP/IP with Kahn. Roberts was more or less a manager roll of ARPARNET. He couldn't claim any contribution to the invention of packet switching or to TCP/IP. TCP/IP is more important than packet switching, because TCP/IP's inventors Cerf and Kahn have been honored by some top notched prizes: Turing prize, National medal of technology, and medal of freedom.</p>

<p>So let's do some simple math here. MIT's contribution is Kleinrock's 1/3 packet switching. Stanford's contribution is Cerf's 1/2 TCP/IP. Since TCP/IP > packet switching and 1/2 > 1/3. Hence Stanford > MIT in the birth of internet.

[/quote]
</p>

<p>Way to completely ignore Robert's contribution or to the creation of the Arpanet in the first place. I would calculate it as follows:</p>

<p>Roberts and Kleinrock = each 1/2 of the Arpanet = 1
Cerf = 1/2 of TCP/IP = 1/2</p>

<p>Hence, Roberts + Kleinrock is still more than Cerf. That is, unless you want to argue that the development and design of the original Arpanet is just not important at all. Are you prepared to argue that? </p>

<p>
[quote]
By the way, in terms of APARNET, Stanford was one of the 4 NODES. UCLA was the 1st NODE. Stanford was the 2nd NODE. MIT was not even a NODE.

[/quote]
</p>

<p>Wait, I thought you said that the Arpanet wasn't important anyway. So which is it? If the Arpanet is important, then you better give proper credit to both Roberts for designing it and Licklider for inspiring it. </p>

<p>
[quote]
Of course, you can claim Tim Berners-Lee has a tie to MIT. But you can NOT claim MIT has contributed anything to the invention of WWW.

[/quote]
</p>

<p>I never said that he did. I said that Berners-Lee founded the W3C at MIT. There is no W3C office situated at Stanford. </p>

<p>
[quote]
Again. The most recent USNEWS ranking put Stanford, MIT, Berkeley, and CMU tied for #1. But Stanford is the only one ranked #1 in all years. In NRC ranking, Stanford ranked ahead of MIT in CS. In addition to that, there are plenty of other reasons to claim Stanford CS is better than MIT CS. 1) Stanford has 18 ties (faculty, former faculty, or graduates) to Turing awards, far more than MIT. 2) Stanford has created lots of milestone level inventions in IT which I have listed on this thread. Again, MIT's inventions in IT are less impressive to say the least.

[/quote]
</p>

<p>Really? I would like to see evidence that Stanford is the only one ranked #1 "in all years". Do you have evidence of this? </p>

<p>I aso see that in the new 2008 USNews graduate edition (just released today), MIT is ranked higher than Stanford in EE and compE (still tied in CS because they didn't rerank CS this year). So Stanford 'wins' in the NRC, and MIT 'wins' in USNews. So what are you trying to prove?</p>

<p>I see that you also glided over what I said before regarding Berkeley and Caltech. Are you afraid to say that the rankings say that Berkeley and Caltech are clearly not as good as MIT is? After all, MIT beats Berkeley and Caltech in both the NRC and USNews in nearly all categories. </p>

<p>
[quote]

And this has a huge direct impact on its US NEWS ranking in engineering. For example, the 3 categories US NEWS used for engineering school ranking: the # of ph.d produced, the total graduate students enrolled, and the total research expenditures in engineering obviously line up with the size, and helps MIT secure the #1 ranking.

[/quote]
</p>

<p>Uh, no, not really. Stanford actually has MORE total graduate engineering students enrolled. Yes, more (about 3200 at Stanford compared to about 2600 at MIT). </p>

<p>
[quote]
MIT has most patents because it is the biggest engineering school. It also has the history advantage. This is something that separates MIT from its competitors such as Stanford, Berkeley, and Caltech

[/quote]
</p>

<p>I would hardly blame it on a specific 'history advantage' per se. MIT is only 3 years older than Berkeley is, and about 30 years older than Stanford or Caltech is. That's not exactly a whole lot of time. Furthermore, I highly doubt that MIT built up a large patent advantage in its first 30 years of existence. Keep in mind that MIT almost went bankrupt in its early years. It was only during WW2 and afterwards that MIT (as well as Stanford) really hit its stride. Before that, Caltech and Berkeley were clearly stronger research schools. </p>

<p>So to me, there is no discernable "history advantage" that MIT has over the other 3 schools.</p>

147

<p>This report presents a preliminary list of the U.S. universities receiving the most patents for invention (i.e., utility patents) during the 2003 calendar year. All campuses are included.</p>

<p>Rank in 2003* Number of Patents in 2003* Organization* (Rank
in 2002) (Number of Patents in 2002)
1 439 University of California (1) (431)
2 139 California Institute of Technology (3) (110)
3 127 Massachusetts Institute of Technology (2) (135)
4 96 University of Texas (5) (93)
5 85 Stanford University (4) (104)</p>

148

<p>I want to just say anyone who says Princeton and Harvards engineering schools are on par with MIT when normalized for size is an idiot. Princeton's engineering is the furthest building off of campus, every single student basically is required to own a bike just so they can get to classes. My brother went there, and when I was applying to schools said this specifically, Princeton's engineers are the most disrespected students on campus. He liked Princeton fine, but refused to allow me to go there for engineering because as he described it, engineering is just not a priority there at all, and I would have been miserable there.</p>

149

<p>"Maybe you are talking about Rodney Brooks? He is a huge name in AI, the director of MIT CS and AI lab. By the way, both Rivest anf Brooks are Stanford trained Ph.d."</p>

<p>Yeah, you're right. I was thinking of Rodney Brooks.</p>

150

<p>


What are you talking about? After coming up with the idea of the Galactic Network, Licklider than became a main mover at DARPA, where he convinced a number of people, including Roberts, Ivan Sutherland, and Bob Taylor, that the concept of an all-encompassing worldwide network was important work to do. Hence, he was the original inspiration for the Internet in the first place! How can you call that 'nothing'?

Again, Licklider contributed nothing to packet switching and TCP/IP, which govern today's internet.

``` Way to completely ignore Robert's contribution or to the creation of the Arpanet in the first place. I would calculate it as follows:

Roberts and Kleinrock = each 1/2 of the Arpanet = 1 Cerf = 1/2 of TCP/IP = 1/2

Hence, Roberts + Kleinrock is still more than Cerf. That is, unless you want to argue that the development and design of the original Arpanet is just not important at all. Are you prepared to argue that? ``` If Roberts or Kleinrock can win top notched prizes such as Turing award, national medal of technology, and medal of freedom, as Cerf and Kahn already did. I might agree with your argument. But honestly, I think their chance is very very small, if any.

Again, MIT's contribution < 1/3 packet switching Stanford's contribution > 1/2 TCP/IP. TCP/IP > packet switching 1/2>1/3 So Stanford > MIT in birth of internet.

By the way, I'm still waiting for you to tell me MIT's contribution to internet after its birth. How can you match Stanford's 56k MODEM DSL broadband connection multiprotocol router YAHOO GOOGLE NETSCAPE SUN work station? Look. After the birth of your baby, you need to take care of him. You can NOT just walk away.

``` Really? I would like to see evidence that Stanford is the only one ranked #1 "in all years". Do you have evidence of this?

I aso see that in the new 2008 USNews graduate edition (just released today), MIT is ranked higher than Stanford in EE and compE (still tied in CS because they didn't rerank CS this year). So Stanford 'wins' in the NRC, and MIT 'wins' in USNews. So what are you trying to prove?



I can not find the link now that Stanford was ranked ahead of MIT CS. But MIT has never been ranked ahead of Stanford in CS by US NEWS, which I am 100% sure.&lt;/p&gt;

&lt;p&gt;Stanford beats MIT in CS for the following reasons:&lt;/p&gt;

&lt;p&gt;1) Stanford has more Turing awards ties (18). MIT can not match that.
Will you be able to privide me MIT's Turing award list?&lt;/p&gt;

&lt;p&gt;2) Stanford beats MIT in NRC ranking in CS&lt;/p&gt;

&lt;p&gt;3) Stanford has created more mile stones: microprosessor, DSL, digital music synthesis, RISC, STANLEY driverless car, Stanford arm, Stanford cart, multiprotocol router, and etc.&lt;/p&gt;

&lt;p&gt;4) Stanford has contributed more to IT industry: HP, GOOGLE, YAHOO, NETSCAPE, SUN, CYSCO.&lt;/p&gt;

&lt;p&gt;5) look at another link: the world leading INFO TECH LABs from survey on IT professionals conducted by bussiness week. Again, Stanford CS is #1.
 &lt;a href="http://www.businessweek.com/1997/25/b353224.htm%5B/url%5D"&gt;http://www.businessweek.com/1997/25/b353224.htm&lt;/a&gt;&lt;/p&gt;
151

<p>The 'X-Lab' List (Overall Ranking)
BUSINESS WEEK's poll included this question: If you were 35 and had just won the first Nobel Prize for Information Technology, triggering invitations to the lab of your choice, which one would you pick? Most researchers did not choose the lab where they work. Here are the results:</p>

<p>



1.      Stanford University</p>

<ol>
<li><pre><code> AT&T Labs
</code></pre></li>
<li><pre><code> Bell Labs (Lucent)
Carnegie Mellon University
</code></pre></li>
<li><pre><code> MIT Lab for Computer Science
Santa Fe Institute
</code></pre></li>
<li><pre><code> Microsoft Research
</code></pre></li>
<li><pre><code> University of California-Berkeley
</code></pre></li>
<li><pre><code> MIT Artificial Intelligence Lab
MIT Media Lab
</code></pre>

<p>

</p></li>
</ol>

<p><a href="http://www.businessweek.com/1997/25/b353221.htm%5B/url%5D"&gt;http://www.businessweek.com/1997/25/b353221.htm&lt;/a&gt;&lt;/p>

152

<p>The 'X-Lab' List (Ranked by Category)</p>

<p>Business Week's poll included this question: If you were 35 and had just won
the first Nobel Prize for Information Technology, triggering invitations to any
lab of your choice, which one would you pick? Most researchers didn't chose the
lab where they work. Here are the complete results, with labs ranked within
four separate research categories: </p>

<p>


COMPUTER SCIENCE

1st Choice: Stanford University

2nd Choices: Microsoft Research, University of California-Berkeley

3rd Choice: Massachusetts Institute of Technology (MIT) Laboratory for Computer Science (LCS)

4th Choice: Carnegie Mellon University

5th Choices: AT&T Labs, Bell Labs (Lucent), MIT Media Lab, Xerox Palo Alto Research Center (PARC)

``` TELECOMMUNICATIONS, NETWORKING, GROUPWARE ```

1st Choice: Bell Labs (Lucent)

2nd Choice: Stanford

3rd Choices: MIT LCS, Xerox PARC

4th Choice: AT&T Labs

5th Choice: University of Southern California

``` ARTIFICIAL INTELLIGENCE, ROBOTICS, SPEECH, INTERFACES ```

1st Choice: Stanford

2nd Choices: Carnegie Mellon University, MIT AI Lab

3rd Choices: Microsoft Research, Xerox PARC

4th Choices: AT&T Labs, IBM Research, MIT Media Lab

5th Choices: CMU Robotics Institute, Swiss Machine Learning Research Institute (IDSIA), MIT LCS, University of Michigan


BIOLOGICALLY INSPIRED: ARTIFICIAL LIFE, GENETIC ALGORITHMS

</p>

<p>1st Choice: Santa Fe Institute</p>

<p>2nd Choice: Stanford</p>

<p>3rd Choice: Swiss Federal Institute of Technology (EPFL, Lausanne)</p>

<p>4th Choices: IDSIA, University of North Carolina, University of Sussex (UK)</p>

<p>5th Choices: Natural Selection Inc., Naval Research Lab, UC-San Diego,
University of Illinois</p>

153

<p>"The 'X-Lab' List (Overall Ranking)</p>

<h2>BUSINESS WEEK's poll included this question: If you were 35 and had just won the first Nobel Prize for Information Technology, triggering invitations to the lab of your choice, which one would you pick? Most researchers did not choose the lab where they work."</h2>

<p>I don't know whether this is a great question...
A lot of people at MIT don't like it there, but are there because of the research opportunities <em>and</em> prestige. If they won the Nobel Prize, they might leave for a place that is not as prestigious but has a higher quality of life. Do you honestly believe that a majority of profs at MIT could not move to Stanford (and vice versa) if they wanted to?</p>

<p>As for US News & World Report, I think you are wrong that Stanford has ranked #1 more often than MIT. I generally peruse the rankings, and from memory it is more common that MIT would be alone at #1 than Stanford, although recently they seem to be tied. </p>

<p>Right now, they are tied at #1 with Stanford, Berkeley, and Carnegie-Mellon in the overall Comp Sci. rankings. MIT is #1 in Computer Engineering (considered separate from Comp Sci.) and #1 at AI and Theory. </p>

<p>AI </p>

<h1>1 MIT</h1>

<h1>2 Stanford</h1>

<p>Programming Languages </p>

<h1>1 Carnegie Mellon</h1>

<h1>2 Stanford/Berkeley</h1>

<p>Systems<br>
1. Berkeley<br>
2. MIT<br>
3. Carnegie-Mellon</p>

<p>Theory
1.MIT
2.Stanford</p>

154

<p>


A lot of people at MIT don't like it there, but are there because of the research opportunities <em>and</em> prestige. If they won the Nobel Prize, they might leave for a place that is not as prestigious but has a higher quality of life. Do you honestly believe that a majority of profs at MIT could not move to Stanford (and vice versa) if they wanted to?

In terms of quality of life, you think Boston is worse than the bay area? My impression from this forum is the opposite though. No matter what you say, in any of these 4 catogories, Stanford beats MIT consistently according to the poll from IT professionals. Notice that in the 4th category: BIOLOGICALLY INSPIRED: ARTIFICIAL LIFE, GENETIC ALGORITHMS, MIT is NOT even in the top 5. You can NOT just blame that Boston's quality of life is not so good.

Notice that in computer science, Berkeley started to beat MIT. Is that because the bay area has a better quality life than Boston, hence it is more attractive than MIT? If your answer is yes, then how come in some other subfields, MIT is ahead of Berkeley?

In terms of faculty reputation by NRC ranking, Stanford (#1, scored 4.97) beats MIT (#2, scored 4.91). Since people (peers in CS) who evaluated the CS programs used a scale of 1-5. 5 means distinguished, 4 means very strong. So Stanford's 4.97 can be explained as 97% of the people who participated the evaluation gave Stanford a 'distinguished' score, while 3% gave Stanford a 'very strong' score. But for MIT, only 91% of the people gave it a 'distinguished' score, while 9% gave it a 'very strong' score. That is the difference. 


As for US News & World Report, I think you are wrong that Stanford has ranked #1 more often than MIT. I generally peruse the rankings, and from memory it is more common that MIT would be alone at #1 than Stanford, although recently they seem to be tied.

</p>

<p>I want to say that again, MIT has NEVER NEVER been ranked ahead of Stanford in US NEWS CS (overall) ranking and Stanford has been ranked #1 for all years. If you can find a single year when it is NOT the case, I am willing to concede. </p>

<p>US NEWS subfields ranking are generally questionable to IT professionals. And it only lists very few fields. For example, US NEWS ranked the subfield "database" before, on which Stanford is extremely strong while MIT is weak. But they droped that ranking in recent years. I don't know why.</p>

155

<p>I'm curious about those who seem to be spending many long minutes researching and assembling posts lauding Stanford on an MIT discussion board, while denigrating the quality of MIT academics (in color even!) -- do y'all think you're convincing any students to go to Stanford instead of MIT? Or do y'all just enjoy arguing at each other, hoping to "win"?</p>

<p>I'm an IT professional and researcher. Graduates of MIT or Stanford would both be right at the top of any list when we were looking to hire. Don't be silly and don't bother wasting your time: they're both fantastic schools, they provide unparalleled educations, they both offer unique academic and social environments, and anyone considering between the two has a surfeit of riches. I congratulate them.</p>

156

<p>If you don't like the post, you don't have to read.</p>

157

<p>I agree with mootmom, for many its the culture of these schools, the city etc.</p>

<p>And datalook, you look like nothing more than a Stanford troll to me, I mean I see you lauding Stanford on like every thread with even the slightest mention of it.</p>

158

<p>You study too much CS and not enough economics, datalook. By acquiring such a thick and greasy patina of bias, you cause people to Bayes update their expectations about you and discount your nonsense babble appropriately. :-D.</p>

159

<p>By quality of life, I meant MIT's rather bleak campus in Cambridge and the fact that generally it is a higher stress environment than any other place. I wasn't referring to Boston. </p>

<p>Anyway, I haven't been following the US News rankings in CS religiously; it is surprising to me that MIT has never been ranked #1 by itself. Right now MIT and Stanford is tied overall.</p>

<p>As for the subspecialty rankings being not respected, I really have no idea. I know in chemistry the subspecialties are fairly accurate. The idea that the US News & World Report is somehow gifting MIT's high ranking is really bizarre. Let's look at the undergrad rankings. Do you think it's an accident that the SAT component is reported as the 25%-75% range instead of just the mean? If it was the mean that was reported, MIT and CalTech would easily outscore Stanford and the ivies. Also, they get hurt by retention rate since it is so much harder to graduate from MIT.</p>

160

<p>Milki and Ben Golub,</p>

<p>It doesn't matter if datalook is a Stanford troll or not. I didn't attend Stanford. I might be more of a Stanford fan. Try to refute my arguments, not attack me.</p>

<p>By the way, I am not even a CS major.</p>